CROJFE - Volume 34, Issue 1

Original scietific paper – Izvorni znanstveni rad

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR Data of Funyu Experimental Forest Masashi Saito, Msahiro Goshima, Kazuhiro Aruga, Keigo Matsue, Yasuhiro Shuin, Toshiaki Tasaka Abstract – Nacrtak In this study, a model to automatically design a forest road considering shallow landslides using LiDAR data was examined. First, in order to develop a shallow landslide risk map of the Funyu Experimental Forest, a slope stability analysis was carried out using the infinite slope stability analysis formula. The soil depth was surveyed at 167 points using simple penetration tests, and the frequency distributions of the soil depth were estimated as logarithmic normal distributions. A soil depth map of the experimental forest was made using the mode values of the lognormal distributions. Then, shallow landslide risk maps were also made for the experimental forest by a slope stability analysis using these soil depth distributions. Finally, an automatic forest road design model was developed with a LiDAR based highly accurate Digital Terrain Model (DTM) and shallow landslide risk map using cubic spline interpolation and dynamic programming. The program has capability of minimizing the earthwork costs while avoiding shallow landslide risk areas. The program can be effectively used to design an environmentally sound low volume road automatically.

Keywords: LiDAR, automatic forest road design, shallow landslide, spline interpolation, dynamic programming

1. Introduction – Uvod To enable more efficient and stable timber extraction from forest resources, as well as sustainable forest management, considering the perspective of public functions such as land and watershed conservation and climate change mitigation, it is necessary to develop forest road networks in Japan. Extensive field investigations, including a preliminary survey, route survey, and cross-section survey, are necessary to design a forest road. This entails considerable time and cost. Work experience is also necessary to select the best route for a forest road from many alternatives. It is not easy even for an advanced engineer to examine both an earthwork and slope failure on site. If a design model could be developed to examine both an earthwork and slope failure and select the best route from many alternatives, it would help forest engineers design forest roads.

Croat. j. for. eng. 34(2013)1

A variety of forest road design supporting models to reduce the necessary workload have been developed using DTMs (Digital Terrain Models). Reutebuch (1988) developed a computer program for the preliminary route location, ROUTES, which estimated grades and distances along a possible road route using DTM. Liu and Sessions (1993) developed a preliminary planning model for road systems that designed the route while estimating the establishment cost, maintenance expense, and transportation cost using DTM. Douglas and Henderson (1988) optimized a forest road route location using a dynamic programming approach, and Suzuki et al. (1998) optimized a forest road route location using a hybrid model that combined the Dijkstra method and genetic algorithms. These models considered only the optimization of horizontal arrangements.

1

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

It is necessary to examine not only the horizontal arrangement but also the vertical arrangement to design the best route. Dynamic programming has been widely used in the literature and in practice to optimize the vertical arrangement (Kanzaki 1973). Heuristic combinatorial optimization models, simulated annealing, genetic algorithms, and tabu search have been studied to minimize the construction and maintenance costs by optimizing the vertical arrangement (Akay 2006; Ichihara at al. 1996; Aruga et al. 2005) using DTM. However, most of these models have demonstrated a lack of accuracy because of the low reproducibility of geographical features. Therefore, it has been difficult to adjust the design to the site.

In order to improve the geographical features, Aruga et al. (2006) developed a forest road design model using LiDAR (Light Detection and Ranging) data, which demonstrated a significant improvement in representing relatively accurate geographical features. However, no comparison with field investigations was given in that paper. Therefore, we developed a forest road design model using LiDAR data for the Funyu Experimental Forest and conducted a comparison with field investigations (Saito et al.

2007). As a result, it was found that the ground surfaces produced by the LiDAR data represented the ground surfaces well and the forest road design results using the LiDAR data were similar to the established forest road.

Slope failure, which is an important constraining factor in forest road design and establishment in the mountainous areas of Japan, was not considered in the study (Saito et al. 2007). Umeda et al. (2007) and Suzuki et al. (2007) carried out numerous field surveys of spur road networks and clarified the terrain conditions required for constructing spur road networks. Umeda et al. (2007) implied that a detailed DTM generated from airborne laser measurements could extract the terrain conditions required for constructing spur road networks. Yoshimura (1997) predicted the collapse risk places from the results of field investigations using fuzzy integration and planned forest road networks considering the environmental impact of failure and soil erosion using the Dijkstra method. However, his study did not focus on forest road design.

In this research, shallow landslide risks were determined using an infinite slope stability analysis.

Then, an automatic forest road design model was developed considering shallow landslide risks while accurately calculating the earthwork volumes and costs using the detailed topographical information of a high resolution DTM generated from LiDAR.

2

2. Study site and method – Mjesto i metoda istraživanja 2.1 Study site – Mjesto istraživanja The study site was around the terminal point of the main forest road at the Funyu Experimental Forest of Utsunomiya University in Japan (Fig. 1 and 2). The vegetation around the study site was a mixed forest composed of cedar, cypress, pine, oak, azalea, and maple. A high resolution (1 m grid) DTM was made by processing the LiDAR data using the intersection angle method, which was a new technique developed by Saito et al. (2008) to create ground surfaces from the raw LiDAR data. This method reproduces geographical features more clearly than the 1 m grid DEM, which an aerial survey company generated using the roller method and manual filtering in the Utsunomiya University Forest.

At the study site, the operation system included falling with a chainsaw, processing with a processor, yarding with a tower yarder, and transporting using a truck with a 4 t loading capacity. On terrain where it

Fig. 1 Funyu Experimental Forest of Utsunomiya University Slika 1. Pokusni šumski object Funyu, Utsunomiya sveučilišta Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

M. Saito et al.

used instead of the 1 m grid DTM because the 1 m grid DTM was too detailed for a slope stability analysis.

The area where the safety rate (F) shown in the following equation became one or less was judged to be the area of slope failure potential.

F=

Fig. 2 Study site with high resolution DTM Slika 2. Područje istraživanja prikazano digitalnim modelom reljefa visoke rezolucije was difficult to construct forest roads, spur roads were constructed from landings to forest roads and a forwarder was used for forwarding. The starting point of the planned forest road was set as the terminal of the main forest road established in 2003, while the terminal point was the terminal of the main forest road established in 2005 (Fig. 2). The length and width of the planned forest road were about 850 m and 4 m, respectively. This forest road was a main road with full bench construction and planned to be used permanently.

c + (g s ´ h -g w ´ hw) ´ cos 2 q tan j g s ´ h ´ cos q sin q

(1)

Where: c soil cohesion, N/m2 gs soil density, kg/m3 h soil depth, m γw water density, kg/m3 hw groundwater level, m q slope angle, ° j soil internal frictional angle, ° In this study c, j, and γs were assumed to be 1 730 N/m2, 30°, and 2 000 kg/m3, respectively, from the classification of the surface soil in the Funyu Experimental Forest as sandy soil (Goshima et al. 2008). γw is 1 000 kg/m3. θ was calculated from the 10 m grid DTM. hw was calculated using the probable rainfall intensity for 1 h calculated by the fair formula from the Automated Meteorological Data Acquisition System’s probable rainfall intensity calculation program for every return period (Public Work Research Institute 2010). The return periods used covered a range of 10 to 100 years, with 10 year increments (Table 1). h was estimated from the values found in a previous study where 167 points were evaluated using simple penetration tests (Goshima et al. 2008). This test can explain a vertical change in the resistance of the soil layer based on the number of times a 5 kg hammer has to fall from a 50 cm height (Nc value) to penetrate the cone 10 cm into the soil. The soil layer depth was determined from an Nc value of 20 or less because the basement geology of the Funyu Experimental Forest belongs to the Neogene Tertiary Formation and the Kanto loam soil layer (Ohsaka et al. 1987). The spatial distribution of the soil depths was estimated using the method by Iida et al. (2005). They paid attention to the inclination and average depth of the water catchment area, and estimated soil depths as logarithmic normal distributions with five classes of slope angles and four classes of average depths for a water catchment area.

2.2 Slope stability analysis – Analiza stabilnosti pokosa

2.3 Process for automatic determination of route Postupak automatskog projektiranja šumske ceste

A slope stability analysis was used to determine the slope failure potential distributions. The infinite slope stability analysis formula with the 10 m grid DTM was

The route determination method using the cubic spline interpolation proposed by Tasaka et al. (1996) balanced cut and fill materials and reduced the earth-

Croat. j. for. eng. 34(2013)1

3

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

Table 1 Rainfall intensities and slope failure areas Tablica 1. Intenzitet padalina i područja pojave klizanja kosina Shallow landslide risk areas, ha Rainfall duration, h

Return period, year

Rainfall intensity, mm/h

Trajanje oborina, h

Vrijeme promatranja, godina

Intenzitet oborina, mm/h

Rizična područja s obzirom na mogućnost pojave klizišta, ha Lognormal distributions

Survey data

Logaritamska normalna distibucija

Terenski podaci

1

10

50.28

51.18

68.42

1

20

59.21

53.34

70.58

1

30

65.16

59.26

76.51

1

40

69.74

57.11

78.66

1

50

73.51

64.65

80.82

1

60

76.77

68.96

85.66

1

70

79.59

51.18

89.44

1

80

82.13

71.66

93.21

1

90

84.45

73.81

95.90

1

100

86.58

75.43

96.98

work volumes. Therefore, we introduced this method to the program for the automatic forest road design model using the DTM generated from LiDAR data

(Saito et al. 2009). In addition, this method has preferred because its process time was very short, even on a personal computer.

Fig. 3 Forest road design process Slika 3. Postupak projektiranja šumske ceste

4

Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

M. Saito et al.

Table 2 Unit costs (Nihon Rindo Kyokai 2004) Tablica 2. Troškovi (Nihon Rindo Kyokai 2004) Item – Radni zahvat

Machine – Stroj

Unit cost, soil

Unit cost, rock

Jedinični trošak, materijal C kategorije

Jedinični trošak, materijal A kategorije

Bucket excavator – Bager

235 yen/m3

339 yen/m3

Smoothing – Profiliranje planuma

Bulldozer – Buldozer

140 yen/m3

140 yen/m3

Compacting – Sabijanje

Bulldozer – Buldozer

108 yen/m3

108 yen/m3

Transporting – Transport

Dump truck – Kamion kiper

435 yen/m3

552 yen/m3

Fill slope greening – Stabiliziranje pokosa iskopa

–

1 083 yen/m2

1 083 yen/m2

Cut slope greening – Stabiliziranje pokosa nasipa

–

1 860 yen/m2

0 yen/m2

Retaining wall – Potporni zid

–

16 000 yen/m2

16 000 yen/m2

Cutting – Iskop

The procedure is as follows (Fig. 3): Þ The starting and terminal points are determined; Þ A contour line is generated from the starting point; Þ The nearest point, P, on the contour line from the terminal is determined; Þ The difference in elevation, h, between the terminal point and P is calculated; Þ The length, l, along the contour line is calculated from the starting point to P;

Þ The difference, h, is distributed proportionally in the distance Δhi for each point i, where these points are located at 20 m intervals along the contour line; Þ Each point moves to a point in the steepest slope direction so that the elevation of the new point becomes the same as the elevation of each point added by the elevation difference on each point, Δhi. The new point is considered to be the passing nomination point;

Fig. 4 Minimum earthwork cost route search process Slika 4. Postupak pronalaženja trase šumske ceste s najmanjim troškovima zemljanih radova Croat. j. for. eng. 34(2013)1

5

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

The cubic spline interpolates between the nomination points at 20 m intervals, and a temporary route is determined.

Then, the program searches for the minimum earthwork cost route based on the temporary route using dynamic programming as follows (Fig. 4): Þ The curvature radius of each nomination point on the internal solution route is determined; Þ Nomination points that have local minimum curvature radius values are determined to be the base points for searching for the minimum earthwork cost route using dynamic programming (Kanzaki 1973); Þ Candidate points are generated on both sides of the base points on the normal line ranging from 1 m to 10 m at 1 m intervals; Þ The minimum earthwork cost route is determined using dynamic programming by selecting from the combinations of these candidate points connected with the cubic spline interpolation.

2.4 Earthwork cost estimation – Procijenjene vrijednosti zemljanih radova In order to calculate the earthwork costs, the unit costs listed in Table 2 were used.

The soil type influenced the work efficiency of the forest road establishment. The road width was 4.0 m, the fill slope was 1:1 (45°), and the cut slope was 1:0.8 (51°) for soil and 1:0.3 (73°) for rock. When the slope length became 3 m or more, the slope was assumed to be made by a retaining wall with a slope of 1:0.2 (Fig. 5). The earthwork volumes were estimated using the average end section method. Cut and fill areas were estimated from the differences between the forest road cross sections designed by the program and the ground surfaces of DTM while dividing each section between adjacent candidate points into 10 individual sections. It was assumed that a soil pit or waste site was 2 km away from the construction site.

In a previous study designed section length was 200 m and its earthwork volume was about 3 600 m3 (18 m3 per meter) (Saito et al. 2007). In this study, the designed section was extended to 850 m. The extended 650 m section was measured by LiDAR in 2003 before the forest road establishment and again after the forest road establishment in 2005. Therefore, the earthwork volumes of the established forest road were estimated from the difference between the LiDAR data measured before and after the forest road establishment. The forest road was constructed with a full bench. Therefore, it was assumed that the differences in the DTM between 2003 and 2005 were cutting volumes.

6

Fig. 5 Cross-section considering soil depth Slika 5. Poprečni presjek s obizrom na dubinu tla

2.5 Shallow landslide risk consideration Rizična područja s obzirom na mogućnost pojave klizišta In this research, the shallow landslide risk map was overlapped on the road map to search for the route while avoiding a forest road failure. A failure was assumed to occur when the forest road passed over shallow landslide risk areas. In this case and the forest road was then assumed to be established again. It was also assumed that the restoration costs would be the same to 10 times as much when passing over shallow landslide risk areas based on the probable rainfall intensities for 100 to 10 years, with 10 year steps.

In order to examine the effect of the usage period on forest road design, some limited usage periods were considered in the analyses, although permanent use was assumed for this main forest road. The usage period for the spur roads was assumed to be 10 years.

In this case, the reconstruction costs for shallow landslide risk areas based on each probable rainfall intensity, excluding 10 years, were set to zero because it was assumed that the spur roads were not to be used for more than 10 years. Furthermore, in order to examine the effect of the strength of the forest road structure on the forest road design, it was assumed that no slope failure occurred by some years probable rainfall intensity. It was difficult to assess the strength of the forest road structure in relation to failure because natural conditions and construction techniques affected it.

Therefore, the 10 year probable rainfall intensity was assumed as the probable rainfall intensity by which no slope failure occurred in just a test case of this study although the effect of the strength of the forest road structure on forest road design should be examined in Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

M. Saito et al.

Table 3 Results of soil depth survey, cm Tablica 3. Rezultati istraživanja dubine tla, cm Catchment depth – Vodonosni sloj Number of samples – Broj uzoraka

35–45 °

45–55 °

51

79

14

0

130.8 (94.4)

86.1 (52.8)

–

311

171

101

58

32

–45

–19.8

29.8

28.1

–

3

6

9

3

0

182.9 (134.5)

79.8 (25.4)

62.1 (121.6)

67.1 (29.7)

–

217

117

68

38

21

–34.1

–37.2

94.1

29.1

–

2

7

6

0

0

83.1 (1.9)

48.3 (111.3)

159.6 (81.7)

–

–

152

81

46

26

14

–68.9

67.3

113.6

–

–

Number of samples – Broj uzoraka

0

1

0

0

0

Survey average, (SD) – Srednja vrijednost (Sd)

–

116

–

–

–

106

55

31

17

9

–

61

–

–

–

Theoretical lognormal distribution

Number of samples – Broj uzoraka Survey average, (SD) – Srednja vrijednost (Sd) Theoretical lognormal distribution Teoretska logaritamski normalna distibucija Difference – Razlika Number of samples – Broj uzoraka Survey average, (SD) – Srednja vrijednost (Sd) Theoretical lognormal distribution Teoretska logaritamski normalna distibucija Difference – Razlika

2 000–20 000 m

25–35 °

151.2 (89.8)

Difference – Razlika

200–2 000 m

15–25 °

16

Teoretska logaritamski normalna distibucija

20–200 m

0–15 ° 266 (95.5)

Survey average, (SD) – Srednja vrijednost (Sd) 0–20 m

Inclination – Nagib

Theoretical lognormal distribution Teoretska logaritamski normalna distibucija Difference – Razlika

a future study. In this case, the reconstruction cost for slope failure caused by the 10 year probable rainfall intensity was set to zero.

3. Results and discussion – Rezultati s raspravom 3.1 Soil depth – Dubina tla With regard to the soil depth, large influences were expected from terrain factors such as the elevation, inclination, water catchment area, etc. Negative correlations were found between the soil depth and slope angles for the average depth of the water catchment area (Table 3). Although the survey points were not evenly distributed on the geographical features, the results were similar to our expectations.

In order to determine the spatial distribution of soil depths, the method by Iida et al. (2005) was applied to the Funyu Experimental Forest.

Croat. j. for. eng. 34(2013)1

Fig. 6 shows a comparison of the measured values of soil depth and the theoretical lognormal distribution at an inclination of 25–35° and an average depth for the water catchment area of 0–20 m. Table 3 lists the averages of the survey results, the mode values of the lognormal distributions, and the differences for each category. The soil depth comparisons were relatively consistent, and this method could approximate the soil depth by a lognormal distribution, although the maximum and average errors were 152.0 cm and 15.0 cm, respectively (Table 3). Then, a soil depth map for the experimental forest was made using the mode values of the lognormal distributions as the estimated values of soil depth (Fig. 7).

3.2 Slope failure – Klizanje kosina Shallow landslide risk area distributions were estimated using the estimated soil depths and the different probable rainfall intensities (Table 1). As an example, the shallow landslide risk areas predicted

7

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

Fig. 6 Example of lognormal distribution and soil depth frequency distribution of survey results Slika 6. Primjer logaritamske normalne distibucije i distribucije dubina tla na istraživanom području

Fig. 7 Estimated soil depth map, white frame: study site Slika 7. Karta procjenjenih dubina tla, bijeli okvir: područje istraživanja

using the 50 year probable rainfall intensity are shown in Fig. 8. The shallow landslide risk areas estimated using the mode values of the lognormal distributions

8

Fig. 8 Shallow landslide risk map based on 50 year probable rainfall intensity Slika 8. Karta pojavnosti klizišta na temelju predviđenog 50 godišnjeg intenziteta padalina Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

M. Saito et al.

Fig. 9 Forest road design considering soil depth, right: close-up of frame in the figure on the left Slika 9. Projektiranje šumske ceste na temelju dubine tla, desno: uvećani prikaz iz okvira s lijeve strane as soil depths were 12% of the Funyu Experimental Forest, 538.77 ha, while those estimated using the average soil depth of the survey results were 15%. 16% of the shallow landslides that occurred in 1998 were predicted using the mode values of the lognormal distributions as soil depths (Fig. 8), while 11% were predicted using the average soil depth of the survey results. Thus, using the mode values of the lognormal distributions as soil depths improved the shallow landslide risk prediction.

3.3 Forest road design considering soil depth Projektiranje šumske ceste s obzirom na dubinu tla The estimated earthwork volume for the established forest road was 13 487 m3, while that estimated by the program was 14 162 m3, which was 21.8 m3 per meter. Thus, the program estimated the earthwork volume accurately. Although the forest road was conCroat. j. for. eng. 34(2013)1

structed with a full bench, the planned forest road after this section was designed by the program with a balance of cut and fill materials in order to reduce the earthwork volumes and costs (Saito et al. 2007).

The horizontal alignments designed by the program considering the soil depth did not change significantly compared to those for the established forest road. In Fig. 9, »The soil depth was not considered« means that the soil depth was not considered in the route determination, but costs were estimated with considering soil depth.

However, some sections shown in Fig. 9 were changed so that the forest road avoided the thick soil depth areas where the earthwork volumes, and thus earthwork costs, were relatively high, even though the unit costs were lower. Although the gradients of the forest road considering soil depth became large in areas to avoid these thick soil depths, these were within forest road regulations. The maximum gradi-

9

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

Fig. 10 Profile considering soil depth Slika 10. Uzdužni profil šumske ceste s obzirom na dubinu tla ent was 12.2% on the section from 268 m to 308 m (Fig. 10).

In a comparison of cross-sections, the forest road considering soil depth passed over areas of thin soil depth, where the amount of soil to be cut was small.

Moreover, because the majority of the cutting slopes consisted of rocks, little slope protection was necessary (Fig. 11).

In a comparison of the earthwork volumes, the amount of rock increased, and the amount to be transported increased (Table 4).

Therefore, the earthwork costs increased (Table 5).

However, the slope protection cost was reduced, and the total cost was also reduced. The program could find the minimum cost route by searching for areas with thin soil depths.

Fig. 11 Cross section, left: soil depth was considered; right: soil depth was not considered Slika 11. Poprečni presjek šumske ceste, lijevo: u obzir je uzimana dubina tla; desno: dubina tla nije uzimana u obzir

10

Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

Table 4 Earthwork volumes, m3 Tablica 4. Količina zemljanih radova, m3

Cutting volume, soil Volumen iskopa C kategorije Cutting volume, rock Volumen iskopa A kategorije Cutting volume, total Ukupni volumen iskopa Filling volume Volumen nasipa

Soil only

Soil and rock

Materijal C kategorije

Materijal B kategorije

2 366.44

1 977.86

2 256.56

2 745.22

4 623.00

4 723.08

4 681.63

4 481.99

58.64

241.09

Transporting volume Volumen transportiranog materijala

M. Saito et al.

designed by the program bypassed only the shallow landslide risk areas using the 10 and 20 year probable rainfall intensities, the routes were almost the same shape and hardly any differences were observed in a comparison of the profiles and cross-sections.

In order to examine the effect of the usage period on forest road design, it was assumed that the forest road was only used for 10 years. The reconstruction costs for slope failure caused by each probable rainfall intensity, excluding 10 years, were set to zero. The forest road alignments did not change significantly because the forest road designed with a consideration of shallow landslide risk areas using just the 10 year

Table 5 Costs considering soil depth Tablica 5. Troškovi s obzirom na dubinu tla Soil only

Soil and rock

Materijal C kategorije

Materijal B kategorije

¥ 2 339 098

¥ 2 450 479

¥ 19 570 977

¥ 16 309 148

¥ 25 626

¥ 21 839

Earthwork cost Troškovi zemljanih radova Slope protection cost Troškovi zaštite pokosa Cost per meter Troškovi po dužnom metru

3.4 Forest road design considering slope failure Projektiranje šumske ceste s obzirom na klizanje kosina Fig. 12 shows the results for a forest road design with a shallow landslide risk map. The establishment costs of a forest road design without avoiding shallow landslide risk areas was about 5 000 yen per meter cheaper than when trying to avoid the shallow landslide risk areas (Table 6).

However, the distance of forest road on the shallow landslide risk areas using the 10 year probable rainfall intensity was significantly reduced when avoiding the shallow landslide risk areas (Table 7). Therefore, the restoration costs were reduced, and the total costs were also reduced (Table 6). Because the forest road Croat. j. for. eng. 34(2013)1

Fig. 12 Forest road design considering shallow landslide risk Slika 12. Projektiranje šumske ceste s obzirom na mogućnost pojave klizišta

11

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

probable rainfall intensity only bypassed the shallow landslide risk areas based on the 10 year probable rainfall intensity (Table 7). However, the establishment costs were reduced compared to a forest road that considered shallow landslide risk areas based on all of the probable rainfall intensities and were similar to those for a forest road designed without considering shallow landslide risk areas because the route did not bypass the shallow landslide risk areas based on the 20 year probable rainfall intensity (Table 6).

Furthermore, it was assumed that a slope failure did not occur by the 10 year probable rainfall intensity, in order to examine the effect of the strength of the forest road structure on the forest road design. As a result, the forest road designed by the program did not bypass shallow landslide risk areas based on the 10 year probable rainfall intensity, but only bypassed shallow landslide risk areas based on the 20 year probable rainfall intensity (Table 7). Although, when not considering the shallow landslide risk areas based on the 10 year probable rainfall intensity, the forest road alignment was similar to that when considering shallow landslide risk areas based on all of the probable rainfall intensities, the establishment costs were reduced compared to those of the forest road that considered shallow landslide risk areas using all of the probable rainfall intensities, and were similar to those for the forest road designed without considering shallow landslide risk areas (Table 6).

Thus, the effects of the usage period and strength on the forest road design were not significant at this study site. However, this method could be used to design a forest road considering shallow landslide risk areas based on the usage period and strength of the forest road. For instance, from the viewpoint of usage periods, main forest roads should avoid shallow land-

slide risk areas based on the probable rainfall intensities for 30 years or less, while spur roads should avoid shallow landslide risk areas based on the probable rainfall intensities for 10 years or less. From the viewpoint of structures, main forest roads should avoid shallow landslide risk areas based on the probable rainfall intensities for 50 years or more, while spur roads should avoid shallow landslide risk areas based on the probable rainfall intensities for 20 years or more.

4. Conclusions – Zaključci In this study, the frequency distributions of soil depth were estimated as logarithmic normal distributions and a soil depth map for the experimental forest was created using the mode values of these lognormal distributions. The program developed in this study could find the minimum cost route by considering the soil depth and searching for areas with thin soil depth on the forest road. However, this study used only two classifications, that is, sandy soil and soft rock, even though there were various types of soils and rocks in various areas. In addition, the earthwork costs were different; for example, the excavating costs for hard rock increased because of the need for a breaker. Thus, it is important to clarify the soil type, but it is difficult to predict the soil type in large areas. Therefore, a technique for predicting the soil type in large areas by remote sensing should be researched.

Shallow landslide risk maps were also made in the experimental forest by a slope stability analysis using these soil depth distributions. Using the mode values of the lognormal distributions as soil depths improved the shallow landslide risk prediction. Finally, an automatic forest road design model that considered shal-

Table 6 Costs considering shallow landslide risk Tablica 6. Troškovi s obzirom na mogućnost pojave klizišta

Establishment cost Troškovi izgradnje Reconstruction cost Troškovi obnove Cost per meter Troškovi po dužnom metru

12

Failure was not considered Klizanje kosina nije uzimano u obzir

Failure was considered Klizanje kosina je uzimano u obzir

Only 10 years were considered U obzir je uzimano 10 godišnje razdoblje

Only 20 years were considered U obzir je uzimano 20 godišnje razdoblje

¥ 18 759 627

¥ 19 852 451

¥ 18 956 214

¥ 18 822 390

¥ 9 353 400

¥ 4 063 920

¥ 2 064 422

¥ 1 086 428

¥ 32 880

¥ 27 972

¥ 24 586

¥ 23 309

Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

M. Saito et al.

Table 7 Length of forest road on shallow landslides risk areas, m Tablica 7. Duljina cesta na područijima na kojima postoji mogućnost pojave klizišta, m Only 20 years were considered

Failure was not considered Klizanje kosina nije uzimano u obzir

Failure was considered Klizanje kosina je uzimano u obzir

Only 10 years were considered U obzir je uzimano 10 godišnje razdoblje

U obzir je uzimano 20 godišnje razdoblje

10 year probable rainfall intensity Predviđeni 10 godišnji intenzitet padalina

35.2

10.2

9.2

35.2

20 year probable rainfall intensity Predviđeni 20 godišnji intenzitet padalina

18.8

18.7

18.8

9.6

low landslide risks was developed that accurately calculated the earthwork volumes and costs using the detailed topographical information of a high resolution DTM generated from LiDAR in this research. The program could minimize the earthwork costs while avoiding shallow landslide risk areas. The program could easily design an environmentally sound lowvolume road automatically.

The effects of the usage period and risk of failure on forest road design were also examined in this study.

Forest roads were designed avoiding some types of shallow landslide risk areas using different probable rainfall intensities. This simple method of weighing the probable rainfall intensity and restoration costs could be used to design a forest road while considering shallow landslide risk areas with the usage period and risk of failure of the forest road. However, it would be difficult to clarify the effects of these weights on a main forest road or spur road design. Therefore, a future study should be conducted to clarify how many years of probable rainfall intensity and how much reconstruction costs should be used for various types of roads and areas.

In addition to shallow landslide risk, collapses that occur as the result of forest road construction should be considered (Yoshimura 1997). Although collapses caused by forest road construction are important for forest road design, they were not considered in this study because natural conditions and construction techniques affected them and their assessment was difficult. However, a high-resolution DTM might help to predict these collapses (Umeda et al. 2007). Thus, this study should also be extended in the future to consider collapses that occur as the result of forest road construction using a high-resolution DTM. Croat. j. for. eng. 34(2013)1

5. References – Literatura Akay, A.E., 2006: Minimizing total costs of forest roads with computer-aided design model. Academy Proceedings in Engineering Sciences, SADHANA 31(5): 621–633.

Aruga, K., Sessions, J., Akay, A.E., 2005: Heuristic planning techniques applied to forest road profiles. J For Res 10(2): 83–92.

Aruga, K., Tasaka, T., Sessions, J., Miyata, E.S., 2006: Tabu search optimization of forest road alignments combined with shortest paths and cubic splines. Croat J For Eng 27(1): 37–47.

Douglas, R.A., Henderson, B.S., 1988: Computer assisted forest road route location. Proceedings of the Council of Forest Engineering, 10th Annual Meeting, High Technology in Forest Engineering. Syracuse, New York, pp 201–217.

Goshima, M., Shuin, Y., Tasaka, T., Aruga, K., Matsue, K., Naito, K., 2008: Relation between shallow landslide on hill slopes and distribution of soil depth, soil structure in Funyu forest, Utsunomiya University. Bull Utunomiya Univ For 44: 15–32.

Ichihara, K., Tanaka, T., Sawaguchi, I., Umeda, S., Toyokawa, K., 1996: The method for designing the profile of forest roads supported by genetic algorithm. J For Res 1(1): 45–49.

Iida, T., Tanaka, K., 1997: The relationship between topography and soil depth measured with the portable penetration test apparatus. Trans Jpn Geomorphol Union 18(1): 61–78.

Liu, K., Sessions, J., 1993: Preliminary planning of road systems using digital terrain models. J For Eng 4(2): 27–32.

Kanzaki, K., 1983: On the decision of profile line of forest road by dynamic programming. J Jpn For Soc 55(4): 144–149.

Nihon Rindo Kyokai, 2004: Rindo Hikkei Gijyutsuhen Tokyo, 279 p.

13

M. Saito et al.

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

Ohsaka, A., Tukamoto, Y., 1987: A preliminary study on critical soil depths on forested slopes. J Jpn Revegetational Technol Soc 12(3): 1–6.

Pubic Work Research Institute, 2010: Hydrologic Engineering Research Team http://www.pwri.go.jp/jpn/seika/amedas/top.htm (accessed on April 1, 2010).

Reutebuch, S., 1988: ROUTES: A computer program for preliminary route location. USDA Gen. Tech. Publication.

PNW–GTR–216, Portland, 18 p.

Saito, M., Aruga, K., Matsue, K., Tasaka, T., 2007: Development of the forest road design technique using LIDAR data of the Funyu Experimental Forest. J For Plan 13 (Special issue): 203–206.

Saito, M., Aruga, K., Matsue, K., Tasaka, T., 2008: Development of the filtering technique of the inter section angle method using LIDAR data of the Funyu Experimental Forest. J Jpn For Eng 22(4):265–270.

acceptable type using high-resolution DTM. J Jpn For Eng 23(4): 273–276.

Suzuki, H., Ichihara, K., Noda, I., 1998: Road planning in forest for recreation. J Jpn For Eng Soc 13(3): 151–160.

Suzuki, H., Umeda, S., Yamaguchi, Y., 2007: Influence of terrain conditions on forestry road construction on ridges. J Jpn For Eng Soc 22(3): 133–142.

Tasaka, T., Ochi, S., Matsuo, T., 1996: Planning method of terrain acceptable forest-roads applying with spline function. Bull Utunomiya Univ For 32:17–26.

Umeda, S., Suzuki, H., Yamaguchi, S., 2007: Considerations in the construction of a spur road network. J Jpn For Eng Soc 22(3):143–152.

Yoshimura, T., 1997: Development of an expert system planning a forest road based on the risk assessment. PhD. Thesis, Kyoto University, Faculty of Agriculture 82 p.

Saito, M., Aruga, K., Matsue, K., Tasaka, T., 2009: Development of the automatic forest roads design system as a terrain

Sažetak

Istraživanje automatskog projektiranja šumskih cesta s obzirom na plitka klizišta pomoću LiDAR-a na pokusnom objektu Funyu U radu je istražen model automatskog projektiranja šumskih cesta s obzirom na plitka klizišta pomoću LiDAR-a.

Istraživanje je provedeno na pokusnom objektu Funyu, Sveučilišta Utsunomiya u Japanu, za koje je napravljen, upotrebom »intersection angle method«, digitalni model terena visoke rezolucije (1 m mreža). Prvo, kako bi se napravila karta plitkih klizišta na području istraživanja, napravljena je analiza stabilnosti nagiba pomoću formule za izračuna stabilnosti (1). Dubina tla je mjerena na 167 mjesta pomoću jednostavnih penetracijskih testova, te je frekvencija dubine tla procjenjivana na osnovu logaritamske normalne distribucije. Karta dubine tla na pokusnom objektu napravljena je korištenjem vrijednosti logaritamske distribucije. Također, napravljena je karta plitkih klizišta za pokusni objekt pomoću analize stabilnosti nagiba s obzirom na dubinu tla. Nakon toga, pomoću programa za automatsko projektiranje, na osnovu digitalnog modela reljefa, postavljena je trasa ceste koristeći metodu interpolacije zemljanih radova po dužnom metru. Takva metoda je korištena zbog kratkog vremena obrade podataka na osobnom računalu, te zbog toga što omogućava smanjenje zemljanih radova, prebacivanjem materijala iz iskopa u nasip šumske ceste. Kao pomoć pri postavljanju trase šumske ceste, karta plitkih klizišta je postavljena preko digitalnog modela reljefa te se na takav način trasa ceste mogla postaviti tako da se izbjegnu mjesta na kojima bi došlo da značajnog oštećenja i propadanja šumske ceste. Nakon postavljene trase šumske ceste napravljen je izračun troškova zemljanih radova, uzimajući u obzir opasnost od klizišta. Pretpostavljeno je na temelju predviđanja intenziteta padalina u sto godišnjem razdoblju, podijeljenom na 10 perioda, da bi troškovi obnove ceste bili jednaki trošku izgradnje ili čak do deset puta veći prilikom prelaska ceste preko klizišta. Kako bi se istražio utjecaj vremena upotrebe šumske ceste na projektiranje ceste, za analizu je odabrano nekoliko ograničenih vremena upotrebe, iako je za ovu cestu predviđeno neograničeno vrijeme upotrebe. Kao rezultat, utjecaja vremena upotrebe i kvaliteta projektiranja nisu značajni za ovo područje istraživanja. U budućim istraživanjima bi se trebalo razjasniti kroz koliko godina predviđeni intenzitet padalina utječe na troškove obnove različitih vrsta cesta u različitim područjima. Međutim ova metoda se može koristit za projektiranje šumskih cesta s obzirom na opasnost od klizišta, na osnovu vremena upotrebe i nosi-

14

Croat. j. for. eng. 34(2013)1

Study of Automatic Forest Road Design Model Considering Shallow Landslides with LiDAR ... (1–15)

M. Saito et al.

vosti šumske ceste. Tako, korišteni program ima mogućnost da prilikom projektiranja šumske ceste smanji troškove zemljanih radova izbjegavajući područja na kojima postoji mogućnost pojave klizišta. Učinkovito korištenje programa omogućuje automatsko projektiranje šumskih cesta na okolišno prihvatljiv način s minimalnim zahvatima u okoliš gdje se cesta gradi.

Ključne riječi: LiDAR, automatsko projektiranje šumskih cesta, klizišta, interpolacija između profila, dinamičko programiranje

Authors’ address – Adresa autorâ: Assist. Prof. Masashi Saito, PhD. * e-mail: m_saito@shinshu-u.ac.jp Shinshu University Faculty of Agriculture Department of Forest Science 8304, Minamiminowa-Village Kamiina-County 399-4598 JAPAN

Received (Primljeno): November 17, 2011 Accepted (Prihvaćeno): September 11, 2012 Croat. j. for. eng. 34(2013)1

Msahiro Goshima, MSc. Assoc Prof. Kazuhiro Aruga, PhD. e-mail: aruga@cc.utsunomiya-u.ac.jp Assoc Prof. Keigo Matsue, PhD. e-mail: matsue@cc.utsunomiya-u.ac.jp Prof. Yasuhiro Shuin, PhD. e-mail: shuin@cc.utsunomiya-u.ac.jp Prof. Toshiaki Tasaka, PhD. e-mail: tasaka@cc.utsunomiya-u.ac.jp Utsunomiya University Faculty of Agriculture Department of Forest Science 350 Mine Utsunomiya 321-8505 JAPAN * Corresponding author – Glavni autor

15

Original scietific paper – Izvorni znanstveni rad

Designing a Forest Road Network Using Mixed Integer Programming Akbar Najafi, Evelyn W. Richards Abstract – Nacrtak Forest roads are an essential yet costly part of forest management, and optimization methods are important tools for planning road systems to support harvesting. This paper presents a Mixed Integer Programming (MIP) optimization model to design a forest access system consisting of logging roads for trucking and access spurs for skidding. The network designed is hierarchical in the sense that the two transport systems require significantly different road standards, and timber may only be transferred from access spurs to forest roads. All timber must be transported from harvest sites to exit nodes that connect the forest road network to public roads. A dense network of potential connections is formed by overlaying a regular grid onto the forest, and then calculating costs of inter-node connections using GIS topographical data. Feasible arcs thus determined are input to the optimization model. The model minimizes total cost of road construction and maintenance, skidding and whole transportation in forest. It can be used to develop road system alternatives to support the process of planning the total access system. The model performance is explored on a study area in a mountainous region, where a persistent access network for partial harvesting is required. High quality solutions were achieved in reasonable computational time.

Keywords: optimization, forest harvesting cost, forest road network designing, access spur, harvest access planning

1. Introduction – Uvod Harvesting is one of the most important forest activities, for which purpose thousands of kilometers of roads and access spurs are constructed and millions of dollars are spent annually. Access roads may be used for other forestry, development and construction purposes, such as fire protection, silviculture, or recreational use. Nonetheless, their main economic purpose is for the extraction of timber. These roads have a large construction cost as well as maintenance costs (Najafi et al. 2008). It is thus important to develop methods to design road systems that carry out this purpose at minimal cost and that also minimize negative impacts such as erosion and water sedimentation.

Network design models can help forest management planners choose appropriate placement, design standard, and construction period for road segments.

Such models are characterized by construction and transport decisions chosen from amongst a finite number of feasible road projects and a finite number of Croat. j. for. eng. 34(2013)1

landing points for timber. Road costs, capacities, requirements for timber extraction, linkages between the forest road network and public roads that link to customers, and defined potential transportation systems grouped by type and their road standard requirements are the inputs used to form problem constraints.

Depending on the situation, the harvesting sites may be determined a priori, or the choice of where to harvest may be included in the optimization model along with the road network design decisions.

To create a forest road network design model, one must first systematically discretize the landscape to design a finite number of construction decisions, or potential road segments. These functions are typically accomplished with a Geographical Information System (GIS) by overlaying a grid of road intersection points, harvest location points, and extraction points on the landscape.

From these, a network of potential arcs for road placement is created. Second, one must acquire information about the forest topography, such as slope, soil type and ground bearing strength, to determine feasibility of

17

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

placing roads on the network arcs, and to estimate costs.

Furthermore, volume of cut at each harvest site and skidding cost on each link must be estimated. Again, this information is usually supported by a GIS system.

To explore feasible decisions thoroughly, the point grid and consequently the network of potential arcs should be dense so as not to preclude good opportunities. An optimization model will then be formulated to minimize total network costs, with constraints to ensure network connectivity, access to all harvesting sites, and optimal routes to exit points.

This problem is related to the integrated harvest and road construction problem, an important tactical planning problem that arises when cutting and road construction schedules are integrated over a medium-term planning horizon. If the road network has already been designed, the problem is to optimally schedule harvest and construction activities to create a plan that is costefficient and meets timber flow requirements. On the other hand, if a harvest schedule is first fixed, the problem is to design an optimal road network system to provide access to harvesting sites and routes to extract timber to main highways. In the case of single entry clearcutting, it is better to include both harvesting and road construction in an integrated model, since the capital cost of road construction is a significant factor in these problems. For example, Jones et al. (1986) showed that integrated models, which included both harvesting and road construction decisions in an integrated model, produced plans with costs reduced by as much as 30% over those that included only harvesting.

Solutions to the harvest-scheduling problem and road construction problem have been found using two strategies: exact mathematical optimization using Mixed Integer Programming (MIP) and heuristic solution methods. One of the earliest harvesting, road construction and transportation models is the Integrated Resources Planning Model (IRPM) presented by Kirby et al. (1986). This is an MIP optimization model to simultaneously select road construction and harvest projects, while minimizing costs of road construction and transportation of timber products. Guignard and Yan (1993) used Lagrangian substitution schemes to solve the IRPM. Subsequently, Guignard et al. (1998) improved the model by adding logical inequalities, lifting of inequalities and careful selection of Branchand-Bound branching priorities based on »doublecontracting« variables.

The appeal of exact methods such as integer programming is that their solutions are optimal, but the disadvantage is that problem size is limiting. In its original form, and using the 1980’s computer hardware and branch-and-bound technology, the MIP ap-

18

proach was considered to be suitable for only small problem instances. Jones et al. (1986) and Kirby et al.

(1986) tested the IRPM integer programming models and concluded that they were capable of solving only modest-sized problems. Even with the improvements of Guignard et al. (1998), time to solve realistic problem instances was prohibitive. Accordingly, many researchers have applied heuristic methods, which seek to approximate an optimal solution at a reasonable computational cost, to these problems. For example, Weintraub et al. (1995) created an iterative procedure with linear programming and heuristics to solve harvest and road construction problems. They relaxed the integrality constraints on road construction variables, and applied rounding heuristic rules to fractional variables in the solution. Clark et al. (2000) developed a three-stage heuristic to solve the harvest and road construction problem. Others have used various metaheuristics such as genetic algorithms, evolutionary programs, tabu search, and simulated annealing to address similar forest management problems. Lu and Eriksson (2000) used a genetic algorithm to form harvest units, and Falcao and Borges (2001) used evolution programs to generate harvest schedules. Richards and Gunn (2003) used tabu search to solve a forest management problem that includes spatial constraints for maximum opening size, adjacency delay (green up), and road construction decisions. They used parameterized weighting of objective function penalty components to develop a trade-off frontier between lost forest productivity due to timing of harvests, and capital cost of road construction.

Some methods to design forest road networks using GIS and heuristic solution methods have been published. Anderson and Nelson (2004) developed a computer algorithm to generate road networks under a variety of assumptions related to road design standards. Their method mimicked the processes an engineer might use to design a road network to access a number of stands, and does not optimize harvest timing or construction decisions. In contrast, the NETWORK 2001 method of Chung and Sessions (2001) uses metaheuristics (genetic algorithms and simulated annealing) to optimize road networks that access a given set of stands and uses digital terrain data to estimate costs of constructing various road links. They also, as in this work, deal with two road standards, with spur (skidding) roads for each stand. Their method designs the main road routes (trucking standard roads) using a genetic algorithm (GA) first, then adds spur roads as branches using simulated annealing (SA). Heinimann et al. (2003) and Newnham (1995) are two other examples of methods to create a road netCroat. j. for. eng. 34(2013)1

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

work using digital terrain data and a dense network of potential arcs, using a heuristic method.

Despite the documented difficulty of solving these and other spatial forest planning models using integer programming, there has recently been a recurrence of interest in MIP. Crowe et al. (2003), Goycoolea et al.

(2005), and Gunn and Richards (2005) have re-visited the adjacency problem in forest harvesting using integer programming models with new formulations. In addition to developing improved formulations for these models, these researchers have found that commercially available solvers have been much improved, and that problems which were previously intractable are now solvable using modern solvers with sophisticated branch and cut methodology. Their work has motivated us to develop an MIP model to design a transportation network of road and skidding trails for a forest to be harvested using multiple entry selection.

This paper presents a rudimentary Mixed Integer Programming (MIP) model that is able to design an access system consisting of roads and access spurs for skidding, and which minimizes the total cost of road construction, transportation costs, and skidding costs.

2. Methods – Metode The forest management problem we address is to design a least cost access network for forests to sup-

A. Najafi and Evelyn W. Richards

port continuous selective harvesting system. Cutting sites are pre-determined, and they are selection harvested on a fixed return cycle (10 years). Timber from each harvesting site is extracted using skidding systems to the banks of a spur road, and then transported to associated highway road systems using trucks.

Since selective harvesting in a number of sites will persist over time, the access network must be entirely constructed during first 10 years and remain in place.

Skidding trails are constructed using cut and fill without paving. Their maximum allowable gradient is 25% and trail width is 4 meters. Roads are constructed to a much higher standard, where cut and fill beds are appropriately drained and paved with three compacted layers. Their maximum allowable gradient is 10%, and road width is 5.5 meters. The higher quality roads have a significant construction cost but lower unit transportation cost when compared to skidding trails.

The two transportation systems are hierarchical. By this, we mean that skidding trails are not suitable for trucking, and that timber is never shifted from the higher level trucking system to the lower level skidding system. The input network of potential skidding trails and road segments is a very dense system of directional arcs. We also note that performance and applicability of this network design model may not be generalized to situations where access requirements are time dependent. For example, when single-entry

Fig. 1 Systematic nodes placed on the forest map Slika 1. Prostorni razmještaj mreže mjernih točaka Croat. j. for. eng. 34(2013)1

19

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

clearcut harvesting is used, the road network need not be permanent, and it need not be constructed at once.

The network design problem then becomes a multiperiod decision situation, which requires much larger models and is likely to be less »MIP-friendly« than the single period case.

The optimization problem is formulated, using an MIP model, to choose a least cost transportation network that supports all harvesting activities and the two transportation systems. Its objective is to minimize total cost of road construction, maintenance costs, transportation costs, and skidding costs.

2.1 Mathematical Model Definition – Definicija matematičkoga modela To define the network design model, 298 nodes are placed in the center of each real harvesting area and located on the forest map (Fig. 1).

Each node has a known volume of timber cut that is to be extracted in each period. These real world data were already collected by forest inventory group. In this model, each node is defined with two arguments, i and j, indicating the horizontal axis, or columns, and the vertical axis, or rows, respectively (Fig. 2). Additional nodes that are potential road intersections are then added. The distribution of the nodes over the forest can be done in a systematic or a random way. The

examples presented in this paper are based on a systematic placement of nodes on the forest map. Then, potential road segments are defined. For each node, a fixed number of arcs to near nodes are mapped, and their construction costs calculated using digital terrain data. Thus, all potential transportation decisions are based on directed arcs that connect nodes. Subject to terrain restrictions, each arc may be used as a skidding trail, or constructed to a road standard, or omitted from the solution altogether. In this work, there are potentially 20 adjacent nodes to each node, and hence as many as 20 outgoing and 20 incoming arcs at each node, creating a high density design network (Fig. 2).

Finally, exit nodes are chosen, or added if necessary, to identify the connection(s) between the forest road system and the public highway systems.

The objective is to minimize the sum of road network, timber transportation, and skidding costs. Road network costs consist of road construction costs based on the segment length, terrain conditions, and road standard, as well as road maintenance costs which are calculated based on traffic volume over each segment, its length, and its characteristics. Transport costs along each road segment are based on its unit cost of transportation and volume transported. Unit costs vary according to the characteristics of individual road segments. Similarly, cost of skidding timber on access

Fig. 2 Outgoing links at a systematic point Slika 2. Veze među mjernim točkama

20

Croat. j. for. eng. 34(2013)1

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

spurs is based on cost and hourly production of skidders, volume skidded and unit cost on each segment.

The objective function is to minimize total costs: min

∑

roads (i,j)

BC ij × Bij +

∑

arcs (i,j)

( TC ij × TVij + SC ij × SVij + MC ij × TVij )

(1)

Where: roads(i,j) means the set of all undirected segments connecting node i and node j, arcs(i,j) means the set of all directed segments connecting node i to node j.

Coefficients and constants: V Total volume of timber cut in the forest, C Capacity of a skidding spur, Dn Volume to be left at demand exit node n (equals zero if n is not an exit node), Amount of timber generated in node n, Voln Unit cost of skidding on arc (i,j), SCij Unit cost of trucking on road (i,j), TCij Capital cost to construct road (i,j), BCij Unit cost of maintenance for road (i,j).

MCij Binary Decision Variables: Bij Equals 1 if the segment connecting nodes i and j is built as a road, 0 otherwise, Equals 1 if timber is skidded on an access Sij spur between node i and node j, 0 otherwise.

Continuous decision variables: SVij Volume of cut carried on arc ij through access spur k (skidded volume), Volume of cut carried by trucking on arc TVij (i,j).

Network Flow Variables: SVIn Volume skidded into node n from adjacent nodes, Volume skidded out of node n to adjacent SVOn nodes, Volume trucked into node n from adjaTVIn cent nodes, Volume trucked out of node n to adjacent TVOn nodes, SVLeftn Timber volume skidded into node n that is left to be trucked on roads exiting node n.

Constraints:

Rij + Sij ≤ 1, ∀ undirected arcs ( i , j )

(2)

Constraint 2 means that a link can be unused, a road, or an access spur. It cannot be used for both skidding and trucking.

Croat. j. for. eng. 34(2013)1

A. Najafi and Evelyn W. Richards

TVij + TVij ≤ V × Bij , ∀ undirected arcs ( i , j )

(3)

Constraint 3 links the binary variable Bij to TVij and TVji, so that if any trucking is carried out on road (i,j), then Bij must be equal one.

SVij ≤ C × Sij , ∀ i , j

(4)

Constraint 4 links the binary variable Sij to SVij so that if SVij is > 0, then Sij must be equal to 1.

SVI n = SVOn =

TVI n =

TVOn =

∑

×SV jn

(5)

∑

× SVnj

(6)

∑

× TV jn

(7)

∑

× TVnj

(8)

(j,n) in arcs

(n,j) in arcs

(j,n) in arcs

(n,j) in arcs

Constraints 5, 6, 7, and 8 simply calculate volumes skidded and trucked in and out of each node.

SVIn + TVIn + Voln = SVOn + TVOn + Dn

(9)

Constraint 9 is the standard flow conservation equation. Volume entering a node plus timber volume generated at a node must equal the volume leaving plus any demand volume left.

Dn = 0 ∀ interior nodes n

(10)

Dn must be equal to zero if n is an interior node in the network, but may be positive if n is an exit node.

SVIn + Voln = SVOn + SVLeftn ∀ interior nodes n

(11)

TVIn + SVLeftn = TVOn ∀ interior nodes n

(12)

Constraints 11 and 12 maintain the integrity of timber product transportation by the different systems.

Timber can be skidded from one arc to the next, but timber cannot be carried on a road link, off-loaded, and then skidded along the next link.

∑D {e}

e

= V , where {e} is the set of all exit nodes (13)

Equation 13 ensures that all timber from harvesting sites is transported to exit nodes.

Equations 1–13 define the network flow model with hierarchical constraints. This formulation was improved by adding trigger constraints. These constraints are redundant in the sense that they do not change the optimal integer solution to the model. They

21

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

provide improved bounds and performance in the branching used by the mixed integer solver (Guignard et al. 1998).

First, constraints are added to trigger construction of an existing road segment if there is volume skidded to the node and left for trucking.

SVLeftn ≤

∑V × B

ni

(n,i)

,( n , i ) means all undirected

arcs that are incident on node n.

(14)

Second, if there is volume to be harvested and transported at a node, then at least one skid spur or road that is incident to it must be built.

∑( B (n,i)

ni

+ Sni ) ≥ 1 , ∀ n with Vn > 0

(15)

Third, no isolated road segments should be built.

Bij ≤

∑B (j,k)

jk

+ ∑ Bin , when i and j are interior (i,n)

nodes

(16)

2.2 Study area and scenarios – Područje istraživanja i scenariji To examine the model performance, it was used in designing a forest road network in a mountainous area in Iran. The area is located between 53° 16 ′ 57″ E to 53° 21′ 35″ and 36° 12′ 11″ N to 36° 15′ 40″ N (Fig. 3).

The 1 428 hectare mountainous hardwood forest is managed under sustainable forest management. The harvesting method is single tree selection with a continuous return cycle of 10 years. Forest roads and their related skid trails are permanent. The potential network design was constructed by requiring one road link for every 2.27 hectares and one access spur for every 1.14 hectares.

Fig. 4 and 5 show an exploded view of a portion of the region with its feasible road links and feasible access spurs. In all, the model contains 252 harvesting sites, 883 potential spurs, and 440 road segments. One exit node is located at the top centre of the forest. The resulting MIP model has 69 420 constraints, 33 085 variables, and 300 083 non-zero coefficients. To check the validity and performance of the model, we used the same spa-

Fig. 3 Study area Slika 3. Područje istraživanja

22

Croat. j. for. eng. 34(2013)1

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

Fig. 4 Typical road feasibility; due to slope, swampy areas, or no volume to be extracted, there are typically less than 20 links for each node Slika 4. Izvedivost šumske ceste. Zbog nagiba terena, močvarnih područja ili gospodarske neisplativosti svaka mjerna točka obično ima manje od 20 veza s ostalim mjernim točkama

A. Najafi and Evelyn W. Richards

Fig. 5 Access spur feasibility; there may not be 20 links for each node because of steep terrain, swamp areas, or no volume at the systematically placed harvest node Slika 5. Izvedivost traktorskoga puta. Moguće da će zbog nagiba terena, močvarnih područja ili gospodarske neisplativosti biti manje od 20 veza među mjernim točakama

Table 1 Solution Cost Composition, Solution Times and Quality. Rfact is a factor applied to the base road construction cost Tablica 1. Rješenje troškovne sastavnice, utroška vremena i kakvoće. Rfact je faktor koji se primjenjuje pri prikazivanju osnovnih troškova izgradnje Total Cost Rfact

Ukupni troškovi

Trucking Activity Prijevoz

Skidding Cost

Maintenance Cost

Trucking Cost

Road Construction Cost

CPU Time

Final MIP Gap

Troškovi privlačenja

Troškovi održavanja

Troškovi prijevoza

Troškovi izgradnje šumskih prometnica

Računalno vrijeme

Završna MIP praznina

$

$

1000 m3/km

$

%

$

%

1.3

539 392

160 298

165 510

31%

62 516

12%

1.15

496 013

158 586

162 542

33%

54 712

1

452 264

161 226

156 096

35%

0.85

407 011

165 526

148 170

0.7

362 138

166 491

0.55

313 038

0.4

$

%

s

%

208 388 39%

102 978

19%

< 100

<1

11%

182 374 37%

96 386

19%

200

<1

48 368

11%

161 226 36%

86 574

19%

3600

1.67

36%

42 209

10%

140 697 35%

75 934

19%

3600

3.87

143 114

40%

34 963

10%

116 544 32%

67 517

19%

3600

6.52

173 609

131 071

42%

28 645

9%

95 485

31%

57 837

18%

3600

10.06

259 411

193 310

102 092

39%

23 197

9%

77 324

30%

56 798

22%

3600

15.99

0.3

210 244

219 220

67 080

32%

19 730

9%

65 766

31%

57 668

27%

3600

12.63

0.2

162 371

235 438

57 118

35%

14 126

9%

47 088

29%

44 039

27%

3600

15.83

0.1

100 914

251 818

34 279

34%

7 555

7%

25 182

25%

33 899

34%

3600

12.68

Croat. j. for. eng. 34(2013)1

%

23

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

tial configuration and skidding costs, but varied road construction costs by factors 1.3, 1.15, …, to 0.40.

The values, assigned to skidding and road construction costs in the original scenarios, were 500$ and 18 400$, respectively. Then, to further test computational performance, we generated a number of random instances of the problem by using Monte Carlo sampling to assign skidding and road construction costs. Probability distributions were fit to the original skidding and cost data, which were then sampled to create the random instances. The skidding cost distribution fit best to a three-parameter Beta, with minimum 4.00, size parameter 0.839, and shape parameter 2.81. Road construction costs were sampled from the uniform distribution on [2700, 14400]. In these random instances, all potential road segments and skid trails were assumed to be feasible. The resultant models have 9 520 skidding, 9 520 trucking, and 5 340 road construction feasible decision variables. These constitute a much denser network of potential arcs for both transportation systems.

Finally, seven of the original scenarios were modified to add a second exit point at the bottom of the landscape.

The MIP models were created using OPL Studio 4.2 and solved using CPLEX version 10.0, both products of Ilog, Inc. They were solved using a Pentium 4 3.4 GHz CPU computer with 4 GB RAM, running Windows XP Professional. All models were run until the MIP gap was less than or equal to one percent, or one hour of CPU time was consumed, whichever occurred first.

3. Computational results – Računalni rezultati Table 1 shows solution values, their cost composition, solution times and final MIP gap for each of ten base scenarios. The scenarios were achieved by varying the costs of road construction by a multiple called Rfact. Optimality gaps ranged from less than one percent to 16 percent, and solution times ranged from less than 100 seconds to the cutoff point of 3 600 seconds. Solution quality, as measured by the final MIP gap, worsened as the road construction and trucking costs were reduced to 2 times the first case, then when Rfact was further reduced to 0.1. As expected, increasing road costs causes a corresponding decrease in construction and use of hauling roads, with more timber gathered through access spurs. Total trucking activity, as measured in volume-distance units, ranged from 160 298 to 251 818 m3/km. Fig. 6, 7, and 8 show road network solutions achieved for three of the 10

24

Table 2 MIP Optimality Gap (%) for 10 randomly generated instances, with varied road construction costs. Model execution was terminated in one hour Tablica 2. Optimalna MIP praznina (%) za deset nasumično odabranih slučajeva s različitim troškovima gradnje šumskih cesta. Vrijeme računanja modela iznosilo je sat vremena Road Cost Change Factor Različiti troškovi gradnje šumskih cesta Case – Slučaj

0.4

0.55

0.7

0.85

1

1.15 1.3

1

16.3

10.1

3.76 < 1 < 1 < 1 < 1

2

13.45

8.89

2.53 < 1 < 1 < 1 < 1

3

15.05 10.25 5.12 < 1 < 1 < 1 < 1

4

16.4

10.93 4.55 < 1 < 1 < 1 < 1

5

15.33

9.21

4.11 < 1 < 1 < 1 < 1

6

15.9

8.6

2.73 < 1 < 1 < 1 < 1

7

17.69 15.09 7.63 < 1 < 1 < 1 < 1

8

16.06

10.7 10.63 5.2 < 1 < 1 < 1

9

14.88

9.27

2.17 < 1 < 1 < 1 < 1

10

17.43

10.8

3.57 < 1 < 1 < 1 < 1

Mean Aritmetička sredina Minimum

15.849 10.384 4.68 1.42 < 1 < 1 < 1 13.45

Minimum Maximum

8.6

2.17 < 1 < 1 < 1 < 1

17.69 15.09 10.63 5.2 < 1 < 1 < 1

Maksimum

Table 3 CPU time (seconds), 10 random cases, optimality tolerance 1% Tablica 3. Računalno vrijeme (sekunde), 10 nasumično odabranih slučajeva, optimalna tolerancija 1 % Road Cost Change Factor Različiti troškovi gradnje šumskih cesta Case – Slučaj

0.4

0.55

0.7

0.85

1

1.15

1.3

1

3 600 3 600 3 600

400

<1

<1

<1

2

3 600 3 600 3 600

274

<1

<1

<1

3

3 600 3 600 3 600

<1

<1

<1

<1

4

3 600 3 600 3 600

<1

<1

<1

<1

5

3 600 3 600

3600

400

<1

<1

<1

6

3 600 3 600 3 600

<1

<1

<1

<1

7

3 600 3 600 3 600

300

<1

<1

<1

8

3 600 3 600 3 600

3 600

<1

<1

<1

9

3 600 3 600 3 600

<1

<1

<1

<1

10

3 600 3 600 3 600

<1

<1

<1

<1

Croat. j. for. eng. 34(2013)1

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

A. Najafi and Evelyn W. Richards

Fig. 6 Solution showing the road and related access spurs (Rfact = 1.3) Slika 6. Planirana šumska cesta s pripadajućim traktorskim putovima (Rfact = 1,3)

Fig. 7 Solution showing the road and related access spurs (Rfact = 1.0) Slika 7. Planirana šumska cesta s pripadajućim traktorskim putovima (Rfact = 1,0) scenarios. Dark lines indicate road segments, and lighter lines indicate the skidding paths. One can see that the solutions are feasible with respect to continuity of transport system and connection to the exit point.

Table 2 shows MIP gaps for the ten randomly generated cases. Again, as the relative cost of roads decreased, solution quality degraded. Gaps ranged from less than one percent to 17.69 percent. Table 3 shows CPU times for these 10 random cases with Rfact values ranging from 1.3 to 0.4. CPU times ranged from less than one minute to the cutoff point of one hour.

Croat. j. for. eng. 34(2013)1

Table 4 shows solution cost structure and computational performance for seven scenarios with two exit points. Solution times ranged from 44 seconds to one hour, and MIP gaps of less than one percent to 16.54 percent. Total costs were reduced significantly from the original scenarios where only one exit point was allowed. Fig. 9 shows the solution for Rfact equal to 1.0, and Fig. 11 the solution for Rfact equal to 0.4. These solutions are markedly different from the cases of one exit point, with roads clustered at each of the two exit points and considerable reduction in skidding and truck transport costs.

25

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

Fig. 8 Solution showing the road and related access spurs (Rfact = 0.70) Slika 8. Planirana šumska cesta s pripadajućim traktorskim putovima (Rfact = 0,70)

Fig. 9 Solution showing the road and related access spurs (Rfact = 1.00, 2 Exit Points) Slika 9. Planirana šumska cesta s pripadajućim traktorskim putovima (Rfact = 1,00 s dvije izlazne točke)

4. Discussion – Rasprava The model produces valid results for the hierarchical transportation model, with integrity of transport system and road system maintained. Moreover, the results are of very reasonable quality when compared to known upper bounds on optimality.

We have used the MIP gap to assess the quality of the solutions achieved. This is the percentage gap between the best integer solution achieved and the best bound on the optimal solution found by the solver.

26

This does not mean that a solution with MIP gap of 15% is fifteen percent worse than the optimal solution.

Rather, the measure guarantees that it is at least as good as 85% of the optimal.

In all cases, the MIP model was able to produce good solutions to the problem in very reasonable computational time. The final MIP gap was less than 20 percent for all cases in up to one hour of computation time, with many cases solved to less than 1 percent optimality in less than one minute.

Croat. j. for. eng. 34(2013)1

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

A. Najafi and Evelyn W. Richards

Table 4 Road network solutions with 2 exit points; solution cost composition, times and quality; Rfact is a factor applied to the base road construction cost Tablica 4. Planiranje mreže šumskih prometnica s dvije izlazne točke. Rješenje troškovne sastavnice, utroška vremena i kakvoće. Rfact je faktor koji se primjenjuje pri prikazivanju osnovnih troškova izgradnje Total Cost Rfact

Ukupni troškovi

Trucking Activity Prijevoz

Skidding Cost

Maintenance Cost

Trucking Cost

Troškovi privlačenja Troškovi održavanja Troškovi prijevoza

Road Construction Cost CPU Time Troškovi izgradnje šumskih prometnica

Final MIP Gap

Računalno vrijeme

Završna MIP praznina

$

1000 m3/km

$

%

$

%

$

%

$

%

Sec.

%

1.30

321 285

48 590

169 917

52.89

18 950

5.90

63 167

19.66

69 251

21.55

44

1.00

1.15

303 761

50 061

167 659

55.19

17 271

5.69

57 570

18.95

61 260

20.17

121

1.00

1.00

285 625

52 944

158 927

55.64

15 883

5.56

52 944

18.54

57 870

20.26

375

0.91

0.85

266 467

54 037

157 567

59.13

13 780

5.17

45 932

17.24

49 190

18.46

3 367

1.00

0.70

246 787

61 780

141 687

57.41

12 974

5.26

43 246

17.52

48 881

19.81

3 600

7.50

0.55

221 520

75 287

121 593

54.89

12 422

5.61

41 408

18.69

46 098

20.81

3 600

11.07

0.40

188 503

98 322

86 407

45.84

11 799

6.26

39 329

20.86

50 969

27.04

3 600

16.54

The quality of solutions attainable in one hour CPU time varied with the cost structure of the problem instance. That is, when road costs were reduced relative to skidding costs, more CPU time was consumed to achieve solutions of the same quality as for instances with a higher road to skidding cost ratio. This behavior was observed in the original scenarios and in the randomly generated cases.

One of our research objectives was to determine whether it was possible to solve the MIP model in reasonable time for problems that had many feasible rout-

ing choices. We allowed 20 incoming and outgoing arcs at each node of the network. Our first examples were constructed assuming a mountainous terrain, with many node-to-node connections infeasible due to slopes. However, for the ten randomly generated cases, all of these arcs were feasible. These models were solved to approximately the same quality of solution in much the same computational time as the original ten scenarios. This indicates that density of the network is not a limiting factor in attaining a solution. Also, although we have not reported results using other formulations,

Fig. 10 Solution showing the road and related access spurs (Rfact = 0.40, 2 Exit Points) Slika 10. Planirana šumska cesta s pripadajućim traktorskim putovima (Rfact = 0,40 s dvije izlazne točke) Croat. j. for. eng. 34(2013)1

27

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

we found that adding trigger constraints was important in improving MIP performance.

We can see from the network solutions (Fig. 8–10) that the solutions make sense spatially. However, one might question the potential for this model to produce long skidding routes that are feasible in practice. There are no direct constraints on skidding distances in the solution. Instead, arc capacity constraints were used to reduce the likelihood that timber would be skidded too far, by constraining the accumulated flow through each skidding trail. It should be noted that the capacity method works only approximately since timber volumes vary between stands and skidding volumes may accumulate to the imposed capacity over distances which are not acceptable. This implies that it might be necessary to subject the network to some post–optimization fixing if excessive skidding distances occur.

5. Conclusion – Zaključci This model is designed to create access networks for areas to be continuously harvested by selection harvesting, and hence require a permanent road network. In such forests there are generally two choices of road standard, the skidding and the hauling road standards. The cost considerations are that the higher standard roads have higher construction costs but lower transportation costs. Moreover, once timber is skidded to a road location, it must be then transported over hauling roads until it reaches an exit point of the network. To minimize total cost, costs in both networks should be considered simultaneously. This model solves this problem in reasonable computational time and produces realistic solutions of reasonable quality. Solver performance was robust with respect to density of the input network, but sensitive to the cost structure. Solving models with two exit points was not appreciably more difficult than solving those with one exit, but the solutions attained were radically different in structure and in cost.

The examples and the discussion have focused on skidding and hauling transportation systems. This model could also be used to design access systems where any two different transportation systems exist, to create a primary network from which the goods are transported to the secondary network. We did not test systems with more than two transportation systems standards, but the model is easily extended to accommodate more transportation systems and road standards.

The CPLEX, commercial MIP solver was able to solve these problems easily, attaining optimal solutions in less than one minute in most cases, and attain-

28

ing solutions with optimality bound less than 20% in all cases. Forecasting performance for branch and bound methods using a few instances is risky, since results can vary widely. We found that the cost structure had a significant impact on performance, while the network density did not.

In conclusion, we have presented a model and study to show a mixed integer programming method to design an access network for hierarchically constrained transport systems. The system can minimize costs of construction, transportation, and maintenance for a dense network of potential arcs, and can solve the problem in reasonable time. These results are quite positive, and it is important for decision makers to realize that models of this size, which would have been considered intractable a short time ago, can be solved in a reasonable time, and can provide useful decision support.

6. References – Literatura Anderson, E. A., Nelson, J., 2004: Projecting vector-based road networks with a shortest path algorithm. Can J For. Res. 34(7): 1444–1457.

Chung, W., Session, J., 2001: NETWORK 2001 – Transportation Planning Under Multiple Objectives; The International Mountain Logging and 11th Pacific Northwest Skyline Symposium.

Clark, M. M., Meller, R. D., McDonald, T. M., 2000: A ThreeStage Heuristic for Harvest Scheduling with Access Road Network Development. For. Sci. 46(2): 204–218.

Crowe, K., Nelson, J., Boyland, M., 2003: Solving the area-restricted harvest-scheduling model using the branch and bound algorithm. Can. J. For. Res. 33(9): 1804–1814.

Falcao, A.O., Borges, J. G., 2001: Designing an Evolution Program for Solving Integer Forest Management Scheduling Models: An Application in Portugal. For. Sci. 47(2): 158–168.

Goycoolea, M., Murray, A. T., Barahona, F., Epstien, R., Wintraub, A., 2005: Harvest Scheduling Subject to Maximum Area Restrictions: Exploring Exact Approaches. Ope. Res. 53(3): 490–500.

Guignard, M., Ryu, C., Spieldberg, K., 1998: Model tightening for integrated timber harvest and transportation planning.

Eur. J. Ope. Res. 111(3): 448–460.

Guignard, M., Yan, H., 1993: Structural decomposition methods for dynamic multiple machine, multiple hydropower plant optimization. Dec. Sci. Rep. 111: 448–460.

Gunn, E. A., Richards, E. W., 2005: Solving the adjacency problem with stand-centred constraints. Can. J. For. Res. 35(4): 832–842.

Heinimann, H. R., Stückelberger, J., Chung, W., 2003: Improving automatic grid cell based Road Route Location Procedures. High Tech Forest Operation For Mountainous Terrain.

Austria 5–9 October.

Croat. j. for. eng. 34(2013)1

Designing a Forest Road Network Using Mixed Integer Programming (17–30) Jones, J. G., Hyde, J. F. C., Meacham, M. L., 1986: Four analytical approaches for integrating land management and transportation planning on forest lands. Res. Pap. INT-36l, USDA Forest Service, Intermountain Res. Sta. 32 p.

Kirby, M., Hager, W., Wong, P., 1986: Simultaneous planning of wildland management and transportation alternatives.

TIMS Studies in the Management Sciences 21: 371–387.

Lu, F., Eriksson, L. O., 2000: Formation of harvest units with genetic algorithms. For. Eco. Man. 130(1–3): 57–67.

A. Najafi and Evelyn W. Richards

Newnham, R. M. 1995: Roadplan: A tool for designing forest road network. Int. J. for Eng. 6(2):17–26.

Richards, E. W., Gunn, E. A. 2003: Tabu Search Design for Difficult Forest Management Optimization Problems. Can. J. For. Res. 33(6): 1126–1133.

Weintraub, A., Jones, J. G., Meacham, M. L., Magendzo, A., Malchuk, D., 1995: Heuristic procedures for solving mixedinteger harvest scheduling-transportation planning models.

Can. J. For. Res. 25(10): 1618–1626.

Najafi, A., Sobhani, H., Saeed, A., Makhdom, M., Mohajer, M. M., 2008: Planning and Assessment of Alternative Forest Road and Skidding Networks. Cro. J. For. Eng. 29(1): 63–73.

Sažetak

Planiranje mreže primarnih šumskih prometnica primjenom modela mješovitoga cjelobrojnoga programiranja Troškovi su izgradnje i održavanja šumskih prometnica veliki te je zbog toga važno razviti metode projektiranja šumskih prometnica koje u obzir uzimaju minimalan utrošak novčanih sredstava.

Ovaj je članak optimizirani model mješovitoga cjelobrojnoga programiranja (MIP) koji služi za planiranje mreže šumskih prometnica uzimajući u obzir minimalan utrošak novčanih sredstava pri izgradnji i prijevozu drvnih sortimenata.

Problem pri optimizaciji, uz primjenu modela MIP, bio je odabrati najučinkovitiju mrežu šumskih prometnica koja omogućuje sve aktivnosti u pridobivanju drva. Cilj je spomenutoga modela smanjiti troškove gradnje i održavanja šumskih cesta te troškove primarnoga i sekundarnoga transporta drva.

Pri definiranju sustava za planiranje mreže šumskih prometnica odabrano je 298 smještenih točaka u prostoru.

Za svaki točku poznat je drvni volumen za sječu i privlačenje. Kako bi se točke spojile (u buduću šumsku prometnicu), za svaku je sniman broj lukova (poveznica) prema susjednim točkama, a troškovi izgradnje šumske ceste izračunati su pomoću podataka iz digitalnoga modela terena.

Dakle, planiranje je mreže šumskih prometnica temeljeno na smjeru pružanja lukova koji povezuju mjerne točke.

U skladu s terenskim ograničenjima svaki luk (buduća prometnica) može služiti kao primarna ili sekundarna šumska prometnica ili se potpuno izostavlja iz modela.

O ovom radu svaka mjerna točka ima 20 mogućih susjednih točaka, a samim time i 20 odlaznih i 20 dolaznih lukova stvarajući na taj način vrlo gustu mrežu šumskih prometnica. Cilj je minimizirati gustoću mreže šumskih prometnica te troškove primarnoga i sekundarnoga transporta drva.

Primjenjivost modela ispitivana je pri planiranju mreže šumskih prometnica u planinskom području u Iranu. Šumske su ceste i njihovi pripadajući traktorski putovi trajne građevine. U model su ugrađena ukupno 252 radilišta (mjesta sječe), s 883 potencijalna traktorska puta i 440 segmenata šumskih cesta. Izlazna je točka iz sastojine smještena na gornjem dijelu istraživanoga područja. U konačnici model MIP sadrži 69 420 terenskih ograničenja, 33 085 varijabli i 300 083 koeficijenta koji svi zajedno određuju mrežu šumskih prometnica. Pri provjeravanju valjanosti i učinkovitosti modela korišteni su isti prostorni podaci i troškovi transporta uz različite faktore za izračun troškova gradnje (u rasponu 1,3; 1,15,..., do 0,40). Dobiveno je sedam mogućih modela (scenarija) mreže šumskih promentica te je dodana i druga izlazna točka pri dnu istraživanoga područja. Kao što se i očekivalo, povećanjem novčanih sredstava utrošenih u izgradnju šumskih cesta smanjuju se novčana sredstva za izgradnju traktorskih putova te se zbog toga na njima povećava drvni volumen u primarnom transportu. Ukupni transport drvnoga volumena istraživanoga područja kretao se u rasponu od 160 298 do 251 818 m3/km. Ukupni su troškovi značajno smanjeni u usporedbi s prvotnim scenarijem u kojem je bila određena samo jedna izlazna mjerna točka (na gornjem dijelu sastojine). Sustav daje valjane rezultate za hijerarhijski transportni model s cjelovitim transportnim sustavom uz održavanja šumskih cesta. Rezultati su vrlo visoke kakvoće ako ih uspoređujemo s poznatim gornjim granicama optimalnosti. To je postCroat. j. for. eng. 34(2013)1

29

A. Najafi and Evelyn W. Richards

Designing a Forest Road Network Using Mixed Integer Programming (17–30)

otna razlika između najbolje dostignutoga cjelobrojnoga rješenja i najbolje veze s optimalnim rješenjem pronađene putem rješavača (tzv. solver). U svim slučajevima model MIP u mogućnosti je proizvesti dobro rješenje problema u vrlo razumnom računalnom vremenu. Jedan od ciljeva ovoga istraživanja bio je odrediti je li moguće pronaći rješenje problema vezanih uz planiranje trasa modelom MIP u razumnom računalnom vremenu. Stoga su prvi primjeri bili rađeni za planinski teren, s mnogo neostvarivih veza među mjernim točkama zbog ograničavajućega nagiba. Sustav može minimizirati troškove izgradnje, transporta i održavanja kod guste mreže prometnica s mnogo mogućih lukova povezivanja i riješiti problem u razumnom vremenu, što ga čini prikladnim sustavom za potporu odlučivanja u šumarstvu. Ovaj model može biti primijenjen pri planiranju pristupnih pravaca, gdje postoje dva različita transportna sustava, za planiranje sekundarne mreže s kojih se dobra (drvo) transportira na primarnu mrežu šumskih prometnica. Model nije testiran na više od dva različita transportna sustava, ali se vrlo lako može proširiti na druge transportne sustave i prometne standarde. Struktura je troškova važan čimbenik na učinkovitost modela, dok gustoća mreže šumske prometne infrastrukture nije.

Ključne riječi: optimizacija, troškovi pridobivanja drva, gustoća mreže šumskih prometnica, trakorski putovi, planiranje pristupa šumi

Authors’ address – Adresa autorâ: Assist. Prof. Akbar Najafi, PhD. * e-mail: a.najafi@modares.ac.ir Tarbiat Modares University Faculty of Natural Resources Department of Forestry Imam Ave, Noor, Mazandaran Province P.O. Box 64414–356 IRAN

Received (Primljeno): September 17, 2011 Accepted (Prihvaćeno): December 13, 2011

30

Assoc. Prof. Evelyn W Richards, PhD. e-mail: ewr@unb.ca University of New Brunswick Faculty of Forestry and Environmental Management Forestry/Geology, Rm. 201 UNB Fredericton Campus CANADA * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

Original scietific paper â&#x20AC;&#x201C; Izvorni znanstveni rad

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road Network: an Application in the Italian Alpine Region Marco Pellegrini, Stefano Grigolato, Raffaele Cavalli Abstract â&#x20AC;&#x201C; Nacrtak The combination of GIS tools and Analytic Hierarchy Process (AHP) techniques is used to develop a Decision Support System to rank the maintenance priorities of a forest road network according to the actual conditions and needs. The decision-making process is divided into 4 stages. The first stage fixes the objectives of the analysis as the minimization of the sediment production from road surface and the maximization of the social value of the road. The second stage defines the hierarchical structure of the decision problem. At this stage the set of factors (criteria) to maximize each objective and the evaluation methods are defined. At the third stage AHP analysis is applied using a specific application running on ArcGIS, to calculate the evaluation layer that represents the importance of each road according to the set objectives.

The values of the evaluation layer are used at the fourth stage to rank the maintenance interventions according to the given benefit. The methodology has been tested in a forest road network with an extension of 107.8 km including in the analysis the real budget constraints and maintenance costs.

The results show that the integrated use of GIS and AHP analysis represents a valuable tool to rate the importance of the forest road network for the management of a mountain territory and to define priorities among maintenance operations of the road network, in order to maximize the overall benefit with limited economic resources.

Keywords: forest road maintenance, maintenance cost, AHP, GIS, Decision Support System

1. Introduction â&#x20AC;&#x201C; Uvod Routine road maintenance is vital to keep a forest road system serviceable and to maintain the proper working of its drainage system. Many studies show how a well-maintained road can be protected from rapid deterioration, minimizing sediment production (Thompson et al. 2010) and improving the trafficability with a reduction of trucking costs (Feng and Douglas 1993; Talbot and Nitteberg 2011). Each year, a consistent amount of money is spent to upgrade and maintain forest road networks. In order to optimize the use of limited funds, it is of primary importance to set investment priorities while meeting management Croat. j. for. eng. 34(2013)1

and environmental goals. The resulting task is complex because of the many aspects involved in forest road management, including the natural environment and the socio-economic context in which the road network is located. For that reason, the management of forest road networks needs methods to integrate multiple objectives and set priorities across these different goals.

Existing studies generally focus on a single aspect in order to better understand the conditions of the road network and the processes related to this road.

The Washington Department of Natural Resources and Boise Corporation has created an empirically based model (SEDMODL) used to estimate road-relat-

31

M. Pellegrini et al.

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31â&#x20AC;&#x201C;42)

ed sediment production and transport to streams (Dubè et al. 2004). PotoÄ?nik et al. (2006) investigate a traffic management strategy in the preserved forest area of the Pokljuka highland (Slovenia) considering the touristic importance of the roads. Hruza and Vyskot (2010) evaluate a forest road network according to the social-recreation value in order to define the optimal path for touristic trails.

Spatial multi criteria decision techniques provide tools for aggregating the geographical data and the decision-makers preferences into unidimensional value or utility of alternative decisions, combining a set of criteria to achieve a single composite basis for a decision according to a specific objective (Malczewski 1999).

Although there is a variety of techniques for the determination of the weight of a given set of criteria, one of the most promising seems to be the pairwise comparisons developed by Saaty (1980) in the context of a decision-making process known as the Analytical Hierarchy Process (AHP). In the pairwise comparison method, the decision-maker is asked to give the rela-

tive importance to the criteria by comparing them two by two. Schmoldt et al. (2001) describes the basis of the application of the AHP in natural resource and environmental decision-making.

The use of AHP appears to have the potential to help managing the existing road systems where research has not yet uncovered quantifiable relationships between cause and effect, meaning that the synthesis of road inventory data to set investment priorities should depend in part on professional judgment. According to this idea, Coulter et al. (2006) developed a maintenance priority definition methodology that uses AHP analysis in order to minimize the environmental impact to soil and water resources from forest roads. Shiba (1995) used an AHP based approach to improve the development strategy of road network in mountainous areas of Japan considering complex socio-economic problems. AHP analysis and GIS tools have been recently applied to evaluate the needs of forest roads in a mountainous area of Italy (Cavalli et al. 2010) and to evaluate new forest road location alternatives (Abdi et al. 2009).

Fig. 1 General layout of the analysis Slika 1. Izgled analize

32

Croat. j. for. eng. 34(2013)1

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

M. Pellegrini et al.

Table 1 Considered criteria and relative evaluation methods Tablica 1. Kriteriji i relativne metode procjene Criteria – Kriteriji

Evaluation Methods – Metode procjene Objective 1 – Erosion risk – Cilj 1 – Rizik od erozije GIS analysis (Model based on contour line) – Analiza pomoću GIS-a (model temeljen na slojničkim kartama)

Road gradient (iGRADE) – Uzdužni nagib ceste Surface condition (iCOND) – Stanje vozne površine

Visual evaluation (field survey) – Vizualna procjena (terensko istraživanje)

Drainage system (iDRAIN) – Elementi sustava odvodnje

Visual evaluation (field survey) – Vizualna procjena (terensko istraživanje) GIS analysis (based on Network analysis) – Analiza pomoću GIS-a (model temeljen na slojničkim kartama)

Traffic (iTRAFFIC) – Prometno opterećenje Location of the road (iSLOPE) – Položaj šumske ceste

GIS analysis (based on Digital Terrain Model) – Analiza pomoću GIS-a (temeljena na digitalnom modelu terena)

Objective 2 – Social value – Cilj 2 – Socijalna vrijednost Touristic importance (iTOUR) – Turistička uloga Farming importance (iFARM) – Poljoprivredna uloga

GIS analysis (access to touristic sites) – Analiza pomoću GIS-a (pristupačnost turističkim znamenitostima) GIS analysis (access to agricultural/farming sites) Analiza pomoću GIS-a (pristupačnost poljoprivrednim gospodarstvima/farmama)

Productive importance (iPROD) – Gospodarska uloga Operative class (iOPER) – Operativni razred

GIS analysis – Analiza pomoću GIS-a Visual evaluation (field survey) – Vizualna procjena (terensko istraživanje)

The present study combines GIS tools and AHP techniques to rank the maintenance priorities in a forest road network. In particular, the proposed methodology integrates the evaluation of the risk of erosion from road surface and the evaluation of the multifunctionality of forest roads, resulting particularly suitable for areas where the management of the roads have to consider many different functions. The result of the analysis is applied to support the use of the economic resources for the maintenance of a forest road network in the Italian Alpine region.

2. Material and methods – Materijal i metode 2.1 Analysis structure – Struktura analize Any decision-making process begins with the recognition of decision problems. A spatial decision problem is the difference between the desired and the existing state of a real-world geographical system (Mal czew ski 1999). Simon (1960) suggests that any decision-making process can be structured in three major phases: intelligence (is there a problem or an opportunity for change?), design (what are the alternatives?), choice Croat. j. for. eng. 34(2013)1

(which alternative is better?). The procedure passes through the evaluation of the actual status of a forest road network and the actual maintenance strategies to search for the improving elements (Fig. 1). The production of sediment from the gravel road surfaces represents one of the most frequent and hazardous processes directly connected with the presence and maintenance of the forest road network. The minimization of this process is directly related to the correct execution of the maintenance interventions and has been considered as the first objective. The other leading elements in determining maintenance interventions are the type and level of use of the forest road.

Following a multi-functional approach, the maximization of the social values has been considered as second objective in the priorities definition.

The set of criteria and the hierarchical structure of the decision problem have been developed through the examination of the relevant literature regarding the analyzed aspect and through the opinion of experts and stakeholders.

2.2 Criteria evaluation – Kriteriji procjene Minimization of the risk of sediment production from forest road surface (Objective 1): sediment can be

33

M. Pellegrini et al.

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

eroded from all road features but road surface erosion is generally the dominant source of sediment (RamosScharron and MacDonald 2005). A recent review paper on surface erosion and sediment delivery model for unsealed road (Fu et al. 2010) effectively describes the main factors highlighted in literature. These factors include rainfall intensity and duration, snowfall, characteristic of road surface and used materials, road slo pe, traffic, construction standards and level of main te nan ce (MacDonald and Coe 2008). The gradient, or slope, of a road segment influences the erosion rate. Water flows down steeper road segments more quickly, resulting in a greater erosive power and in a higher shear stress (Bilby et al. 1989; Elliot and Tysdal 1999; Luce and Black 1999a). The width of the road and the extent of traffic on a road both influence the amount of erosion produced from the road surface. Researches by Reid and Dunne (1984); Grayson et al. (1993); Luce and Black (1999b); Ziegler et al. (2001a); Sheridan and Noske (2007) were specifically aimed at determining the effects of traffic on road erosion. All of these studies showed increasing erosion rates with increased traffic use. According to the literature, the set of criteria considered for Objective 1 are: road gradient, surface condition, presence of drainage structure, level of traffic and location of the road on the hill-slope (Table 1).

Maximization of the social value of the road network (Objective 2): to evaluate the social value, the different functions that each road can perform have been considered. After discussion with stakeholders of the studied area, we considered the following functions: touristic importance (access to important tourist sites), farming importance (access to farming activities), productive importance (access to productive forests) and operative class (actual constructive parameters of the road and ease of transit) (Table 1). Once the hierarchical structure of objectives and attributes has been established, each criterion can be represented as a raster map layer in the GIS database. Information about surface condition, presence of drainage structures, operative class, touristic and farming importance, were collected during the field surveys and reported as attributes to the layer representing the forest road network. To determine road gradient, hill-slope and level of traffic in a representative layer, three semiautomated analysis procedures have been developed using geo-processing tool for surface analysis and Network Analysis techniques. The analyses have been supported by ArcGIS 10 (ESRI 2011) and ModelBuilder interface. In all the cases, the inputs needed to run the models are the digital terrain model (DTM) and forest type regional map.

Multi-criteria decision analysis requires that the values contained in the various criterion map layers

34

be transformable to comparable units. For this reason all of the data have been converted into standardized values using the maximum score procedures.

X ' ij =

X ij X max

(1)

Where: X’ij the standardized score for the ith object (alternative), th j attribute and xij is the raw score and xmax is the maximum score for the jth attribute.

The scale of standardized scores range from 0 to 1000, where higher overall score values indicate greater benefit. A pairwise comparison matrix was completed according to expert judgment and opinions of different people representing local perceptions regarding the importance of the different criteria. The following AHP analysis was supported by the use of a specific application (AHP 1.1 – Decision support tools for ArcGIS) for multi-criteria analysis based on pairwise comparison (Marinoni 2004a; Marinoni 2004b).

2.3 Evaluation of road conditions and maintenance costs – Procjena stanja i troškovi održavanja šumskih cesta The basis for the development of a road maintenance plan is a thorough understanding of the road system, its characteristics, and its needs. This is accomplished by establishing and maintaining an intensive inventory of the road system. The inventory has to provide the information necessary for identifying and prioritizing the required maintenance such as categorization of roads, identification of drainage structures and their state and information related to the condition of the road surface (Cavalli et al. 2010).

The most important characteristics of each forest road have been collected and organized in a geodatabase structure. Road survey information included road width, surface type, surface condition, traffic limitation, presence and efficiency of drainage structures, functional and operative classification. To estimate the maintenance needs, the drainage system and surface condition were visually evaluated and rated as reported in Table 2.

The cross-drain culvert spacing (CS) is calculated in relation to the road gradient (slope, in %) using the following formula reported in the forest road manual (Oregon Department of Forestry 2000): CS = 800/slope

(2)

The number of cross-drain structures to be installed is then calculated considering the length of Croat. j. for. eng. 34(2013)1

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

M. Pellegrini et al.

Table 2 Adopted code in the evaluation of the road conditions Tablica 2. Procjena stanja šumskih cesta Drainage structure

Rate

Operation

Stanje odvodnih elemenata

Ocjena

Zahvat

0

No – Bez

1

A

2

D

1

No – Bez

2

B

3

C

Present and functional

The drainage system does not need maintenance

Postojeća funkcionalna

Elementi sustava odvodnje koji ne trebaju održavanje

Present and not functional

The drainage system needs maintenance immediately

Postojeća nefunkcionalna

Elementi sustava odvodnje koji trebaju hitnu sanaciju

Missing

The drainage system is missing

Nedostaje

Nedostaju elementi sustava odvodnje

Surface condition Stanje gornjega ustroja Regular

The road surface is functional and the trafficability is efficient. Road fits perfectly its preeminent function. No evident sign of erosion or potholes Vozna je površina funkcionalna, a prometnost učinkovita. Cesta savršeno ispunjava zadanu funkciju. Nema očitih znakova erozije ni udarnih rupa

Normalno

Partially damaged Djelomično oštećeno

Minor rilling is present on the surface. The condition slightly affects the trafficability. Potholes and erosion process are present but not very evident. The road needs regular surface maintenance intervention Na voznoj su površini prisutne vododerine koje u manjoj mjeri utječu na prometnost. Udarne su rupe i erozija prisutne, ali nisu jako uočljive. Na cesti se trebaju izvoditi zahvati obuhvaćeni redovitim održavanjem Severe rilling is present on the surface

Damaged

Trafficability is affected and sometimes the road cannot perform its function. The road needs an extraregular intervention of surface maintenance

Oštećeno

Na voznoj su površini prisutne velike vovoderine koje utječu na prometnost, a cesta ne ispunjavaju svoju funkciju u potpunosti Na cesti se trebaju izvoditi zahvati obuhvaćeni periodičnim održavanjem

each road segment, where drainage structures are miss ing. The result has to be considered only as an in dication of how detailed placement of cross-drainage culverts has to be evaluated taking into account the specific conditions of individual sites. The evaluation of maintenance needs is conducted to understand the current state of the forest road network and the gap between the actual and optimal conditions that represents the area in which the definition of the best maintenance strategy will be defined.

The result of the survey is a map representing all the maintenance operations needed to upgrade the forest road network. The sum of the maintenance interventions on each road section (node to node) represents one possible alternative in the ranking procedures.

Finally four types of possible maintenance operations were defined. The mean cost for each type of operation was determined through project analysis (Table Croat. j. for. eng. 34(2013)1

3) and the derived mean cost has been used to calculate the economic resources needed to complete each intervention (alternatives) on the forest road network.

3. Results – Rezultati The methodology was applied to a test area of 3991 ha located in the »Altopiano dei Sette Comuni« in the North-Eastern part of Italy (latitude N of 45.56–45.52 longitude E of 11.23–11.28). The region is mainly covered by Norway Spruce (Picea abies) and Beech (Fagus sylvatica) forests. The area represents a meaningful case study because the forest road network has to solve many different preeminent functions.

The recreational function is primarily a consequence of the presence of many historical sites related to the First World War and scenic hiking trails in the area. The silvicultural function is also important as

35

M. Pellegrini et al.

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

Table 3 Types of road maintenance operations performed and related costs Tablica 3. Vrste i relativni troškovi izvođenih radova pri održavanju šumskih cesta Operation type

Operational details

Cost

Vrsta radova

Detalji pri izvođenju pojedine vrste radova

Trošak

A, Drainage structure cleaning

2 operators

A, čišćenje elemenata odvodnje

2 radnika

B, Regular surface management**: addition of crushed aggregate and consolidation

1.10 € DS–1*

1 Tractor with grader – Traktor s grejderskom daskom 1 Truck for gravel transportation – Kamion za transport šljunka

B, redovito održavanje**: dovoz rasutoga materijala i ugradnja

1.21 € m–1

3 operators – 3 radnika 1 Tractor with grader – Traktor s grejderskom daskom

C, Extraordinary surface management**

1 Truck for gravel transportation – Kamion za transport šljunka Excavator – Bager

C, periodično održavanje**

6.32 € m–1

3 operators – 3 radnika D, Drainage structure installation D, Ugradnja odvodnih elemenata

Excavator – Bager 52 € DS–1*

2 operators – 2 radnika Materials – Materijal

* DS: drainage structure (standard length 5 m) – *DS: odvodni elementi (standardna duljina 5 m) ** standard road width 3.50 m – **Standardna širina ceste 3,50 m

expected due to the presence of productive forest. Finally, the forest road network has to guarantee access for the farming activities. The length of the primary forest road network inside the area is 107.8 km (of which 80 km are gravel roads) with a density of 26 m/ha.

Lastly, 68.1 km (65.4%) of the road network is accessible without restriction (public roads) while 39.7 (34.6%) presents restrictions.

3.1 Definition of alternatives and maintenance cost analysis – Određivanje varijanti i troškovna analiza održavanja The field survey shows that 12% of the total extension of the forest road network is in good conditions,

while the remaining 88% requires maintenance interventions. 390 maintenance interventions have been iden tified among a total of 44 road segments (alternatives). The majority of maintenance operations involve the improvement of the drainage system, including the installation of new structures (53% of the total extension) and cleaning of the existing ones (23%). Surface maintenance is required on 11 road segments (20%).

The total investment to complete the maintenance of the project was estimated to 68 297 €. Table 4 shows the number of needed interventions and the estimated budget to complete all the interventions. During the observed period, 2008–2010, the management authorities carried out interventions on 27 forest roads with-

Table 4 Required maintenance interventions and estimated costs Tablica 4. Potrebni radovi pri održavanju i procijenjeni troškovi Interventions, n

Total cost, €

Učestalost izvođenja pojedine vrste radova, n

Ukupni troškovi, €

98

107.8

B (Regular surface management) – B (Redovito održavanje)

6

10 350

C (Extraordinary surface management) – C (Periodično održavanje)

5

34 617

281

14 612

390

68 297

Operation type – Vrsta radova A (Drainage structure cleaning) – A (Čišćenje elemenata sustava odvodnje)

D (Drainage structure installation) – D (Ugradnja odvodnih elemenata) TOTAL – UKUPNO

36

Croat. j. for. eng. 34(2013)1

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

M. Pellegrini et al.

Table 5 Criteria weights as resulted from pairwise grid evaluation Tablica 5. Procjena kriterija udvojenoga vrednovanja Objective 1 – Erosion risk – Cilj 1 – Rizik od erozije

Objective 2 – Social value – Cilj 2 – Socijalna vrijednost

Sub-criteria – Potkriterij

Weight – Težina

Sub-criteria – Potkriterij

Weight – Težina

iGRADE

0.5661

iTOUR

0.4236

IDRAIN

0.2012

iOPER

0.2903

ICOND

0.1319

iFARM

0.1745

iSLOPE

0.1007

iPROD

0.1114

Consistency ratio* (CR): 0.0434

Consistency ratio* (CR): 0.0750

Faktor dosljednosti* (CR): 0,0434

Faktor dosljednosti* (CR): 0,0750

*Revision of preference values is recommended if CR > 0.1 *Preporučuje se preispitivanje vrijednosti ako je CR > 0,1

Table 6 Attribute tables of the features of forest road network Tablica 6. Atributna tablica mreže šumskih cesta i njihovih svojstava Road Id

Sed_Risk

Soc_Val

Eq_Val

Rang

Koristi od šumske ceste

Zaštitna vrijednost

Socijalna vrijednost

Srednja vrijednost

1

50

391

627

2

35

259

3

37

4

Rank

Tot_Cost

C1

C2

C3

509

312

0

0

312

663

461

312

0

1 388

1 702

230

649

440

104

0

0

104

41

200

573

387

312

918

802

2 032

5

58

253

484

368

104

834

0

938

6

40

391

302

347

624

0

7 256

7 880

7

30

403

198

339

728

0

7 713

8 441

8

38

307

288

305

1 976

0

1 693

3 669

9

53

406

442

302

936

0

7 565

8 501

10

28

309

259

298

104

0

0

104

11

42

154

215

281

104

0

0

104

Ukupni trošak

3.2 Rating and ranking of alternatives Ocjenjivanje i vrednovanje varijanti

efit score referred to the relative objective for each cell of the road surface. The combination of the two layers at the highest level of the AHP analysis, produced the final evaluation layer that represented the overall benefit score for each cell. The evaluation layer maps made as a result of the application of the AHP process are presented in Fig. 2.

Table 5 shows the output weights for the different factors taken into consideration, after the AHP pairwise grid evaluation.

Application of the spatial AHP analysis led to the creation of the two raster layers that represent the ben-

Through the use of a zonal statistic tool, the cell values on the evaluation layer were summarized within the forest road network features and the statistics relative to the benefit value of each road have been calculated. Table 6 is an example of the attribute table

in the area for a total of 30.7 km of roads maintained, with the total expense of 87 000 €. According to this, the assumed annual budget for regular maintenance equals 29 000 € (budget constraint).

Croat. j. for. eng. 34(2013)1

37

M. Pellegrini et al.

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

Fig. 2 Maps representing the evaluation layers considering different objectives Slika 2. Karte predstavljaju procijenjene slojeve s obzirom na različite ciljeve of the features of forest road network after the application of the AHP analysis. Roads (ROAD ID) are ranked according to their benefit value, where SED_RISK represents the risk value of sediment production from the road surface, SOC_VAL represents the social value of the road and EQ_VAL represents the combination of the two values (evaluation layer). C1, C2 and C3 represent the costs of different maintenance operations, while TOT_COST represents the total cost to upgrade the road.

3.3 Definition of the optimal maintenance strategies – Utvrđivanje optimalnih strategija održavanja Information contained in the AHP output table (Table 6) can be used to rank the maintenance needs and thus define the priority intervention that should be funded in order to have the highest benefit. For example roads that present a higher erosion risk require more efforts to be maintained in good condition; on the other hand, according to the importance of the territory accessibility, roads with higher social value should be preferably maintained in good conditions.

38

Fig. 3 shows the mapping of the forest road network highlighting the parts that will be maintained with the budget of 29 000 € for regular maintenance, in accordance with different objectives. When only the social value of the roads is considered, the maintenance operations will regard a road extension of 40 km of which 31 km (77%) are actually open to traffic. Considering only the minimization of the erosion, the maintenance operations will regard 8.5 km of which 4.2 km (49%) are open to traffic. In this case, maintenance is concentrated on few roads that are evaluated as more problematic because the condition of the road is critical and maintenance interventions are more expensive.

Finally, considering the combination of the two objectives, the maintenance operations will regard an extension of 29 km of which 25.5 km (89%) are open to traffic.

Evaluating the correspondence between the priorities resulting from the AHP analysis and the actual maintenance strategies, the highest correspondence (65%) is reached with equal consideration of the two objectives, where 18.8 km of the road considered a priCroat. j. for. eng. 34(2013)1

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

M. Pellegrini et al.

Fig. 3 Maps representing the optimal feasible maintenance strategies considering different objectives Slika 3. Karte predstavljaju optimalne strategije održavanja s obzirom na različite ciljeve ority had maintenance intervention in the last three years. In contrast, the consideration of the erosion risk has the lowest correspondence (25%) with 2.2 km that have been maintained. Finally, when the priorities defined considered only the social value, a correspondence of 56% was found, with 22.7 km that have been maintained.

4. Discussion and conclusion – Zaključci s raspravom The economic resources necessary to completely upgrade the forest road network are equal to 67 000 €.

Additionally, the yearly regular maintenance cost is 29 000 €. These values highlight the inadequacy of the

Table 7 Description of maintenance strategies considering different objectives Tablica 7. Opis strategija održavanja s obzirom na različite ciljeve Total interventions

Interventions on public roads

Correspondence with actual management

Ukupno zahvata

Zahvati na javnim cestama

Podudaranje sa stvarnim upravljanjem

km

km

%

Considering social value – S obzirom na socijalnu vrijednost

40

31

56

Considering erosion risk – S obzirom na rizik od erozije

8.5

4.2

25

Considering both objectives – S obzirom na oba cilja

29

25.5

65

Croat. j. for. eng. 34(2013)1

39

M. Pellegrini et al.

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

current annual budget to meet the maintenance requirements of a forest road network in the considered mountainous area. The lack of economic resources affects the »modus operandi« of the management authority that constantly has to choose which roads to keep in good condition. The current strategy guarantees constant maintenance mainly on the roads open to traffic, where road standards and level of traffic need to be higher.

Field surveys suggest that the management strategy should be reconsidered. Erosion and consequent production of sediment from the gravel road surface is a frequent process, directly connected with the worsening of the road conditions. In particular, the general lack of drainage structures seems to be one of the most important factors leading to a consistent delivery of sediment from forest road surface. Due to this situation, the opportunities for structural improvement of the management strategies have been evaluated.

In this context the integrated use of GIS tools and AHP analysis proved to be a valuable tool to better understand the ongoing processes and to give guidelines for determining the maintenance operations. The proposed methodology integrates two different aspects through the pairwise comparison process; the evaluation of the erosion risk and the evaluation of the social value of the roads considering a total of eight criteria. Additionally, the model can be adapted to the preferences of the stakeholders, who can specify which function of the forest road network should be considered preeminent for the specific vocation and needs of the analyzed area.

The resulting evaluation layer has been used to understand the benefits of the required maintenance operations and to define the priorities. The distribution of the priorities considering the combination of the two objectives justifies the maintenance strategy currently practiced, as it indicates that the available budget (29 000 €) is in large part (89%) allocated to the maintenance interventions on the public roads, and that 65% of the resources involve roads that have been recently maintained. On the other hand, the risk of erosion seems to be considered minor because only 25% of the roads with high risk of erosion have been recently maintained.

In conclusion, the use of integrated GIS tools and AHP analysis shows that different aspects can be effectively integrated. Consequently, this approach could be used to improve the efficiency of administration and management of maintenance planning, especially considering that the existing forest road system needs to evolve toward a paradigm where other ben-

40

efits (e.g. recreational value) and priorities (e.g. environmental aspects) are included. Methods for the consideration of these objectives should be developed.

Acknowledgement – Zahvala This study has been developed within the Test site Asiago of the NEWFOR project financed by the European Territorial Cooperation »Alpine Space« (5-3-2FRA).

5. References – Literatura Abdi, E., Majnounian, B., Darvishsefat, A., Mashayekhi, Z., Sessions, J., 2009: A GIS-MCE based model for forest road planning. Journal of Forest Science 55 (4): 171–176.

Bilby, R. E., 1985: Contributions of road surface sediment to a western Washington stream. Forest Science 31(4): 827–838.

Cavalli, R., Grigolato, S., 2010: Influence of characteristics and extension of a forest road network on the supply cost of forest woodchips. Journal of Forest Research 15(3): 202–209.

Cavalli, R., Cappellari, E., Grigolato, S., 2010: Metodologia per la valutazione delle esigenze di viabilità silvo-pastorale in un contesto montano (Method for assessment of forest road network requirement in a mountain area). L’Italia Forestale e Montana 65(3): 313–330.

Coulter, E. D., Sessions, J., Wing, M. G., 2006: Scheduling forest road maintenance using the analytic hierarchy process and heuristics. Silva Fennica 40(1): 143–160.

Dubè, K., Megahan, W., Mccalmon, M., 2004: Washington Road Surface Erosion Model (WARSEM) Manual. Department of Natural Resources, State of Washington, 189 p.

Elliot, W. J., Tysdal, L. M., 1999: Understanding and reducing erosion from insloping roads. Journal of Forestry 97(8): 30– 34.

ESRI, 2011. ArcGIS 10. Redland, CA: Environmental System Research Institute. <http://webhelp.esri.com> (Accessed 28 Febbruary 2012).

Feng, Zhen-Wei, Douglas, R. A., 1993: Logging truck vehicle performance prediction for efficient resource transportation system planning: Computer Modelling Approach. Journal of Forest Engineering 4(2): 7–18.

Fu, B., Lachlan, T. H., Newham, L. T. H., Ramos-Scharron, C. E., 2010: A review of surface erosion and sediment delivery models for unsealed roads. Environmental Modelling and Software 25 (1): 1–14.

Grayson, R. B., Haydon, S. R., Jayasuriya, M. D. A., Finlayson, B. L., 1993: Water quality in mountain ash forests – separating the impacts of roads from those of logging operations. Journal of Hydrology 150(2–4): 459–480.

Hruza, P., Vyskot, I., 2010: Social-Recreation Evaluation of Forest Roads and their Suitability for Trails: Towards a ComCroat. j. for. eng. 34(2013)1

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

M. Pellegrini et al.

plex Approach. Croatian Journal of Forest Engineering 31(2): 127–135.

roads, St John, US Virgin Islands. Earth Surface Processes and Landforms 30(10): 1283–1304.

Luce, C. H., Black, T. A., 1999a: Sediment Production from Forest Roads in Western Oregon. Water Resources Research 35(8): 2561–2570.

Reid, L. M., Dunne, T., 1984: Sediment production from forest road surfaces. Water Resources Research 20(11): 1753– 1761.

Luce, C. H., Black, T. A., 1999b: Changes in erosion from gravel surfaced forest roads through time. In Proceedings of the International Mountain Logging and 10th Pacific Northwest Skyline Symposium. International Union of Forestry Research Organizations and Oregon State University, Corvallis, Oregon, p. 204–218.

Saaty, T. L., 1980: The Analytic Hierarchy Process: Planning, Priority Setting, Resource Allocation. McGraw-Hill, New York.

MacDonald, L. H., Coe, D. B. R., 2008: Road sediment production and delivery: processes and management. In: Proceedings of the First World Landslide Forum, International Programme on Landslides and International Strategy for Disaster Reduction. United Nations University, Tokyo, Japan, p. 381–384.

Malczewski, J., 1999: GIS and Multicriteria Decision Analysis. New York: John Wiley.

Marinoni, O., 2004a: Implementation of the analytical hierarchy process with VBA in ArcGIS. Computers and Geosciences 30(6): 637–646.

Marinoni, O., 2004b: Some words on the analytic hierarchy process and the provided ArcGIS extension ext_ahp. <http:// arcscripts.esri.com/details.asp?dbid=13764> (Accessed 28 Febbruary 2012).

Oregon Department of Forestry, 2000: State Forest Program – Forest Road Manual. <http://www.oregon.gov/ODF/ STATE_FORESTS/roadsmanual.shtml> (Accessed 28 February 2012).

Potočnik, I., 2006: Road traffic in Protected Forest Areas – Case Study in Triglav National Park, Slovenia. Croatian Journal of Forest Engineering 27(2): 116–121.

Ramos-Scharron, C. E., MacDonald, L. H., 2005: Measurement and prediction of sediment production from unpaved

Schmoldt, D. L., Kangas, J., Mendoza, G. A., Pesonen, M., 2001: The Analytic Hierarchy Process in Natural Resource and Environmental Decision Making. Dordrecht, The Netherlands: Kluwer Academic Publishers.

Sheridan, G. J., Noske, P. J., 2007: A quantitative study of sediment delivery and stream pollution from different forest road types. Hydrological Processes 21(3): 387–398.

Shiba, M., 1995: Analytic hierarchy process (AHP) based multi attribute benefit structure analysis of road network system in mountainous rural areas of Japan. International Journal of Forest Engineering 7(1): 14–50.

Simon, H. A., 1960: The new science of management decision. New York: Harper and Row. Talbot, B., Nitteberg, M., 2011. Using transloading times in determining the effect of reduced road standards on the delivered cost of timber. In: Pushing the boundaries with research and innovation in forest engineering. FORMEC 2011 Proceedings. Graz, October 9–13.

Thompson, M., Sessions, J., Boston, K., Skaugset, A., Tomberlin, D., 2010: Forest Road Erosion Control Using Multiobjective Optimization. Journal of the American Water Resources Association 46(4): 712–723.

Ziegler, A. D., Sutherland R. A., Giambelluca, T. W., 2001: Interstorm surface preparation and sediment detachment by vehicle traffic on unpaved mountain roads. Earth Surface Process and Landforms 26(3): 235–250.

Sažetak

Određivanje prioriteta pri održavanju šumskih cesta pomoću višekriterijskih modela odlučivanja: primjena u alpskom području u Italiji U ovom radu kombinacijom GIS-a i analitičkoga hijerarhijskoga procesa (AHP) razvijen je sustav potpore pri odlučivanju (DSS) i vrednovanju prioriteta u održavanju mreže šumskih cesta.

U procesu donošenja odluka čini se da AHP ima mogućnost unaprijediti upravljanje postojećim prometnim sustavima gdje još u potpunosti nisu određeni mjerljivi odnosi između uzroka i posljedica oštećenja.

Predložena metodologija integrira procjenu rizika od erozije s vozne površine šumske ceste uz određivanje multifunkcionalnosti šumskih cesta, što se pokazalo posebno pogodnim za područja gdje se pri upravljanju šumskim cestama moraju uzeti u obzir mnoge različite funkcije koje promatrana šumska cesta obnaša.

Sustav potpore pri odlučivanju podijeljen je u četiri faze. Prva i druga faza bave se ciljevima analize i definiraju hijerarhijsku strukturu rješavanja problema. U tim su fazama definirani čimbenici (kriteriji) kako bi se naglasili ciljevi Croat. j. for. eng. 34(2013)1

41

M. Pellegrini et al.

Spatial Multi-Criteria Decision Process to Define Maintenance Priorities of Forest Road... (31–42)

analize te su definirane metode procjene. Skup kriterija i hijerarhijska struktura DSS-a razvijeni su istraživanjem relevantne literature i analiziranjem mišljenja stručnjaka i ostalih interesnih skupina. Kriteriji koji su korišteni pri procjeni erozijskih procesa su: uzdužni nagib ceste, stanje gornjega ustroja, stanje sustava (elemenata) odvodnje, učestalost prometa (prometno opterećenje), položaj šumske ceste s obzirom na poprečni nagib. Kriteriji korišteni pri procjeni socijalne vrijednosti šumskih cesta su: turističko značenje (pristupačnost turističkim znamenitostima), kategorija šumske ceste, poljoprivredno značenje (pristupačnost poljoprivrednim gospodarstvima/farmama) i gospodarska važnost (pristup šumskoj površini).

U trećoj je fazi primijenjena analiza AHP-a, uz uporabu specijalne aplikacije u GIS-u, pri izračunu (ocjeni) pojedinih slojeva koji predstavljaju važnost svake ceste s obzirom na zadane ciljeve. Vrijednost pojedinih slojeva korištena je u četvrtoj fazi pri rangiranju zahvata u održavanju šumskih cesta.

Navedena metodologija testirana je na mreži primarnih šumskih prometnica u alpskom području u Italiji ukupne duljine od 107,80 km.

Na terenu su prikupljeni podaci o najvažnijim obilježjima pojedine šumske ceste. Terenska je izmjera uključivala širinu šumske ceste, tip gornjega ustroja, stanje vozne površine, prometna ograničenja, pojavnost i učinkovitost elemenata sustava odvodnje, funkcionalnu i operativnu raščlambu. Potreba za održavanjem elemenata sustava odvodnje i stanje vozne površine procjenjivana je vizualno.

Definirane su četiri vrste radnih operacija pri održavanju šumskih cesta te su projektnom analizom ustanovljeni prosječni jedinični troškovi svake radne operacije. Stvarni je proračunski limit određen kao iznos sredstava koja se najčešće utroše pri održavanju šumskih cesta.

Rezulat je analize niz karata na kojima su prikazani prioriteti održavanja za svaku istraživanu šumsku cestu.

Usporedba navedenih rezultata sa stvarnim stanjem šumskih cesta na terenu pokazala je da je integrirana upotreba GIS-a i AHP-a vrijedan alat pri vrednovanju gospodarske važnosti pojedine šumske ceste i pri definiranju prioriteta u održavanju mreže primarnih šumskih prometnica radi povećanja ukupnih koristi uz minimalni utrošak novčanih sredstava.

Prikazana se metoda može primijeniti za poboljšanje učinkovitosti pri upravljanju i planiranju održavanja šumskih cesta, uzmajući u obzir činjenicu da se postojeća mreža šumskih cesta mora razvijati prema pretpostavci da su ostale koristi (npr. rekreativna vrijednost) i prioriteti (npr. ekološke značajke) uključeni.

Ključne riječi: održavanje šumskih cesta, troškovi održavanja, AHP, GIS, DSS

Authors’ address – Adresa autorâ:

Received (Primljeno): April 17, 2012 Accepted (Prihvaćeno): December 10, 2012

42

Marco Pellegrini, PhD.* e-mail: marco.pellegrini@unipd.it Stefano Grigolato, PhD. e-mail: stefano.grigolato@unipd.it Prof. Raffaele Cavalli, PhD. e-mail: raffaele.cavalli@unipd.it University of Padua Department of Land, Environment, Agriculture and Forestry Viale dell’Università 16 35020 Legnaro (PD) ITALY * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

Original scietific paper – Izvorni znanstveni rad

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region in Romania Adrian Enache, Martin Kühmaier, Karl Stampfer, Valentina Doina Ciobanu Abstract – Nacrtak Sound development of forest infrastructure represents the backbone for sustainable forest management. However, planning forest roads, which nowadays must fulfill multiple conflicting objectives, is not an easy task. A GIS based model was developed for supporting decision making in forest road engineering. The tool allowed assessment of forest infrastructure scenarios based on multiple criteria analyses, considering stakeholders’ interests, economic, ecological and social aspects. First, the decision problem was clearly structured and then criteria and sub-criteria were weighted. Then, forest road scenarios were defined and quantitative and qualitative assessments regarding infrastructure and harvesting systems were performed. In the end, utility analysis for each scenario was conducted, the forest road variant with the highest utility score being selected as the most suitable option for implementation. The model was tested and validated in a mountain forest area from Brasov County, Romania. Reduction of mean skidding distance from 864 m to 255–268 m was reported, leading to an increase in productivity of timber extraction from 7.5 m3/h to 11.7 m3/h and to an increased contribution margin from 21.2 €/m3 to 25.1 €/m3. Enhancement of forest infrastructure reduced CO2 emissions re timber harvesting and transport from 8.52 kg/m3 to 7.3 kg/m3. This study showed how multiple attribute utility theory could be used in assessing different forest road options based on a participatory approach.

Keywords: forest roads, multiple criteria decision making, utility analysis, decision support tool, participatory approach

1. Introduction – Uvod Enhancing forest infrastructure has always been a topic of interest among specialists in their quest to provide sound approaches for improving forest accessibility in the context of sustainable forest management (SFM). Several studies have been published regarding automation of road locating (Akay et al. 2005; Aruga 2005; Rogers 2005; Stückelberger et al. 2007) or regarding the impact of forest roads on the environment (Coulter 2004; Akay 2004). However, most of these studies are based on assessments of only one objective function. Recently, Kühmaier et al. (2010) developed a multi-attribute spatial decision support tool for selecting the best suited harvesting systems, taking into account ecological, economic and social aspects. In addition, recent studies have shown that forest roads Croat. j. for. eng. 34(2013)1

fulfill multiple functions; they are of strategic importance in forest operations, they allow access to remote areas, in cases of natural hazards, for tourism and recreational activities (Popovici et al. 2003; Stampfer 2007; Kühmaier et al. 2010). Sustainable development of the forest infrastructure requires harmonization of road planning, designing, construction and maintenance with operational harvesting plans. Thus, planning of forest road routes and skid trails should be approached simultaneously (Pentek et al. 2007). Consideration of environmental and social aspects from the early stages of planning have also been acknowledged (Popovici et al. 2003; Gumus et al. 2008; Ciobanu et al. 2011), underlining the necessity of performing impact assessments when developing forest infrastructure.

Dürrstein (1998) proposed an extensive approach system of cost-efficiency analysis of forest road options,

43

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43â&#x20AC;&#x201C;60)

underlining the importance of participatory process in decision making, while Heinimann (1998) stressed the planning phase should consider assessment of the technical feasibility of alternatives, environmental impact and public involvement in decision making.

Though, dealing with so many variables and constraints is not an easy task for road planners and decision makers. In Romania, Zarojanu (2006, 2007) described a multi-criteria analysis model for optimizing the selection of most suitable forest road option, focusing on the technical aspects of the roads and only marginally addressing the environmental and social aspects. However, these aspects are particularly important for the development of the Romanian forestry sector. The former national strategy (Ministry of Environment and Forests â&#x20AC;&#x201C; MEF 2011a) envisaged the expansion of forest infrastructure in conjunction with GIS, timber harvesting technologies and environmen-

tal constraints. Whilst several strategic actions were established in this respect (i.e. developing secondary forest infrastructure; fostering the utilization of environmentally friendly harvesting technologies), none of these have been implemented so far on large scale in Romania. Moreover, although the road density in Romanian forests is very low (6.5 m/ha, Enescu 2011), the rate of forest road network expansion is also very low (Bereziuc et al. 2003; MEF 2011b). Skidding is the main method for timber extraction, using winch tractors, skidders, gravitational hauling, ox or horse harnesses, while very few forwarders or cable yarders are used. The mean skidding distance at the national level is 1.8 km (Popovici et al. 2003), consequently with a very low productivity in timber extraction. A peculiarity of the Romanian forest sector is the timber sales procedure: on stump or at the road side. In the first case, timber is sold on stump at auctions. Contractors

Fig. 1 Schematic representation of processes in the decision making tool Slika 1. Shematski prikaz procesa u alatu za odluÄ?ivanje

44

Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

then harvest and sell the timber either at the road side or directly to the mill. This behavior triggered empiric development of skid trails, leading to increased residual stand damage and soil erosion. In the second case, harvesting operations are externalized by forest owners, which then sell the timber at the road side.

A well-developed forest infrastructure is a prerequisite for sustainable development of the forest sector.

Thus, the traditional Romanian behavior of building valley forest roads should be changed toward building slope roads that fulfill multiple objectives (Enache et al. 2012). In Romania for decades significant emphasis

Table 1 Criteria and sub-criteria used to measure the performance of forest road alternatives Tablica 1. Kriteriji i potkriteriji za određivanje performansi varijanata šumskih cesta Efficiency Scale (Indicator)

Kriterij

Potkriterij

Ljestvica učinkovitosti (pokazatelj)

A. Upravljanje

Sub-criterion

A. Management

Criterion

A1. Independence from neighors

1= only own property, 0= over neighbors property

A1. Neovisnost o susjedstvu

1 = samo preko vlastite imovine, 0 = preko susjedne imovine

A2. Accessibility for execution of forest operations

% of areas in the 300 m corridor from forest roads

A2. Pristupačnost za izvođenje šumskih radova

Postotak područja do 300 m udaljena od šumskih cesta

A3. Accessibility for game management

Maximum distance from the road to the furthest point in the project area

A3. Pristupačnost za lovno gospodarenje

Najveća udaljenost od ceste do najudaljenije točke u projektnom području

Unit

Objective function

Jedinica Objektivno djelovanje – %

max

m

min

ha

min min

A4. Loss of productive land (road bed clearance)

B. Costs C. Environment protection

C. Zaštita okoliša D. Socijalni čimbenici i rizici

D. Social factors and risks

B. Troškovi

Road length X Opening width A4. Gubitak produktivne površine (oduzeto planumom Duljina ceste X Oduzeta širina ceste) B1. Road construction costs

Annuity of investment effort

€/year

B1. Troškovi izgradnje cesta

Renta od uloženoga napora

€/godini

B2. Road maintenance costs

Total yearly maintenance costs

€/year

B2. Troškovi održavanja cesta

Ukupni godišnji troškovi održavanja

€/godini

B3. Harvesting costs

Total yearly harvesting costs

€/year

B3. Troškovi pridobivanja drva

Ukupni godišnji troškovi pridobivanja drva

€/godini

C1. Protection of ecological valuable areas

Total cumulated distance to ecological valuable areas

C1. Zaštita ekološki vrijednih područja

Ukupno zbrojena udaljenost do ekološko vrijednih područja

C2. Air pollution

CO2 Emissions from harvesting machineries and timber trucks

C2. Onečišćenje zraka

Emisija CO2 od strojeva za pridobivanje drva i kamiona

C3. Visual disturbance of landscape

Number of curves, serpentines, intersections

C3. Vizualno narušavanje krajolika

Broj krivina, serpentina, raskrižja

D1. Accidents in forest operations

Time needed for first aid teams to arrive at accident location

D1. Nesreće na šumskim poslovima

Vrijeme potrebno za dolazak hitne pomoći na mjesto nesreće

D2. Risk of soil erosion and landslides

Risk factor calculated based on models of soil erosions

D2. Rizik od erozije tla i klizišta

Čimbenici rizika izračunati na temelju modela erozije tla

D3. Accessibility for touristic/local/cultural purpose D3. Pristupačnost za turističke, lokalne, kulturne interese

Cumulated distance to the points of interest Zbrojena udaljenost do točaka interesa

D4. Accessibility in case of forest fires

Proportion of areas in the 200 m corridor from forest roads

D4. Pristupačnost u slučaju požara

Postotak područja do 200 m udaljena od šumskih cesta

D5. Accessibility in case of wind- throws

Proportion of areas in the 300 m corridor from forest roads

D5. Pristupačnost u slučaju vjetroizvala i snjegoloma

Postotak područja do 300 m udaljena od šumskih cesta

Croat. j. for. eng. 34(2013)1

min min

m

max

kg/m3

min

no.

min

min

min

–

min

m

min

%

max

%

max

45

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

has been put merely on technical design of forest roads and only recently (Zarojanu 2007; Ciobanu et al. 2011; MEF 2012) environmental aspects have started to be considered in forest road planning.

In this particular context, the aim of this study was to develop an integrative decision support tool for evaluating forest road options based on economic, environmental and social constraints, considering multiple stakeholders’ interests. The main focus was to guide decision makers in selecting the most suitable forest road option using GIS and multiple criteria analyses.

The conceptual model was developed and applied in a mountain forest area in Romania and could be used in other areas with similar local conditions.

2. Material and Methods – Materijal i metode Forest road engineering involves complex decision problems and conflicting objectives that need to be handled simultaneously. Technical feasibility, environmental soundness, social acceptance and economic affordability are the four main pillars on which forest infrastructure must be built (Stampfer 2007). The conceptual model, which includes all these aspects in the decision process and shows the flow of the main processes, is shown in Fig. 1. Based on this model, workflows can be performed individually or simultaneously, depending on the level of automation and on the available data sets. A master table containing sections with input and output data for each work flow process of the decision tool was created. Data from GIS databases, DEMs and forest management plan, as well as results of GIS analyses, cost appraisal, environmental impact evaluation and intermediate results of workflow processes were used as input in the overall utility analysis of the forest road alternatives.

2.2 Multiple criteria utility model for alternatives evaluation – Multikriterijski model korisnosti za vrednovanje varijanata For the evaluation of forest road alternatives based on stakeholders’ preferences, multiple attribute utility theory (MAUT) has been used. One of the most applied MAUT formulas is the linear additive utility function (Kangas et al. 2008, Greene et al. 2010).

Ui =

m

∑a j= 1

j

⋅ c ji

Ui – the overall utility of alternative i, cji – performance of alternative i with respect to criterion j (normalized value), aj – importance weight (preference) of criterion j.

The sum of preference weights is required to be 1.

As sub-criteria are characterized by different efficiency scales with different measurement units, as a first step, all cardinal values of the sub-criteria were normalized to a common comparable scale. In order to do so, the score range procedure was applied, resulting in local values of each indicator, which follow an interval utility scale (Kangas et al. 2008). vi =

c i − min( c ) max( c ) − min( c )

where: vi – normalized value of criterion i, ci – cardinal value of alternative i in the natural scale.

2.1 Structuring a complex decision problem Strukturiranje problema kompleksnoga odlučivanja The complex management problem of enhancing forest infrastructure requires good structuring, with clearly defined goals and objectives. Thus, based on multiple criteria decision analysis tools (Coulter 2004; Lexer et al. 2005; Green et al. 2010), the decision problem has been defined as follows: »which is the most suitable variant of the forest road that should be implemented considering multiple stakeholders’ interests?«.

The problem has then been hierarchically decomposed into four main objectives (criteria) and fifteen subobjectives (sub-criteria) used in the evaluation of different forest road options (Table 1).

46

Fig. 2 Utility function for road construction costs sub-criterion Slika 2. Funkcija korisnosti za potkriterij trošak izgradnje cesta Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

The best alternative was assumed to have the value 1, while the worst had the value 0. It was possible to define thresholds above or below which the function had a constant value. This was particularly important when referring to sub-criteria with a very big variance between alternative values. For example, in case of road construction costs sub-criterion, the linear utility function had the value 1 below 60 000 € and the value 0 above 100 000 € (Fig. 2).

Another possibility would be to use the ratio scale approach, where the criterion values for each alternative are divided with the maximum value among alternatives. Again, the best alternative has the value 1, while the worst has the value 0 (Kangas et al. 2008).

vi =

ci max( c )

Normalized values of each sub-criterion, for each alternative, were then multiplied with their respective importance weights and then summed up, resulting in the final score of an alternative (Greene et al. 2010).

The importance weights for each criterion have been derived based on preferences of different groups of stakeholders (e.g. forest owner, forest manager, environmental protection agency, forest contractor, other authorities) from the Romanian forest sector, through a nationwide survey. Stakeholders were requested to fill their field of expertise and to rate their preferences

Fig. 3 Location of the study area, existing infrastructure and cardinal points (scenario Zero) Slika 3. Lokacija područja istraživanja, postojeća infrastruktura i ključna mjesta (nulta varijanta) Croat. j. for. eng. 34(2013)1

47

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

regarding the importance they gave to criteria and sub-criteria, as listed in Table 1. The preference weights were expressed on a ratio scale. The sum of preference weights for criteria had to be 100%. The preference weights given to sub-criteria of each criterion had also to sum 100%. The response rate in the survey was about 26% from the 103 successfully delivered survey forms to the stakeholders. The details regarding methodology and results of the stakeholder consultation were comprehensively presented in another study.

2.3 Study location – Istraživano područje This study was conducted in a forest area of approximately 903 ha, located in Brasov County (45o55’ North Latitude and 25o90’ East Longitude), Romania (Fig. 3).

The forest is owned by the community of Tarlungeni and managed by Ciucas Autonomous Forest Administration. The forest is located at an altitude of 900–1600 m above sea level. According to the forest management plan, the most common forest types in this area are: mountain beech forests on shallow soils with mull flora; mixed fir-beech forests with mull flora of medium productivity; and beech forests with Festuca Table 2 Key figures about study area Tablica 2. Glavni podaci o istraživanom području Forest area Šumsko područje Average growing stock Prosječna drvna zaliha Total annual growth Ukupni godišnji prirast Total annual allowable cut (AAC) Ukupni godišnji sječivi etat (GSE) AAC thinning Proreda GSE-a AAC final cuts Glavni prihod GSE-a Average timber price on stump Prosječna cijena drva na panju Average timber price at road side Prosječna cijena privučenoga drva Costs of felling-delimbing-sorting Troškovi rušenja, kresanja, razvrstavanja Cleared road bed corridor width Oduzeta širina za planum ceste

48

902.7 ha 296.6 m3/ha 7 198 m3 4 308 m3 1 120 m3 3 188 m3 25.3 €/m3 42.6 €/m3

altissima. The geology is marly flysch, sandstones and massive conglomerates. The hydrological network includes permanent water streams, with maximum flow in spring and minimum in winter. Based on the Köppen classification presented in the forest management plan, the study area is part of the climatic province Dfck, characterized by a boreal climate (D) with precipitation varying between 750 mm and 1000 mm throughout the year (f), with average temperatures for at least three months >10oC (c) and for at least four months >7oC (k).

The average annual temperature is 7.8oC, the average number of days with a snow layer is 71 and the average number of days without frost is 173. One fifth of the studied forest area is located on gentle slopes (<20%), while approximately 10% is located on steep terrain (>55%). The annual allowable cut is about 4310 m3.

Other key figures are presented in Table 2.

2.4 Field survey – Terenska mjerenja Field survey is a necessary stage prior to planning new forest roads and skid trails. Its purpose is to quantitatively and qualitatively evaluate the existing forest infrastructure. Data regarding quality of existing roads have been collected (e.g. damaged road bed, damaged bridges) and inspected elements have been categorized by causative factors. The geographic coordinates were recorded for each identified issue. In addition, a thorough survey of the project area was also performed, identifying and mapping cardinal points for road planning (i.e. possible locations of landing areas, good stream crossing points, ecologically important areas, touristic/local/cultural points of interest), while skidding trails have been mapped using a GPS for recording intervals of 5 seconds. Trails widths, stream crossings, skidding trail segments going entirely through water streams, soil erosion, average gradients and lengths of steep trails have been recorded in the data collection protocol.

For field data collection, the following instruments were used: GPS Garmin 60 CSx GPSMAP for recording geographic coordinates and for mapping skid trails network; Meridian clinometer for slope measurements; Handheld Algiz 7 rugged Tablet PC for the data collection protocol and a Laser LTI TruePulse 360 optic device for distance measurements. For data processing, ESRI® Arc GIS Desktop 10 and Microsoft Office Excel® were used.

7.0 €/m3

2.5 Qualitative assessment of forest infrastructure Kvalitativna procjena šumske infrastrukture

12.0 m

This phase refers to the calculation of several structural indices of the forest road network (Pentek 2005; Bereziuc et al. 2008) based on the data from the field Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

survey and GIS database: road density index, road distance, relative openness, geometric mean skidding distance and actual mean skidding distance. The last two indices have been derived both with analytic formulas and from analysis with ESRI® Arc GIS tools. In the latter case, the first hypothesis for deriving the geometric mean skidding distance was to assume that timber harvested from a forest management unit is concentrated into its centre of gravity and the skidding distance was calculated to these points. However, due to large sizes of management units, the accuracy of this method was very low. So, the assumption was used that harvested timber was concentrated at points located at 100 × 100 m from each other, thus resulting a grid of points based on which the mean skidding distance has been finally determined. In order to obtain the actual mean skidding distance, a correction factor was applied to mean skidding distance depending on local topography (kg).

Studied literature (Segebaden 1964; Amzica 1971; Pentek 2005) mentioned correction factors varying between 1.05–1.70. For the purpose of this study, the correction factor kg was established to 1.50. Relative openness was calculated with classical formulas and by GIS analysis using the buffer method. Statistics regarding levels of forest accessibility were also performed. The automation of work flow processes was performed in Model Builder ™ extension from ESRI® ArcGIS Desktop 10.

Structural indices of the skid trails network were derived in the same way as for the road network.

2.6 Assessment of harvesting systems – Procjena sustava pridobivanja drva The most common harvesting methodology used in Romania is trunks and masts (Ciubotaru 1998), a method similar to the tree-length system in which trees are felled, topped and delimbed at the felling site and then extracted either as full trunks (masts) or as multiple of assortments at the road side. Extraction is usually done by winch tractors (U651) or skidders (TAF) manufactured in Romania. Pre-skidding is a specific operation in timber extraction for Romanian harvesting conditions with low density of forest roads and long skidding distances (Oprea et al. 2008). This is usually done by horse or ox harnesses at distances up to 150–200 m (Ciubotaru 1998), and refers to the transport of timber from stump to the closest skid trail.

However, currently used harvesting systems in the study area are the TAF 657 skidder for final cuts and the winch tractor U651 for thinnings.

2.6.1 Assessment of productivity – Procjena produktivnosti Productive system hour (PSH) is a parameter used in calculation of timber extraction costs. For the purCroat. j. for. eng. 34(2013)1

pose of this study, productivity of harvesting systems was calculated as follows: Since there were no specific local productivity models available, PSH of the U651 winch tractor was determined based on a logarithmic regression function derived from existing time norms (Ciubotaru 1996) that consider the following variables: group of tree species (i.e. coniferous, broadleaves), average tree volume and mean skidding distance.

PSH of the TAF 657 skidder was determined based on a recent local productivity model (Duta 2012), developed for hard-to-reach mountain regions, with similar topographic, site and infrastructure conditions as in this study.

2.6.2 Costs of harvesting systems – Troškovi sustava pridobivanja drva The cost per system hour was calculated for the TAF 657 skidder and the U651 winch tractor using the FAO cost calculation scheme adapted by Holzleitner et al. (2011a), considering an interest rate of 6.5% and including the operator’s costs. The input data used for calculations were the result of discussions with representatives of local forest administration.

2.6.3 Soil erosion and transport of sediments Erozija tla i transport sedimenata Soil erosion and transport of sediments represent key issues in the study area. During the field survey, records were made of damages on residual stands due to timber skidding and areas with massive soil erosion (e.g. depths >150 cm) and sediment transport through water streams. Moreover, several segments of skid tracks were identified as going entirely through permanent water streams. A recent study in harvesting plots with similar site conditions (Sparchez et al. 2009) showed that most of the trees located in a buffer zone of 5 m along skid trails were damaged. An average value of soil dislocation of 40.5 m3/ha was also reported, depending on the type of soil, harvesting method, average tree volume and local topography. In addition, Duta (2012) developed a model for quantifying soil erosion in timber skidding, based on soil type and slope grade of skid trails, reporting an average soil dislocation of 0.713 m3 per running meter of skid trail.

2.6.4 CO2 emissions – Emisija CO2 The Kyoto Protocol calls for active action of all EU member states in reduction of greenhouse gas emissions (2002/358/CE). Under these circumstances, the impact of road construction, harvesting machinery and timber trucks on CO2 emissions was evaluated for each infrastructure scenario. For determining the CO2 emissions from timber transport, the assumptions

49

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

were as follows: a CO2 output factor for diesel engines of 2.65 kg/l, an average fuel consumption on forest roads of 2.05 l/km and a truck payload of 25 m3 (Holzleitner et al. 2011b). Based on several studies on emissions from forest operations (Berg and Karjalainen 2003; Johnson et al. 2005; Markewitz 2006), the evaluation of CO2 emissions from timber extraction was done considering a CO2 output factor of 2.65 kg/l, PSHs of U651 winch tractor and TAF657 skidder, and fuel consumption rates of 7.5 l/h for the U651 winch tractor and 10.0 l/h for the TAF657 skidder.

Regarding road construction impact on CO2 emissions, Loeffler et al. (2008) reported a rate of 3.8 t CO2/km

of forest road built in mixed profile on slopes with gradients less than 50%, for a CO2 output factor of 2.73 kg/l of diesel. Karjalainen and Asikainen (1996) noted a value of 3.3 t CO2/km for forests road built in Finland and for a CO2 output factor of 2.66 kg/l.

2.7 Forest road scenarios – Varijante šumskih cesta The focus of this paper was on the effect that enhancement of forest infrastructure alone had on the current management practices, without considering any changes in harvesting systems. However, the improvement of forest infrastructure created the conditions for adapting current timber extraction practices

Fig. 4 Infrastructure scenarios proposing new forest roads (FR1–FR3) Slika 4. Infrastrukturne varijante predložene novim šumskim cestama (ŠC1–ŠC3)

50

Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

Table 3 Structural indices of forest infrastructure before and after planning new roads Tablica 3. Strukturni pokazatelji šumske infrastrukture prije i nakon planiranja novih cesta Scenarios – Varijante

Structural indices Strukturni pokazatelji Length of road network, m Duljina mreže cesta, m – out of which new forest roads, m – od toga novih šumskih cesta, m Density index of road network, m/ha Klasična otvorenost mreže cesta, m/ha Road distance, m Razmak cesta, m Geometric mean skidding distance SD0, m Geometrijska srednja udaljenost privlačenja SD0, m Mean skidding distance (grid 100 x 100), m Srednja udaljenost privlačenja (mreža točaka 100 x 100), m Maximum skidding distance (grid 100 x 100), m Najveća udaljenost privlačenja ((mreža točaka 100 x 100), m Actual mean skidding distance, m Stvarna srednja udaljenost privlačenja, m Actual maximum skidding distance, m Stvarna najveća udaljenost privlačenja, m Length of skid trails network, m Duljina mreže šumskih vlaka, m Density index of skid trails network, m/ha Klasična otvorenost mreže traktorskih vlaka, m/ha

to state of the art harvesting systems. These issues will be addressed in a further study.

In order to test and validate the conceptual model, a Zero option (current infrastructure conditions) and other three infrastructure scenarios proposing new roads (FR1–FR3) were developed in GIS and considered for the assessment, assuming in all cases current harvesting and skidding means. The alternatives proposing new forest roads were mapped in ESRI® Arc GIS (Fig. 4), based on contour line maps derived from DEM, considering maximum slope grade of the road, terrain steepness and constraint layers developed from cardinal points collected during the field survey.

The processes were automated in Model Builder™.

2.8 Cost evaluation of forest road scenarios Troškovno vrednovanje varijanata šumskih cesta Since the proposed roads are intended to serve for low annual timber traffic (<1000 t), they were considCroat. j. for. eng. 34(2013)1

Zero – Nulta

FR1 – ŠC1

FR2 – ŠC2

FR3 – ŠC3

11 719

25 795

25 327

24 501

–

14 076

13 608

12 782

13.0

28.6

28.1

27.2

770

350

356

368

192

87

89

92

576

170

191

178

1 402

652

710

826

864

255

287

268

2 104

978

1 065

1 238

71 301

67 121

67 349

69 939

79.0

74.4

74.6

77.5

ered as a low category of secondary forest roads with mixed (fill-cut) cross profiles adapted to natural contour lines, road bed widths of 3.5 m, maximum slope grades of 13% for unloaded trucks and 9% for loaded trucks. According to Enescu (2011), adopting this type of forest road, with a gravel finishing adapted to a low timber traffic volume, could lead to a significant reduction of investment effort (between 20–33%). Therefore, considering Romanian average forest road construction costs of about 100 €/m, the unit construction cost for the study area was estimated to 70 €/m. Maintenance costs were estimated to 2 €/m p.a. for valley roads and 1 €/m p.a. for slope roads. The annuity of forest roads was calculated considering discounted total road construction costs (Pičman and Pentek 1996), for an interest rate of 6.5% and an investment life span of 30 years. Total yearly costs for each scenario were calculated as an algebraic sum of annuity, maintenance costs and discounted earnings from road bed clearance. However, only maintenance costs were

51

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

Table 4 Relative openness of the study area Tablica 4. Relativna otvorenost istraživanog područja Distance to road

Accessible forest area by scenario, %

Udaljenost do ceste

Pristupačnost šumskomu području po varijantama, %

m

Zero – Nulta

FR1 – ŠC1

FR2 – ŠC2

FR3 – ŠC3

100

9%

35%

35%

35%

200

17%

62%

58%

62%

300

25%

82%

76%

79%

500

43%

99%

94%

96%

750

64%

100%

100%

99%

1 200

90%

–

–

100%

>1 200

100%

–

–

–

considered in scenario Zero, as the investment has already been paid off (road network older than 30 years).

For each infrastructure scenario, the total harvesting costs were calculated based on unit costs and PSH of each harvesting system. Incomes from timber sales were calculated in accordance with both timber selling procedures, on stump and at road side, based on prices provided by the local forest administration. In the end, the net profit-loss statement for each scenario and each procedure of timber sale was calculated.

3. Results – Rezultati During the assessment process, scenario Zero was compared with the other scenarios (FR1–FR3) based on the specified criteria and sub-criteria weighted by stakeholders’ preferences.

3.1 Qualitative assessment of infrastructure scenarios – Kvalitativna procjena infrastrukturnih varijanata In scenario Zero, the access in study area is possible through two valley forest roads and one segment of a public road in total length of 11.72 km, which provide an uneven opening of forest stands. Scenarios FR1–FR3 propose between 12.8 km and 14.1 km of new forest roads (Table 3), improving the accessibility in the studied forest area. The road density increased from 13.0 m/ha (scenario Zero) to 27.2–28.6 m/ha (scenarios FR1–FR3), while the actual mean skidding distance reduced from 864 m (scenario Zero) to about 255–287 m (scenarios FR1–FR3). Thus, good premises for improving productivity and cost efficiency of harvesting systems were created.

52

A total of 71.3 km of skid trails were mapped during the field survey (scenario Zero). Most of the skid trails were developed on the line of the steepest slope alongside the stream or creek bed, causing massive soil erosion and transport of sediments, a common case being 1.0–1.5 m deep ravines. For scenarios FR1–FR3, the length of skid trails network decreased from 1.4 km to 4.2 km, depending on the case, due to possibilities of partial using of the existing skid trails in planning new roads (Table 3). Hence, the density index of secondary infrastructure ranges between 74.4 m/ha (scenario FR1) and 79.0 m/ha (scenario Zero).

The relative openness of the study area is presented in Table 4. In case of scenario Zero, about 43% of the area is accessible for buffer strip of 500 m from the roads, while approximately 90% is accessible for 1200 m buffer strip. Admittedly, in case of scenarios FR1–FR3, 94% to 99% of the forest area is accessible for a buffer strip of 500 m, while 58% to 62% is accessible for 300 m buffer strip.

3.2 Assessment of harvesting systems – Procjena sustava pridobivanja drva Assessment of harvesting systems was performed in terms of productivity, costs and impact on the environment for each infrastructure scenario.

3.2.1 Productivity and costs – Produktivnost i troškovi Productive system hour (PSH) of harvesting systems has significantly increased in scenarios FR1–FR3 proposing new roads (Fig. 5), when compared to the current infrastructure situation. PSH of the U651 winch tractor increased from 1.9 m3/h (scenario Zero) to 3.0 m3/h (scenarios FR1 and FR3), while the PSH of TAF 657 skidCroat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

Table 5 CO2 emissions from harvesting systems and timber transport Tablica 5. Emisija CO2 od sustava pridobivanja drva i prijevoza drva Scenario – Varijanta

Indicator

Harvesting system

Pokazatelj

Sustav pridobivanja drva

Zero – Nulta

FR1 – ŠC1

FR2 – ŠC2

FR3 – ŠC3

Winch Tractor U651

11.5

7.2

7.4

7.2

Skidder TAF 657

5.8

3.7

3.8

3.7

Timber transport – Prijevoz drva

1.3

2.8

2.8

2.7

CO2 emissions, kg/m3 Emisija CO2, kg/m3

der improved from 7.5 m3/h (scenario Zero) to 11.7 m3/h (scenario FR1), triggering also important costs reductions in timber harvesting. Scenario FR1 had the lowest costs of timber extraction for both U651 tractor (8.9 €/m3) and TAF657 skidder (5.4 €/m3), when compared to other scenarios (Fig. 6). However, only minor differences were noticed between scenarios FR1–FR3.

3.2.2 Impact on the environment – Utjecaj na okoliš Soil erosion could be limited by reducing the skidding distance and by closing several unnecessary skid trails. The values of dislocated soil due to timber skidding calculated according to Duta (2012) ranged between 47 858 m3 (scenario FR1) and 50 838 m3 (scenario Zero).

Table 6 Cost appraisal of infrastructure scenarios Tablica 6.Troškovna procjena infrastrukturnih varijanata Scenario – Varijanta Road network length, m Duljina mreže cesta, m – out of which new roads, m – od toga novih cesta, m Construction cost, €/m Troškovi izgradnje, €/m Total construction costs, € Ukupni troškovi izgradnje, € Annual interest rate, %

Zero – Nulta

FR1 – ŠC1

FR2 – ŠC2

FR3 – ŠC3

11 719

25 795

25 327

24 501

–

14 076

13 608

12 782

952 560

894 740

70 –

985 320 6.5

Godišnja kamatna stopa, % Life span of investment, years Životni vijek investicije, godine Annuity road construction, € Renta izgradnje cesta, € Maintenance costs, € Troškovi održavanja, € Area of road clearance, ha Površina oduzeta cestom, ha Volume from road clearance, m3 Volumen oduzet cestom, m3 Earnings from road bed clearance, € Zarada od oduzete površine planuma ceste, € Discounted annual earnings road clearance, € Godišnja zarada s popustom od oduzete površine ceste, € TOTAL ROAD COSTS, €/year UKUPNI TROŠKOVI CESTA, €/godini Croat. j. for. eng. 34(2013)1

30 –

75 453

72 945

68 517

23 438

31 655

31 187

30 361

–

16.9

16.3

15.3

–

5 009.3

4 842.8

4 548.8

–

178 376

172 446

161 978

–

13 660

13 205

12 404

23 438

93 448

90 926

86 474

53

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

Fig. 5 Timber extraction productivity, by infrastructure scenario Slika 5. Produktivnost privlačenja drva po infrastrukturnim varijantama

Fig. 6 Timber extraction costs, by infrastructure scenario Slika 6. Troškovi privlačenja drva po infrastrukturnim varijantama

Due to improved infrastructure, a significant reduction in CO2 emissions from harvesting systems could be achieved in scenarios FR1–FR3 (Table 5), from 11.5 kg CO2/m3 (scenario Zero) to 7.2 kg CO2/m3 (FR1 and FR3) in the case of the U651 tractor and from 5.8 kg CO2/m3 (scenario Zero) to 3.7 kg CO2/m3 (FR1 and FR3) in the case of the TAF657 skidder. Regarding CO2 emissions due to timber transport inside the project area (Table 5), values range between 1.3 kg CO2/m3 (scenario Zero) and 2.8 kg CO2/m3 (scenario FR1).

3.3 Cost evaluation of forest road scenarios Troškovno vrednovanje varijanata šumskih cesta

to scenario FR1 (15 574 € p.a.). However, it has to be underlined that all scenarios proposing new roads (FR1–FR3) were profitable. Second, when considering timber sales at the road side, the highest profit was recorded again in scenario Zero (91 300 € p.a.), while the lowest was noted in scenario FR1 (32 756 € p.a.).

Scenarios FR1–FR3 proved again to be all profitable (Table 7). In addition, the contribution margin for harvesting operations increased from 1.33 €/m3 (scenario Zero) to 3.99 €/m3 (scenario FR1). Thus, in terms of their overall profit performance, it would be presumably better to change the selling procedure from stumpage to road side.

Total road cost is an indicator used in the overall utility evaluation of scenarios, with a relevant significance assigned by stakeholders. Therefore, cost analysis was conducted for all infrastructure scenarios (Table 6). The highest total road costs are required by scenario FR1 (93 448 € p.a.), while scenario Zero has the lowest costs (23 438 € p.a.). When considering only new roads, the lowest road cost scenario is FR3 (86 474 € p.a.).

Net profit-loss statements were calculated for all scenarios and each timber sales procedure (Table 7).

First, in the case of timber sales on stump (current practice), the highest net profit was noted for scenario Zero (85 584 € p.a.), which did not involve any construction costs, while the lowest profit was attributed

Provided that the current stumpage sales method is used, investment in new roads (FR1–FR3) would make no sense from the forest owner’s point of view, since all profit would represent the forest contractors’ profit. Therefore, investing in new forest infrastructure would only make sense if the timber sales procedures were replaced by selling timber at the road side. In this situation, the profits would be to the benefit of the forest owner. Furthermore, subsidies between 50% for private owned forests and 100% for local community forests are available for investing in forest infrastructure through EU Rural Development Programme (MARD 2012). Thus, contribution margin of forest administration could increase from 21.2 €/m3 (scenario Zero) up to 25.1 €/m3 (scenario FR1) (Table 7).

54

Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

Table 7 Profit and loss statement by infrastructure scenarios Tablica 7. Dobit i gubitak po infrastrukturnim varijantama Scenario – Varijanta Zero – Nulta

FR1 – ŠC1

Timber felling-delimbing-sorting, €

FR2 – ŠC2

FR3 – ŠC3

30153

Rušenje, kresanje, razvrstavanje drva, € Timber extraction cost, € Troškovi privlačenja drva, €

38 647

27 180

27 835

27 611

–

11 467

10 812

11 036

Savings timber extraction, € Uštede pri privlačenju drva, € Net income timber sales on stump, €

109 022

Neto prihod od prodaje drva na panju, € Income timber sales at the road side, €

183 537

Prihodi od prodaje privučenoga drva, € Net income timber sales at the road side, € Neto prihod od prodaje privučenoga drva, € Total road costs, €/year Ukupni troškovi cesta, €/godini Total road costs (€/year) with EU incentives Ukupni troškovi cesta (€/godini) s poticajima EU-a

114 738

126 204

125 549

125 773

–23 438

–93 448

–90 926

–86 474

–23 438

–17 995

–17 981

–17 957

NET PROFIT-LOSS STATEMENT – NETO DOBIT I GUBITAK Timber sales on stump, €/year Prodaja drva na panju, €/godini Timber sales at road side, €/year Prodaja privučenoga drva, €/godini Timber sales at road side and EU incentives, €/year Prodaja privučenoga drva i poticaji EU-a, €/godini

85 584

15 574

18 097

22 549

91 300

32 756

34 624

39 300

91 300

108 209

107 568

107 817

TOTAL CONTTRIBUTION MARGIN FOR FOREST ADMINISTRATION UKUPNA KONTRIBUCIJSKA MARŽA ZA UŠP 3

Timber sales on stump, €/m Prodaja drva na panju, €/m3

Timber sales at road side, €/m3 Prodaja privučenoga drva, €/m3 Timber sales at road side and EU incentives, €/m3 Prodaja privučenoga drva i poticaji EU-a, €/m3

3.4 Utility analysis and decision making – Analiza korisnosti i odlučivanje Based on stakeholders’ preferences regarding the importance of defined criteria and sub-criteria, the overall utility value of each scenario was calculated (Table 8). Stakeholders’ consultation showed that the most important sub-criteria were: accessibility for performing silvicultural operations (20%), protection of ecologically important areas (14%) and road construcCroat. j. for. eng. 34(2013)1

19.87

3.62

4.20

5.23

21.20

7.60

8.04

9.12

21.20

25.12

24.97

25.03

tion costs (11%). The least important one was the accessibility for touristic, local or cultural points of interest (1%). According to MAUT, the best alternative is the one with the highest score in total. With a total score of 0.682, scenario FR3 is the alternative that would best satisfy stakeholders’ preferences and thus it would be recommended for implementation (Fig. 7).

The decision support tool for evaluating forest road alternatives presented in this study was tested and

55

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

Table 8 Multiple utility analysis of infrastructure scenarios Tablica 8. Višestruka analiza korisnosti infrastrukturnih varijanata Final scoring of alternatives – Konačno bodovanje varijanata Code

Criteria

Weight

Šifra

Kriterij

Važnost

A1 A2 A3 A4 B1 B2 B3 C1 C2 C3 D1 D2 D3 D4 D5

Neighbors independence Neovisnost o susjedstvu Accessibility for forest operations Pristupačnost za šumske poslove Accessibility for game management Pristupačnost za lovno gospodarenje Loss of productive land (road clearance) Gubitak šumskoga zemljišta (širina ceste) Construction costs Troškovi gradnje Maintenance costs Troškovi održavanja Harvesting costs Troškovi pridobivanja drva Protection of ecologically valuable areas Zaštita ekološki vrijednih područja CO2 emissions – Emisija CO2 Visual disturbance due to Curves/Intersections Vizualni poremećaj zbog krivina i raskrižja Fewer accidents with personal injuries Manji broj nesreća s osobnim ozljedama Risks of soil erosions and/or landslides Rizik od erozije tla ili klizišta Accessibility for touristic/local/cultural interest Pristupačnost za turističke, lokalne, kulturne interese Accessibility in case of forest fires Pristupačnost u slučaju požara Accessibility in case of wind-throws/snow Pristupačnost u slučaju vjetroizvala i snjegoloma Total score – Ukupni rezultat

SCENARIO – Varijanta Zero – Nulta

FR1 – ŠC1

FR2 – ŠC2

FR3 – ŠC3

UUV

WUV

UUV

WUV

UUV

WUV

UUV

WUV

8%

1.0

0.079

1.0

0.079

1.0

0.079

1.0

0.079

20%

0.0

0.000

1.0

0.201

0.9

0.179

1.0

0.192

4%

0.0

0.000

1.0

0.045

0.9

0.041

0.8

0.034

5%

1.0

0.049

0.0

0.000

0.0

0.002

0.1

0.004

11%

1.0

0.110

0.6

0.067

0.7

0.074

0.8

0.086

7%

1.0

0.070

0.0

0.000

0.1

0.004

0.2

0.011

8%

0.0

0.000

1.0

0.078

0.9

0.073

1.0

0.075

14%

1.0

0.143

0.0

0.000

0.2

0.027

0.2

0.027

5%

0.0

0.000

0.9

0.043

0.8

0.040

1.0

0.047

6%

1.0

0.057

0.3

0.014

0.0

0.000

0.3

0.019

3%

0.0

0.000

0.9

0.025

0.7

0.021

1.0

0.029

3%

0.0

0.000

1.0

0.026

0.6

0.017

0.7

0.018

1%

0.0

0.000

1.0

0.026

0.8

0.017

0.8

0.018

3%

0.0

0.000

1.0

0.013

0.9

0.011

1.0

0.010

2%

0.0

0.000

1.0

0.033

0.9

0.030

1.0

0.033

100%

6.0

0.507

10.6

0.651

9.5

0.614

10.6

0.682

* UUV - unweight utility values – neponderirane vrijednosti korisnosti; WUV – weighted utility values – ponderirane vrijednosti korisnosti

validated based on a participatory process. Considering multiple stakeholders’ interests, all scenarios proposing new roads (FR1–FR3) performed better in overall terms than scenario Zero. Based only on normalized utility values of each sub-criterion (before weighting each sub-criterion with stakeholders’ preferences), the total scoring showed that scenarios FR1 and FR3 were

56

ranked equal first (Table 8). When stakeholders’ preferences were considered, FR3 was the best performing scenario. Thus, it could be concluded that stakeholders’ preferences do have significant importance. Therefore, sensitive analyses were conducted in order to show how changes in stakeholders’ preferences for specific criteria or sub-criteria could affect the final Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

results. As an example, the sensitivity analysis regarding accessibility for forest operations sub-criterion was performed. Fig. 8 shows that scenario FR3 would perform best for preference weights up to 60% given to this sub-criterion. If the preference weight were above 60%, than scenario FR1 would be recommended for implementation. Regardless the preference given by the stakeholders to this sub-criterion, scenario Zero had the lowest score. Similarly, sensitivity analyses could be performed for all other criteria and sub-criteria.

4. Discussions and conclusions – Rasprava i zaključci

Fig. 7 Final score of scenarios after multiple utility analyses, based on sub-criteria A1–D5 Slika 7. Konačni rezultati varijanata nakon višestruke analize korisnosti, temeljene na potkriterijima A1–D5

Fig. 8 Sensitive analysis regarding performance of accessibility for forest operations sub-criterion Slika 8. Osjetljiva analiza o djelovanju potkriterija pristupačnost za šumske poslove Croat. j. for. eng. 34(2013)1

The aim of this study was to develop a decision support tool for evaluating different forest road options before technical design, using a participatory approach and multiple criteria analyses. Based on clearly defined criteria and sub-criteria, qualitative and quantitative assessments of forest infrastructure scenarios were performed. The conceptual model of the decision support tool showed a clear flow of processes and how the evaluation of forest road options could be done. The main processes refer to locating new roads, assessment of productivity and appraisal of cost efficiency in timber extraction, evaluation of impact on the environment and finally, the utility analysis of infrastructure scenarios. The model was tested and validated in a mountainous forest located in Romania. A suitable road variant based on stakeholders’ preferences was recommended for implementation. Thus, the importance of the preliminary planning and assessment phase in forest road engineering was highlighted.

The multiple attribute utility theory (MAUT) proved to be an appropriate tool for evaluating forest road alternatives because, among others, it also allowed sensitivity analyses regarding the importance of stakeholders’ preferences in the final score of alternatives. In comparison to the analytic hierarchy process (AHP) used by Coulter (2004), which is a more complex tool requiring expert judgments based on pairwise comparisons, MAUT was preferred in this study for its simplicity in use and its proven practicality in the development of decision support tools in the forestry sector (Lexer et al. 2005; Kangas et al. 2008).

In addition, this study continued and extended the work of Zarojanu (2006; 2007), comprehensively and soundly addressing the economic, ecological and social aspects in selecting the most suitable forest road option, as recommended in the literature by Dürrstein (1998) and Heinimann (1998). Thus, this model proved

57

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

its utility for supporting decision making in forest road engineering and could be used in other regions with similar topographic, forest site and social-cultural conditions. The decision support tool presented in this study could be improved by further process automation and by extending it with the assessment of the impact of new harvesting systems that could be introduced in the study area in the overall utility analysis of the infrastructure scenarios.

Acknowledgements – Zahvala This paper is supported by the Sectoral Operational Programme Human Resources Development (SOP HRD), ID76945 financed from the European Social Fund and by the Romanian Government.

Ciubotaru, A., 1996: Elemente de proiectare si organizare a exploatarii padurilor (Elements of designing and organizing timber harvesting). Lux Libris Publishing House, Braşov, 234 p.

Ciubotaru, A., 1998: Exploatarea padurilor (Timber harvesting), Lux Libris Publishing House, Braşov, 351 p.

Coulter, E. D., 2004: Setting forest road maintenance and upgrade priorities based on environmental effects and expert judgment. PhD Dissertation, Oregon State University Corvallis, 199 p.

Duta, I., 2012: Researches on wood collection in hard-to-reach mountain regions under specific conditions of Soveja III Production Unit, Vrancea County. PhD Thesis, Transilvania University of Brasov, Faculty of Silviculture and Forest Engineering, 189 p.

5. References – Literatura

Dürrstein, H., 1998: Opening up of a mountainous region – decision making by integration of the parties concerned applying cost efficiency analysis, Proceedings of the FAO – Seminar on Environmentally Sound Forest Roads and Wood Transport, Sinaia, Romania.

Akay, A. E., Karas, J. R., Sessions, J., Yuksel, A., Bozali, N., Gundogan, R., 2004: Using high resolution digital elevation model for computer aided forest road design. Proceedings of International Society for Photogrammetry and Remote Sensing Congress Istanbul 2004 VII(B7): 1682–1750.

Enache, A., Ciobanu, V. D., Pertlik, E., 2012: Approaches regarding environmental impact assessment of forest roads with a special emphasis on Romanian forestry sector. Bulletin of the Transilvania University of Brasov. Series II – Forestry.

Wood industry. Agricultural food engineering 5(54): 63–72.

Akay, A. E., Sessions, J., 2005: Applying the Decision Support System TRACER to Forest Road Design. Western Journal of Applied Forestry 20(3): 184–191(8).

Enescu, A. H., 2011: Research about the usefulness and economic efficiency of forest transport network development, in the mountain area, with summary arranged roads and their characteristics. PhD Thesis, Transilvania University of Brasov, Faculty of Silviculture and Forest Engineering.

Amzica, A., 1971: Contributii la studiul desimii optime a retelei de drumuri auto forestiere din Romania. [Contributions to the study of the optimum forest road network density in Romania]. PhD Dissertation. Polytechnic Institute of Brasov, Faculty of Forestry, 246 p.

Aruga, K., 2005: Tabu search optimization of horizontal and vertical alignments of forest roads. Journal of Forest Research 10(4): 275–284.

Bereziuc, R., Alexandru, V., Ionescu, M., 2003: Eficienţa economică a programării şi urmăririi automate a lucrărilor în execuţia drumurilor forestiere (Economic efficiency of scheduling and automated follow-up of works in forest road construction). Bucovina Forestiera XI(1): 3–11.

Bereziuc, R., Alexandru, V., Ciobanu, V., Ignea, Gh., 2008: Elemente pentru fundamentarea normativului de proiectare a drumurilor forestiere (Elements for the substantiation of the normative of forest roads designing). Transilvania University of Brasov Publishing House, 393 p.

Berg, S., Karjalainen, T., 2003: Comparison of greenhouse gas emissions from forest operations in Finland and Sweden. Forestry 76(3): 271–284.

Ciobanu, V. D., Alexandru, V., Borz, S. A., Mihaila, M., Dumitrascu, A. E., 2011: Calculus and evaluation methods for forest roads execution impact upon the environment. International Journal of Energy and Environment 5(5): 686–693.

58

Greene, R., Luther, J. E., Devillers, R., Eddy, B., 2010: An approach to GIS-based multiple criteria decision analysis that integrates exploration and evaluation phases: Case study in a forest – dominated landscape. Forest Ecology and Management 260(12): 2102–2114.

Gumus, S., Acar, H. H., Toksoy, D., 2008: Functional forest road network planning by consideration of environmental impact assessment for wood harvesting. Environmental Monitoring and Assessment 142(1–3): 109–116.

Heinimann, H., 1998: Oppening up planning taking into account environmental and social integrity. Proceedings of the FAO – Seminar on Environmentally Sound Forest Roads and Wood Transport, Sinaia, Romania.

Holzleitner, F., Stampfer, K., Visser, R., 2011a: Utilization rates and cost factors in timber harvesting based on long-term machine data. Croatian Journal of Forest Engineering 32(2): 501–508.

Holzleitner, F., Kanzian, C., Stampfer, K., 2011b: Analyzing time and fuel consumption in road transport of round wood with an onboard fleet manager. European Journal of Forest Research 130(2): 293–301.

Johnson, L. R., Lippke, B., Marshall, J. D., Comnick, J., 2005: Life-cycle impacts of forest resource activities in the Pacific Croat. j. for. eng. 34(2013)1

An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60) A. Enache et al.

Northwest and Southeast United States. Wood and Fibre Sciences 37 (Corrim Special Issue): 30–46.

Karjalainen, T., Asikainen, A., 1996: Greenhouse gas emissions from the use of primary energy in forest operations and longdistance transportation of timber in Finland. Forestry 69(3): 215–228.

Kangas, A., Kangas, J., Kurttila, M., 2008: Decision support for forest management. Springer Science + Business Media B.V., 223 p.

Kϋhmaier, M., Stampfer, K., 2010: Development of a MultiAttribute Spatial Decision Support System in Selecting Timber Harvesting Systems. Croatian Journal of Forest Engineering 31(2): 75–88.

Lexer, M. J., Vacik, H., Palmetzhofer, D., Oitzinger, G., 2005: A decision support tool to improve forestry extensions service for small private forest owners in southern Austria. Computers and Electronics in Agriculture 49: 81–102.

Loeffler, D., Jones, G., Vonessen, N., Healey, S., Chung, W., 2008: Estimating Diesel Fuel Consumption and Carbon Dioxide Emissions from Forest Road Construction. In: McWilliams, Will; Moisen, Gretchen; Czaplewski, Ray, comps. Forest Inventory and Analysis (FIA) Symposium 2008, 11 p.

Markewitz, D., 2006: Fossil fuel carbon emissions from silviculture: Impacts on net carbon sequestration in forests. Forest Ecology and Management 236(2–3): 153–161.

Oprea, I., Derczeni, R., Iordache, E., Giannoulas, V.J., 2008: Technological lines for logging in reduced accessibility conditions. Bulletin of the Transilvania University of Brasov, Series II – Forestry. Wood industry. Agricultural food engineering 1(50): 21–24.

Pentek, T., Picman, D., Potocnik, I., et al., 2005: Analysis of an existing forest road network. Croatian Journal of Forest Engineering 26(1): 39–50.

Pentek, T., Nevecerel, H., Picman, D., Porsinsky, T., 2007: Forest road network in the Republic of Croatia - Status and perspectives, Croatian Journal of Forest Engineering 28(1): 93–106.

Picman, D., Pentek, T., 1996: The influence of forest roads building and their maintenance costs on their optimum density in low lying forests of Croatia. In: Proceedings of the FAO – Seminar on Environmentally Sound Forest Roads and Wood Transport, Sinaia, Romania.

Popovici, V., Bereziuc, R., Clinciu, I., 2003: Extinderea retelei de drumuri pentru accesibilizarea fondului forestier si, in general, a padurii (Extension of road network for improving the access to forests and forest fund). Bucovina Forestiera XI(2): 36–40.

Croat. j. for. eng. 34(2013)1

Rogers, L., 2005: Automating contour-based route projection for preliminary forest road designs using GIS. Master Thesis.

University of Washington, 87 p.

Segebaden, G., 1964: Studies of cross-country transport distances and road net extension. Studia Forestalia Suecica 18, 69 p.

Spârchez, Gh., Derczeni, R., Iordache, E., Drosos, V., 2009: The impact of different carriages on soil and trees during skidding in the Romanian forests. Bulletin of the Transilvania University of Brasov, Series II - Forestry, Wood industry and Agricultural food engineering 2(51): 35–44.

Stampfer, K., 2007: Harvesting Systems for Mountainous Regions – Course Script. Institute of Forest Engineering, Department of Forest and Soil Sciences, University of Natural Resources and Life Sciences - BOKU, Vienna.

Stuckelberger, J., A., 2007: A weighted-graph optimization approach for automatic location of forest road networks. PhD Dissertation, Vdf Hochschulverlag AG an der ETH, Zürich.

Zarojanu, D., Duduman, G., 2006: Consideratii privind stabilirea traseelor de drumuri forestiere (Considerations regarding the establishment of forest road routes). Revista Padurilor 121(6): 46–49.

Zarojanu, D., 2007: Drumuri Forestiere (Forest Roads). University of Suceava Publishing House, 258 p.

Legislation and Policies: 2002/358/CE: Council Decision of 25 April 2002 concerning the approval, on behalf of the European Community, of the Kyoto Protocol to the United Nations Framework Convention on Climate Change and the joint fulfilment of commitments hereunder.

World Wide Web sites: Ministry of Agriculture and Rural Development - MARD (2012): European Agricultural Fund for Rural Development (2007–2013), <http://www.madr.ro/pages/page.php?self=03& sub=0302&tz=030201>, accessed on 15.09.2012.

Ministry of Environment and Forests, 2011a: The policy and strategy for the development of Romanian forestry sector (2001-2010), <http://www.mmediu.ro/paduri/politici_forestiere.htm>, accessed on 07.12.2011.

Ministry of Environment and Forests (2011b): State of Romanian forests - 2010, <http://www.mmediu.ro/paduri/management_forestier.htm>, accessed on 07.12.2011.

Ministry of Environment and Forests, 2012: Normative PD003-11 regarding designing of forest roads, <http://www. mmediu.ro/beta/wp-content/uploads/2012/08/2012-08-07_ paduri_normativproiectaredrumuriforest.pdf>, accessed on 15.09.2012.

59

A. Enache et al. An Integrative Decision Support Tool for Assessing Forest Road Options in a Mountainous Region ... (43–60)

Sažetak

Integracijski alat za odlučivanje pri procjeni varijanata šumskih cesta u planinskom području Rumunjske Razuman je razvoj šumske infrastrukture okosnica za održivo gospodarenje šumama. Međutim, današnje planiranje šumskih prometnica mora ispuniti više sukobljenih ciljeva, što nije jednostavan zadatak. Model temeljen na GIS-u razvijen je za potporu odlučivanja u inženjeringu šumskih cesta. Alat dopušta procjenu šumskih infrastrukturnih varijanata na temelju analize različitih kriterija s obzirom na interese sudionika, gospodarske, ekološke i socijalne aspekte. Prvo, problem pri odlučivanju jasno je strukturiran, a zatim su ponderirani kriteriji i potkriteriji.

Nakon toga su definirane varijante šumskih cesta te su izvedene kvantitativne i kvalitativne procjene infrastrukture i sustava pridobivanja drva. Na kraju je provedena analiza korisnosti za svaku varijantu; varijanta šumske ceste s najvišom ocjenom korisnosti odabrana je kao najprikladnije rješenje za provedbu. Model je provjeren i potvrđen u planinskom šumskom području županije Braşov u Rumunjskoj. Šumsko se područje nalazi na nadmorskoj visini od 900 do 1600 m. Jedna petina promatranoga šumskoga područja nalazi se na blagim padinama (<20 %), dok se oko 10 % nalazi na strmom terenu (>55 %). Postignuto je smanjenje srednje udaljenosti privlačenja s 864 m na 255 – 268 m, što dovodi do povećanja produktivnosti privlačenja sa 7,5 m3/h na 11,7 m3/h te do povećanja kontribucijske marže s 21,2 €/m3 na 25,1 €/m3. Unapređenje šumske infrastrukture smanjuje emisiju CO2 prilikom privlačenja drva i transporta s 8,52 kg/m3 na 7,3 kg/m3. Ovo istraživanje pokazuje kako se multikriterijska analiza korisnosti može upotrijebiti u procjeni različitih varijanata šumskih cesta temeljenih na zajedničkom pristupu. Multikriterijska analiza korisnosti (MAUT) pokazala se kao prikladno sredstvo za procjenu varijanata šumske ceste jer, među ostalim, uključuje analizu osjetljivosti s obzirom na preferencije sudionika u konačni rezultat varijanata. Alat za podršku odlučivanju prikazan u ovom istraživanju može biti poboljšan daljnjim procesom automatizacije i njegovim proširenjem za nove sustave privlačenja koji bi mogli biti uključeni u područje istraživanja.

Ključne riječi: šumske ceste, multikriterijsko odlučivanje, analiza korisnosti, alat za odlučivanje, zajednički pristup

Authors’ address – Adresa autorâ: Adrian Enache, MSc.* e-mail: adrian.enache@unitbv.ro Prof. Valentina Doina Ciobanu, PhD. e-mail: ciobanudv@unitbv.ro Transilvania University of Brasov Department of Forest Engineering Forest Management Planning and Terrestrial Measurements Sirul Beethoven 1 500123, Brasov ROMANIA

Received (Primljeno): December 12, 2012 Accepted (Prihvaćeno): April 10, 2013

60

Martin Kühmaier, PhD. e-mail: martin.kuehmaier@boku.ac.at Prof. Karl Stampfer, PhD. e-mail: karl.stampfer@boku.ac.at University of Natural Resources and Applied Life Sciences Vienna Department of Forest and Soil Sciences Institute of Forest Engineering Peter Jordan Straße 82 A-1190 Vienna AUSTRIA * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

Original scietific paper – Izvorni znanstveni rad

Computer Simulation of the GSM Signal Availability for Data Transmission on the Territory of the Forest Enterprise Tibor Žatko, Milan Koreň, Ján Tuček Abstract – Nacrtak The work deals with the creation of digital maps of the intensity of GSM/EDGE signal on the territory of the Forest Enterprise at the Technical University in Zvolen, Slovakia. Simulation of electromagnetic radiation intensity was calculated using the prediction program TEMS Parcell, based on technical data of transmitters and receivers, as well as attenuation due to the environment. Calculated values were verified in field measurements by means of testing apparatus placed in the measuring vehicle. The output vector layer contains areas categorized according to the suitability of signal intensity for field use with data or voice services.

Keywords: mobile GIS, data communications, GPRS, forestry

1. Introduction – Uvod With the development of various technologies and information systems, increasing need arises to work in the field with mobile devices for collecting data, interacting with digital maps, or for remote access to applications and databases. For interactive work with data sources, it is necessary to have a wireless connection with sufficient transmission capacity. A wide range of services in this area offers a choice between different systems as radio relay network, cell site, WiFi, Bluetooth, etc. A technology of terrestrial mobile radio communications network seems to be the best solution for wireless data transmission in the field. Its signal covers more than 90% of the Slovak Republic.

A great advantage of these systems is their high mobility and virtually universal reach, which is limited only by the availability of a signal. Cellular networks of national operators correspond to the Global System for Mobile Communications (GSM) standard, which was developed by the European Telecommunications Standards Institute (ETSI).

Coverage area and the services offered depend on the priorities set by the provider of telecommunication services. Most of operators prioritize areas with high population density, which follows logically from their commercial interests. Wireless technologies, supported by cellular radio communication networks, are Croat. j. for. eng. 34(2013)1

used outside the populated areas, in mountain areas with very difficult access, in areas with a complex geomorphologic structure as well as in remote forest areas. In order to define the areas qualifying for use of data or voice services, it is useful to have the digital map that enables to determine precisely the GSM/ EDGE signal intensity in the area of interest.

Any base transceiver station covers a limited area (cell), up to 37 km. High transmission capacity of the network is achieved by using multiple transmission frequencies. Mobile networks are constructed to have sufficient capacity for the needs of the areas they cover. Urban areas must comply with higher standards of quality and capacity than rural ones. Network capacity can be increased in two ways: using a larger number of channels per one cell, or by reducing the cell size. Cellular radio networks were designed to provide continuous communication services when moving the receiving station. When passing between the cells, mobile networks ensure the change automatically. The mobile station measures the level of signal of neighbouring base stations. Based on these measurements it decides which cell to switch it to.

According to the way of spreading, radio waves can be divided into direct, reflected, scattered, bent or with re fraction. In the spreading of radio waves in the given fre quency band, the following elements must

61

T. Žatko et al.

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

also be taken into account: reflection of waves from the Earth’s sur face, terrain obstacles, diffraction (i.e. bending waves) and refraction in the upper layer of the atmosphere. Tracks of radio waves are curved only on such ob stacles whose dimensions are smaller than the wavelength of the electromagnetic field. Bitirgan et al.

(2011) work ed out a model of radio signal spreading in the forest.

Performance of communication systems are affected by certain limiting factors that have a negative impact on signal transmission. Radio signals can be limited, for example, by attenuation, distance, noise, channel capacity and interference. On the territory with rugged morphology or in built-up area, the signal is affected by static objects (hills, vegetation, buildings) and movable obstacles (vehicles, trees branches movement). Leakage of the radio signal is divided into long-term (attenuation of the signal due to spreading, shading) and shortterm (extending multi-way and Doppler effect).

Leakages caused by spreading may be characterized as signal attenuation between the transmitter and receiver. In open area, the intensity of electromagnetic waves decreases with the square of the distance. If the signal between the transmitter and receiver spreads directly, the received signal level decreases relatively slowly. A common problem for mobile services is that mobile station is shadowed during the travel by various objects like buildings, trees or moving truck. This results in a rapid reduction of the level of the received signal. Decrease in signal level, relative to the basic course, may be up to several tens of decibels. Sarabande and Koh (2002) noted that the method of ray tracing is modelling quite precisely the intensity of radio signal even to larger distances from the transmitter. Based on the results, the radio signal attenuation depends on the surface of forest stand, and it increases with wavelength.

Short-term leakages cause sudden and sharp signal fluctuations. It occurs frequently in situations when the mobile station has no direct visibility of the base transceiver station.

The received signal is formed by reflections from the objects, but none of the ways of signal spreading is dominant. Each reflected wave is directed to the receiver on differently long roads. The result is that the received signals come to the receiver at different moments of time, with different amplitude and different phase.

This is the multi-way signal spreading that causes multi-way leakages of signal. Cavalcanti et al. (2002) and Tamir (1997) studied multi-way spreading of radio signals under specific conditions of forest environment.

Mobile services are characterized by a great dynamics in the network. Flexibility of movement with

62

the receiver is very high. Mobile station can be used when walking as well as in vehicles moving relatively quickly. In this case, it is necessary to take into account the signal leakage caused by the Doppler shift of wavelengths. Meng et al. (2009) examined in detail leakages of radio signal in forest environment. They described the physical process of spreading and analytical techniques of modelling radio signal attenuation in the forest. They performed experiments with the change of factors that affect the signal intensity.

GSM networks were built mainly for the provision of voice services. They are able to provide 9 kilobits/ sec data flow. Therefore, new GPRS (General Packet Radio Service) standards for data transmission over GSM network were developed. Operators had to make a technical innovation and modifications on their own networks to ensure an interconnection of mobile cellular network and public switched telephone network.

GPRS is built on packet switching, so it is possible to reach transmission rate of 14.4 kilobits/sec. Since GPRS network enables integration of several communication channels for one user, the resulting rate grows by the number of used channels. Usually 4-6 channels merge, according to operator’s settings and technical level of the mobile station. It is theoretically possible to achieve transmission rate of up to 86.4 kilobits/sec. However, it must be taken into account that IP-based applications consume a portion of resources for own purposes, and that final transmission rate depends on the quality and intensity of the radio signal. The actual rate of channel switch is, therefore, highly variable according to interference and signal leakage. The signal decreases to minimum in areas with low level of signal or high interference.

Another improvement of transmission rate was brought by the EDGE technology. EDGE was deployed in GSM network as an extension of the GPRS network.

This standard is usually called the network of 2.75 generation. At present EDGE is implemented in all networks of Slovak operators, and it is successfully used for data transmission, especially in rural areas, where the networks of the third generation do not reach. The structure is similar to GPRS, but EDGE increases the network bandwidth several times. Transmission rate achieved depends on several factors: the number of allocated time-slots, code scheme of network, user location, network capacity and signal quality. When using the coding scheme MCS-9, the transmission rate can theoretically reach 354 kilobits/sec. Other factors that must be taken into account are the network capacity and number of users communicating at the same time. If simultaneously more users use data services, the resources are shared between them. The actually Croat. j. for. eng. 34(2013)1

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

achieved high quality speed is about 260 kilobits/sec with a delay of about 200 ms.

The aim of this work was to create a digital map of the intensity of GSM/EDGE signal on the territory of forests belonging to the Forest Enterprise at the Technical University in Zvolen. Wireless computer network connection is needed for downloading real time corrections during precise measurements of position by global navigation satellite systems (GNSS), data transmission from monitoring stations, on-line access to databases and other research and practical activities in forest territory. The resulting maps will make possible to identify areas where services of mobile radio networks can be used, and on the other hand, to exclude parts where it is not appropriate to plan activities and measurements based on the GSM network support. Accuracy of prediction model of GSM signal quality was verified by field measurements of signal intensity and transmission rate.

2. Materials and methods – Materijal i metode We developed GSM signal intensity model, located in the orographic area Kremnické hills. The area is located in the central part of Slovakia and it is mainly formed of lava bodies originating from volcanic activities with prevalence of andesitic rocks. From the geomorphologic point of view, this area is strongly influenced by endogenous and exogenous factors as evidenced by the considerable variability of relief. Forests cover the majority of the study area. In the forests, broadleaved tree species (72.9%) prevail, of which beech accounts for 36.7%, oak for 19.3% and hornbeam for 9.7%. Coniferous trees cover 27.1%, of which spruce accounts for 11.3%, pine for 7.4% and fir for 6.7%. Digital maps of forest stands and inventory data were provided by the Institute of Forest Resources and Informatics, National Forest Centre in Zvolen.

Data from the company Orange Slovakia were the basic source for the elaboration of simulation. In creating a prediction of the level of GSM signal on the territory of the Forest Enterprise, simulation was based on the exact data of base stations located near the study area. For this purpose, the Orange Slovakia gave us consent for using the necessary data from the corporate database, which became the basis for the prediction model. There are six transmitters of the Orange Slovakia near the study area. Four are located on pylons in the cadastral areas of the near municipalities.

Two transmitters are located on high buildings. For the purposes of simulation, we input exact information about technological devices as they are actually set on Croat. j. for. eng. 34(2013)1

T. Žatko et al.

the transmitter. In addition to the exact location and altitude, it is necessary to specify the number of sectors (cells), height of the antenna above the ground, the type of antenna, its emission characteristics, orientation and tilting, transmitting capacity of device and frequency spectrum.

To verify the simulation results, we used the measured values of signal intensity using the measuring device in selected parts of the study area. For measurements, we used a vehicle to which a device was installed for scanning, monitoring and storing the measured values of GSM signal (Fig. 1). Two GNSS receivers were also a part of scanning for determining the current position. A GNSS receiver was directly connected to the scanning apparatus, so a geographical location was assigned automatically to each measured value of the GSM signal. The second GNSS receiver was connected to the computer with mobile geographical information system (GIS) and digital map of the territory of the Forest Enterprise for navigation and monitoring the planned route of measurement. According to the location of base stations, it could be expected that the northern and north-western part of the territory conditions would not be favourable for communication and use of mobile stations. We planned the route with respect to geographical coverage, time-consuming measurements, battery capacity in measuring device, the amount of measured data, and the expected attenuation of radio signals caused by the distance from the transmitter, shading by relief, and land cover.

We performed simulation of the coverage of the territory by GSM signal in the environment of TEMS Parcell programme. TEMS Parcell is a computer application designed to support planning of cellular radio networks. It enables to display a prediction layer of signal intensity and to combine it with other thematic layers with the aim to achieve the best representation of the result. Simulation accuracy depends mainly on the quality and completeness of input data and perfect tuning of prediction model. In the calculation, the model considered many environmental factors that contribute to the reduction of electromagnetic waves spreading. The calculation was based on the principles of electromagnetic waves spreading, terrain, obstacles and land use. A layer of land use defined categories for groups of objects on the surface such as vegetation, water areas, road network, urban areas differentiated according to built-up areas, etc. that participate in the attenuation of the signal spread, and must therefore be included to the calculation. The value of radio signal attenuation was assigned to each category of the land use layer. For forest stands, the value of the attenuation is around 20 dB, but the exact

63

T. Žatko et al.

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

Fig. 1 Measuring set and GIS in mobile measuring vehicle Slika 1. Uređaji za mjerenje i GIS u mobilnom vozilu za mjerenje value is determined in real measurements and finalization of the model, according to regional area and vegetation zones. Forest stands in land use categories can be divided to more detailed groups: broadleaved stands, coniferous stands and mixed stands. Thereafter, values of radio signal attenuation can be defined separately for each type of forest stand. It contributes to obtaining more exact results of the prediction of radio signal intensity.

Many radio signal spreading models are used to predict signal attenuation during channel transmission.

They differ from each other in the method of determining leakages and input parameters. All models take into account the distance between the transmitter and receiver, as a basic critical parameter. We can give as an example Lee’s model used in ground communication, which takes into account obstacles between the base station and mobile stations, or Okumura-Hata model for densely built-up areas. Objects like buildings, trees, hills, cause further signal attenuation. Radio waves in these cases do not spread to the receiver only in straight line, but migrate by other ways (Doboš et al. 2002).

We used a set of tools included in TEMS (Test Mobile System) from Ericsson for measurements of radio

64

signal quality and data processing. The package contains advanced applications for monitoring, evaluating and planning the radio access networks.

Scanning and monitoring of the radio network is a demanding process, in which a large amount of data is monitored at the same time. For this purpose, we used the testing mobile station TEMS Pocket. It is able to provide information on the quality and level of signal, errors in data transmission, the actual software setting of network as well as to monitor radio communication channel and to obtain information about surrounding cells, base stations, determine their distance and so on.

The application TEMS Investigation was used to collect and process data from field monitoring of the radio network. It is a tool for detecting and confirming the problems, optimisation, measuring and maintaining cellular radio networks. The set consisted of a scanning device, GNSS receiver, and laptop and software application. The testing mobile station TEMS Pocket and GNSS receiver were connected by cable to the computer and they provided application data on the exact position and monitored values. Scanned data were processed by the application TEMS Investigation. It enables to display the actual position of the measuring device on the map and graphical course of data on the communications between mobile and base station. At the same time, all obtained data were stored for further processing. From the results of measurement, we could determine unequivocally defects or problems of the network coverage, as well as intensity and quality of the radio signal.

The measurements were carried out in IDLE mode and DEDICATED mode simultaneously. The IDLE mode is the mode when there is no connection between the mobile station and base station. In this state, the mobile, if switched on and logged to the network, does not transmit. Practically, it only scans signals transmitted from the surrounding base stations, and based on the quality of the signal it makes reselection between the cells. DEDICATED mode is the state that occurs in the implementation of the connection. A phone call may serve as an example, when participants communicate in duplex mode. This means that it transmits and receives messages. For the measurement, we chose a call on testing number to maintain the station in DECICATED mode during the whole measurement. A software robot responds to testing number, which repeats in cycles a programmed message and does not end the connection spontaneously.

In contrast to the IDLE mode, in DEDICATED mode during the whole period of connection, the mobile station communicates with the base station and Croat. j. for. eng. 34(2013)1

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

constantly responds to changing radio conditions.

While driving, the changes occur very quickly and the signal intensity may change in jumps. For this case, both parties must respond flexibly and adapt the transmission power to the requirements of a situation.

Performance adaptation plays an important role in observing necessary qualitative parameters of the connection. GSM specifications define several classes of mobile stations according to their maximum transmission power. In order to achieve the least possible interchannel interference and maintain the performance, both the mobile station and base station try to communicate with the least possible energy consumption so as to maintain an acceptable signal quality. The signal level can move in increments of 2dB downwards or upwards. The mobile station and base station measure continuously the strength of the signal quality (bit errors) and adjust the power output. A function DTX (Discontinuous Transmission) is also used to reduce the transmission power. The mobile station does not transmit permanently (e.g. when the recipient of the call is silent), but only at certain time intervals necessary to maintain the connection. These reductions of power output have a positive effect on the battery life in mobile stations.

Due to these differences, both kinds of measurement were carried out. During the measurements, we monitored the scanned parameters of the signal and paid attention to areas, where the signal quality dropped below the level, when the mobile station was not able to communicate with the base station, and the connection in DEDICATED mode was interrupted. In such situations, we had to wait to get back to the field with the signal, and to be able to arrange a new connection.

On the maps of scanned data in TEMS Investigation, these parts in DEDICATED mode can be distinguished as sections with the absence of measurement points. It is particularly evident in areas with poor quality.

Six control measurements of transmission data rate were made based on the projected positions close to permanent research areas. Testing of transmission data rate took place in stationary position, when the measuring vehicle was on a designated point. After establishing the data connection, a file with well-defined size was downloaded from the test server. Based on the time consumed for downloading the test file to the computer, the average bit rate for data receiving (download) was calculated.

An actual position of the vehicle was monitored by means of GNSS receiver Garmin 60CSx with external antenna fixed on the vehicle roof. The positional accuracy of several meters was sufficient for our needs.

The device was connected to the laptop with installed Croat. j. for. eng. 34(2013)1

T. Žatko et al.

program ArcPad from ESRI, Inc. ArcPad supports vector and raster layers, creating points, lines and area elements, and also setting their attributes. This makes possible to carry out effectively the collection, updating and editing of geographical data in the field. The attribute data can be assigned to geometric features.

In practice, we checked the movement of the monitoring vehicle on the background map and recorded the set of points of the vehicle route.

During the final stage, we analyzed the created radio signal model and data collected from field measurements. For this purpose we used ArcGIS of the ESRI, Inc. The model of radio signal intensity as well as data collected by field measurements were imported to formats supported by ArcGIS system. Raster of the radio signal intensity was classified, and the boundaries of the categories were converted to vector data. The results of field measurements were divided into 25 log files, which were converted to the shape-file format. Finally, an overlay of thematic layers was made and output map compositions were created.

3. Results – Rezultati A model of radio signal intensity in a raster representation with a 100 m resolution was the output of TEMS Parcell simulation model (Fig. 2). Each cell was assigned the value of GSM signal intensity, calculated according to the algorithm of the prediction model.

The value was in units of dBm (decibel related to 1 miliwatt). Signal levels were divided into the categories used in radio telecommunication practice: Þ Indoor (less than or equal to –65 dBm): signal is sufficiently strong to penetrate buildings without windows and ensure a smooth connection. Þ Indoor window (–66 dBm to –82 dBm): signal level is sufficient for communication in buildings with windows. Þ In Car (–83 dBm to –89 dBm): communication in vehicles is ensured. Þ Outdoor: (–90 dBm to –92 dBm): signal level allows the participant to communicate outdoor. Þ Car kit (–93 dBm to –99 dBm): an external antenna for improved receiving needed. Þ No signal (greater than or equal to -100 dBm): communication is impossible.

English names of groups are symbolic. They indicate the scope and possibilities of the use of radio signals at a given level of intensity. Dividing into categories is only approximate, because each area or building has a different structure that causes differences in signal attenuation.

65

T. Žatko et al.

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

Thematic layers (borders of forest stands, road network, rivers and reservoirs, settlements, orthophoto) were added to the output maps. It increased their informative value, transparency and simplified the search. In the map, areas can be identified with available voice and data services of the company Orange Slovakia. The resulting layer of the radio signal intensity was overlapped with the boundaries of forest stands. The percentage of the level of radio signal was calculated for each unit of spatial forest arrangement and stored in the attribute table.

The rate of transmitted data is closely connected with the intensity and quality of the radio signal. Table 1 shows the transmission rate and intensity of the radio signal in the IDLE and DECICATED mode found on site. The best signal was found close to the urban areas, where good quality signal comes from nearby transmitters. An increased transmission rate was also achieved there. The results of measurements in the northern part of the Forest Enterprise demonstrate impaired level of the radio signal. The northernmost point of measurement was interesting. According to the map of coverage by radio signal, there should not be any signal. After careful examination of the data from the measured area, we found that this connection was achieved through a base station located at the distance of 20 km. This is also proved by low transmission rate caused by bad signal quality, although the intensity of signal reaches relatively good values. Stand opening and low age classes of tree species caused this discrepancy. This enabled signal penetration to the stand and improvement of its own intensity in a short section. It is also probable that for such small area, we were unable to predict the simulation of signal intensity correctly, as it was calculated with the resolution of 100 m.

Fig. 2 Map of GSM/EDGE signal availability for data communication (squares – measurement positions; circles – location of transmitters; light gray – forests with a low level of the GSM signal; dark gray – forests with a sufficient level of the GSM signal) Slika 2. Karta dostupnosti signala GSM/EDGE za podatkovni prijenos (kvadrati – lokacije izmjere; krugovi – lokacije odašiljača, svjetlosivo – šume s niskom pokrivenošću GSM-ovim signalom; tamnosivo – šume sa zadovoljavajućom pokrivenošću GSM-ovim signalom)

Table 1 Transmission rates and intensity of radio signal at static measurements Tablica 1. Brzina prijenosa i intenziteta radijskoga signala statičkim mjerenjem Signal intensity in IDLE mode, dBm

Signal intensity in DEDICATED mode, dBm

Brzina prijenosa, KB/s

Intenzitet signala u IDLE modu, dBm

Intenzitet signala u DEDICATED modu, dBm

Včelien 1

3.47

–100

–99

2

Včelien 2

3.68

–88

–94

3

Sviniarka

5.35

–76

–92

4

Žliabky

5.30

–85

–86

5

Jablonka

5.38

–94

–83

6

Budča

14.75

–72

–74

Number of measurement

Name of measurement

Transmission rate, KB/s

Mjerno mjesto

Ime mjernoga mjesta

1

66

Croat. j. for. eng. 34(2013)1

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

The resulting map of the coverage by GSM/EDGE signal provides an overview of the areas, where data transmission via the mobile communication network can be used. For example, in demarcation of forest ownership boundaries by GNSS devices and access to RTK, correction is inevitable. Current services provide users with three formats for correcting the accuracy of positional data on the Earth’s surface (RTCM 2.3, RTCM 3.0, RTCM CMR +). In the worst case, when the use of RTCM 2.3 is required, the transmission rate of about 600b will be needed. Measurement results confirm that even at relatively low quality or intensity of signal, mobile devices have no problem to transmit this amount of data. Coverage by the mobile communication services is a decisive factor. To transmit the required volume of data in GPRS / EDGE network, a code scheme (MCS-1) is sufficient, where minimal transmission rate reaches 8.8 kilobits/sec.

The output of TEMS Parcell is a grid of radio signal level with the resolution of 100 meters. We calculated the arithmetic mean and standard deviation of signal intensity measurements for corresponding grid cells.

The arithmetic means were converted into six categories used in radio telecommunications practice. During measurement in IDLE mode of the total number of 5.719 measured points, 2.903 were included to the cells of predicted radio signal level values (Table 2). In DEDICATED mode, 6.882 points of the total number of 11.743 measured points coincided with the grid cells of simulated signal level values (Table 3). The overall accuracy of prediction of radio intensity levels in IDLE mode was 53% and the overall accuracy of prediction of radio intensity levels in DEDICATED mode was 36%.

Discrepancies in the simulation in DEDICATED mode were somewhat greater than in IDLE mode.

They were caused by prediction model adjustment mainly to IDLE mode. Another reason was that during the communication in DEDICATED mode between mobile and base station, an active control of power output was made, which modified the intensity of the transmitted signal on both sides and adapted the signal to surrounding conditions.

We found that in marginal areas, where the signal intensity of selected transmitters dropped to the minimum (especially in the northern part of the Forest Enterprise), testing mobile stations received signals from base stations being distant also several kilometers from the Forest Enterprise, which were not considered in the prediction. This signal spread from high-placed transmitters. However, the signal quality was very low and unstable. Based on the measurement results and data provided by the Company Orange Slovakia, the results of the simulation were corrected in marginal zones and Croat. j. for. eng. 34(2013)1

T. Žatko et al.

Table 2 Predicted and measured signal intensity levels in IDLE mode (Count – number of measurements; Avg – arithmetic mean; Std – standard deviation) Tablica 2. Predviđena i izmjerena razina intenziteta signala u IDLE modu (zbroj – broj mjerenja; X – aritmetička sredina; SD – standardna devijacija) Cell number

Predicted level

Measured signal intensity Intenzitet izmjerenoga signala

Broj stanice

Predviđena Count razina Zbroj

Min

Max

Avg

Std

Level

Min

Maks

X

SD

Razina

3

1

395

–88

–52

–66.7

6.1

2

4

1

152

–109

–57

–71.5

8.6

2

5

1

73

–72

–55

–64.6

5.0

1

9

1

252

–80

–48

–61.3

9.6

1

12

1

92

–92

–45

–61.5

14.7

1

18

2

33

–83

–77

–79.1

2.0

2

19

2

55

–100

–55

–76.0

14.1

2

20

2

210

–92

–56

–72.1

6.9

2

21

2

29

–93

–83

–87.2

3.5

3

22

2

82

–99

–70

–80.8

8.1

2

23

2

9

–88

–77

–79.2

3.5

2

24

2

51

–97

–75

–82.8

7.6

2

25

2

97

–95

–52

–63.2

13.1

1

29

2

10

–95

–89

–92.0

2.2

4

32

3

18

–100

–91

–94.4

3.0

5

35

3

20

–89

–83

–85.2

2.4

3

36

3

145

–102

–79

–91.8

4.7

4

37

3

20

–98

–57

–89.4

9.1

3

40

3

54

–101

–84

–90.4

5.0

4

44

3

265

–92

–73

–82.0

3.8

2

47

4

18

–96

–90

–93.7

2.2

5

49

4

2

–99

–98

–98.5

0.7

5

52

4

17

–101

–91

–97.0

3.7

5

56

4

25

–98

–78

–90.9

4.5

4

58

4

609

–106

–71

–84.9

8.8

3

64

5

3

–87

–87

–87.0

0.0

3

66

5

10

–108

–89

–97.6

4.8

5

69

5

16

–101

–89

–97.6

3.7

5

71

5

133

–105

–73

–84.5

7.7

3

74

5

8

–99

–99

–99.0

0.0

5

0

6

2816

–110

–72

–93.9

7.5

5

67

T. Žatko et al.

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

Table 3 Predicted and measured signal intensity levels in DEDICATED mode (Count – number of measurements; Avg – arithmetic mean; Std – standard deviation) Tablica 3. Predviđena i izmjerena razina intenziteta signala u DEDICATED modu (zbroj – broj mjerenja; X – aritmetička sredina; SD – standardna devijacija) Measured signal intensity

Cell Predicted number level

Intenzitet izmjerenoga signala

Broj Predviđena Count stanice razina Zbroj

68

Min

Max

Avg

Std

Level

Min

Maks

X

SD

Razina

3

1

613

–88

–59

–72.2

6.4

2

4

1

306

–95

–58

–74.5

5.7

2

5

1

234

–89

–59

–74.1

5.7

2

9

1

701

–91

–57

–74.7

4.8

2

12

1

284

–95

–60

–76.9

7.7

2

18

2

113

–92

–70

–81.5

5.0

2

19

2

115

–96

–56

–77.3

10.2

2

20

2

370

–98

–53

–76.4

9.0

2

21

2

66

–100

–82

–91.2

4.0

4

22

2

163

–93

–68

–82.0

6.1

2

23

2

12

–93

–77

–86.2

5.2

3

24

2

140

–102

–77

–87.6

5.0

3

25

2

329

–100

–64

–76.3

6.7

2

29

2

13

–102

–99 –100.9

1.4

6

32

3

53

–96

–90

–94.4

0.9

5

35

3

76

–100

–81

–91.5

4.3

4

36

3

296

–101

–79

–92.5

4.1

4

37

3

53

–106

–73

–95.5

9.6

5

40

3

54

–93

–86

–89.9

2.2

3

42

3

9

–101

–97 –100.3

1.4

6

44

3

866

–106

–75

–89.4

5.7

3

47

4

44

–104

–89

–97.7

4.4

5

52

4

31

–102

–95

–96.9

1.5

5

56

4

85

–106

–71

–91.1

9.6

4

58

4

1390

–106

–75

–92.8

6.0

4

64

5

7

–92

–86

–88.7

2.6

3

65

5

2

–91

–90

–90.5

0.7

4

66

5

14

–106

–106 –106.0

0.0

6

69

5

33

–102

–91

–97.2

3.8

5

71

5

410

–106

–77

–93.9

6.1

5

0

6

4861

–110

–74

–96.9

6.4

5

a new map layer illustrating the coverage by GSM signal was created for the given territory.

With the help of visual interpretation of differences displayed in the map, we found that most differences occurred in three types of regions: Þ In urban areas, where the signal changes depend on the density and height of buildings. Þ In areas with big variability of stand heights, caused by regeneration, and mainly in forest stands of lower age classes that do not reach the height of mature stands as well as on forest roads going from the south to north. This phenomenon can be attributed to the fact that the simulations calculated for the sectors of base stations, which broadcast signals, were directed to the north and also included attenuation of forest region as a whole, not taking into account gaps in the forest stands formed by roads. Þ Radio signal intensity could be affected by humidity, as it was raining during the measurement. Similarly, seasonality might have had some effect, because the measurements were carried out in non-vegetation period. A study assessing the impact of environment humidity on the intensity of GSM signal confirms the conclusions (Helhel et al. 2008).

The data obtained confirmed the reliability of the prediction model. The quality of the model was limited by the scope and accuracy of input data. Relatively high general character of the output layer of the signal intensity was caused by two main factors. First, it was incorrect raster layer of digital elevation model with the resolution of 100m and not properly elaborated layer of landscape cover with a small number of categories, when only one category defines the forest areas without more detailed specification of stands. With the aim of making the calculation more accurate in forested regions, the structure of forest stands should be described more consistently, mainly including tree species composition, as well as size and age structure of forests.

Considering that broadleaved tree species prevail, the simulation should also consider defoliation and the actual humidity that contribute to the variability of results. Contemporary prediction models do not work with such detailed input parameters. We cannot incorporate the exact short-term leakages to the calculations, as they have stochastic character, and therefore cause quick fluctuation of the signal level in short sections.

4. Conclusions – Zaključci Nowadays, the networks of the third generation (UMTS - Universal Mobile Telecommunications SysCroat. j. for. eng. 34(2013)1

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

tem) are preferred for data transmission via mobile radio networks. With using the protocol HSPA (High Speed Packet Access), transmission rate in these networks reaches several megabits. A main factor limiting their use in the field work is that the operators concentrate these networks in urban areas.

In comparison with GSM networks, they only cover a minimal part of forested areas. GSM network operates in lower frequency bands and spreads better in the environment. For the purposes of mobile communication in forestry practice and research, the GMS network is still unanimously the most available source.

Due to the above reasons, the created layer of GSM/ EDGE signal intensity is good material for using in forest practice and for obtaining information on the availability of data transmission services in the area of interest.

Prospective areas of mobile communication in forestry are, for example, an access to RTK corrections in high precision measurements of locations by GNSS equipment, data transmission from measuring stations to research plots, online access to databases and applications. Information on the availability and quality of the radio signal is an important factor for effective planning and introducing of wireless data transmission technologies to forest research and practice.

Acknowledgements – Zahvala We would like to thank Orange Slovakia Company for providing documents and renting the devices and software used in the implementation of our work. Our thanks go also to the Institute of Forest Resources and Informatics, National Forest Centre Zvolen for data providing.

T. Žatko et al.

We carried out the work within the research project of the Scientific Grant Agency of the Ministry of Education and Slovak Academy of Sciences 1/0764/10 (VEGA) »Research principles and methods for precision forestry«.

5. References – Literatura Bitirgan, M., Yoruk, Y. E., Celik, S., Kurnaz, O., Helhel, S., Ozen, S., 2011: Generation of An Empirical Propagation Model for Forest Environment at GSM900/GSM1800/ CDMA2100. Proceedings of the XXX General Assembly and Scientific Symposium of the International Union of Radio Science, August 13–20, 2011, Istanbul, Turkey. Available at: Ursigass 2011.org/abstracts/ursi/F02-9.pdf.

Cavalcanti, G. P. S., Sanches, M. A. R., Oliveira, R. A. N., 1999: Mobile radio propagation along mixed paths in forest environment. Journal of Microwaves and Optoelectronics 1(4): 42–52.

Doboš, Ľ., Rainbow, J., Marchevský, S., Wieser, V., 2002: »Mobile radio network« published by EDIS Žilina, ISBN 80-7100936-9, 312 p.

Helhel, S., Ozen, S., Goksu, H., 2008: Investigation GSM signal variation of dry and wet climatic effects. Progress in Electromagnetic Research B 1: 147–157.

Meng, Y. S., Lee, Y. H., Ng, B. C., 2009: Study of propagation loss prediction in forest environment. Progress in Electromagnetic Research B, 17: 117–133.

Sarabande, K., Koh, I. S., 2002: Effect of canopy-air interface roughness on HF-VHF wave propagation in forest. IEEE Trans. Antennas Propag. 50(2): 111–121.

Tamir, T., 1977: Radio wave propagation along mixed paths in forest environments. IEEE Trans. Antennas Propag. AP25(4): 471–477.

Sažetak

Računalna simulacija dostupnosti GSM-ova signala za prijenos podataka na području šumarskoga poduzeća Cilj rada i istraživanja provedenoga na Tehničkom sveučilištu u Zvolenu bio je stvoriti digitalnu kartu zastupljenosti GSM-ova odnosno EDGE-ova signala na području kojim gospodari šumarsko poduzeće. Točnost modela za predviđanje kakvoće GSM-ova signala potvrđena je i terenskim mjerenjima intenziteta signala te brzine prijenosa podataka. Simulacija se temeljila na egzaktnim podacima baznih stanica smještenih u blizini istraživanoga područja, a stanice su dobivene od nacionalnoga mobilnoga operatera »Orange Slovačka«. Baza podataka uključuje točnu lokaciju i nadmorsku visinu odašiljača, broj sektora (stanica), visinu antene iznad tla, vrstu antene, njezine emisijske karakteristike, orijentaciju i ekspoziciju, kapacitet odašiljanja uređaja i frekvencijski spektar.

Za provjeru rezultata simulacije upotrijebljene su stvarne izmjerene vrijednosti intenziteta signala pomoću mjernoga uređaja u odabranim dijelovima istraživanoga područja. Pri mjerenju je korišteno vozilo kojemu je ugrađen uređaj za skeniranje, praćenje i spremanje stvarne izmjerene vrijednosti GSM-ova signala (slika 1). Mjerenja su provedena u IDLE modu i DEDICATED modu istodobno. Korišten je skup alata uključenih u TEMS (testni moCroat. j. for. eng. 34(2013)1

69

T. Žatko et al.

Computer Simulation of the GSM Signal Availability for Data Transmission... (61–70)

bilni sustav) iz Ericssona za mjerenje kakvoće radijskoga signala i prijenosa podataka. Paket sadrži napredne aplikacije za praćenje, vrednovanje i planiranje pristupnosti radijskoj mreži.

U radu je napravljena simulacija pokrivenosti teritorija GSM-ova signala u okruženju programa TEMS Parcell.

TEMS Parcell je računalna aplikacija dizajnirana kako bi podržavala planiranje mobilnih radijskih mreža. Program omogućuje da se kroz slojni prikaz predviđa intenzitet signala te da se kombinira s drugim tematskim slojevima, sve radi postizanja najboljih prikaza rezultata. Rezultat TEMS Parcell simulacijskoga modela (slika 2) jest model intenziteta radijskoga signala u obliku rasterske zastupljenosti s rezolucijom 100 m. Svakoj je stanici dodijeljena vrijednost intenziteta GSM-ova signala u jedinicama dBm (decibel u odnosu na jedan milivat), izračunat sukladno algoritmu za model predviđanja. Razine su signala podijeljene u kategorije koje se koriste u radiotelekomunikacijskoj praksi.

Tablica 1 u radu prikazuje brzinu prijenosa i intenzitet radijskoga signala u IDLE i DECICATED modu, izmjerenoga statičkim mjerenjem na lokalitetu. Tijekom mjerenja u IDLE modu od ukupnoga broja 5719 mjerenih točaka 2903 uključene su u stanice predviđanja razine vrijednosti radijskih signala (tablica 2). U DECICATED modu 6882 točke od ukupnoga broja izmjerenih točaka (11 743) poklopile su se s mrežnim stanicama simulirane razine vrijednosti signala (tablica 3). Za odgovarajuće mrežne stanice izračunate su aritmetičke sredine i standardne devijacije intenziteta signala mjerenja. Ukupna točnost predviđanja razine radijskoga intenziteta u IDLE modu iznosi 53 %, a ukupna je točnost predviđanja razine radijskoga intenziteta u DECICATED modu 36 %. Nepodudarnost simulacije u DECICATED modu nešto je veća nego u IDLE modu. One su uglavnom uzrokovane prognostičkim prilagodbama modela u IDLE modu. Drugi razlog obuhvaća tijek komunikacije u DECICATED modu između mobilne i bazne stanice, gdje je napravljena aktivna kontrola snage izlaznih podataka, čime je modificiran intenzitet emitiranoga signala na obje strane te prilagodba signala okolnim uvjetima.

Istraživanjem je utvrđeno da je najbolji signal u blizini urbanih područja u kojima signal visoke kakvoće dolazi iz obližnjih odašiljača, što se istodobno odražava na povećanu brzinu prijenosa. Rezultati mjerenja u sjevernom području kojim gospodari šumarsko poduzeće pokazali su oslabljenu razinu radijskoga signala. U rubnim područjima istraživanja mobilna je stanica primala signale iz baznih stanica udaljenih nekoliko kilometara od šumarskoga poduzeća.

Kakvoća dobivenoga modela intenziteta radijskoga signala ograničena je opsegom i točnošću ulaznih podataka.

Dva su glavna ograničavajuća čimbenika: niska prostorna rezolucija digitalnoga modela reljefa i netočan opis sastojinske strukture šuma. S obzirom na prevladavajuću zastupljenost listopadnoga drveća simulacija treba uzeti u obzir osutost i stvarnu vlažnost jer one pridonose varijabilnosti rezultata. Suvremeni modeli predviđanja ne uzimaju u obzir takve detaljne ulazne parametre.

Nastale digitalne karte omogućit će utvrđivanje područja u kojima se usluge mobilnih radijskih mreža ne mogu koristiti. Informacije o dostupnosti i kakvoći radijskoga signala važan su čimbenik pri učinkovitom planiranju i uvođenju tehnologije bežičnoga prijenosa podataka u znanstvenim istraživanjima u šumarstvu, ali i u šumskoj praksi. Moguća su područja primjene mobilne komunikacije u šumarstvu, na primjer, pristup RTK korekturi visoke preciznosti izmjere lokacija pomoću GNSS-ove opreme, prijenos podataka iz mjernih postaja na istraživačke plohe, on-line pristup bazama podataka te raznim aplikacijama i sl.

Ključne riječi: mobilni GIS, podatkovna komunikacija, GPRS, šumarstvo

Authors’ address – Adresa autorâ: Tibor Žatko, BSc. e-mail: tibor.zatko@orange.sk Orange Slovakia Company Jegorovova 22, 974 01 Banská Bystrica

Received (Primljeno): May 18, 2012 Accepted (Prihvaćeno): January 10, 2013

70

Milan Koreň, PhD.* e-mail: milan.koren@tuzvo.sk Prof. Ján Tuček, PhD. e-mail: jan.tucek@tuzvo.sk Department of Forest Management and Geodesy Forestry Faculty, Technical University in Zvolen T. G. Masaryka 24, 960 53 Zvolen SLOVAKIA * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

Original scietific paper – Izvorni znanstveni rad

Estimation of Trafficable Grades from Traction Performance of a Forwarder Jörg Hittenbeck Abstract – Nacrtak Wood as a renewable resource is getting more and more popular for material use as well as for energy usage. In order to meet the demands, it is important to make previously unused wood potentials accessible to the timber market. One area of interest is the incompletely utilized resources in inclined regions. Problems arise in these areas by the topographic limitations of highly mechanized timber harvesting. These limitations occur from the stability of the machines as well as from the damages done to the soils.

Considerations about the downhill slope forces acting on the machines show a direct relation between the inclination of a grade and the traction coefficient. In theory, it seems possible to calculate a trafficable grade for an accepted level of wheel slip. On the basis of a 25% slip limitation, a model for trafficable grades was developed and tested for typical hillside conditions.

Measurements of traction force vs. slip identified the soil water content and the skeletal rate as the main soil parameters that affect the climbing ability of machines. Test drives in inclined terrain indicated that the approach of calculating limitations (for trafficable grades) from traction tests under level conditions lead to a fairly good prediction.

Keywords: full mechanized harvesting, slope, loess, slip resistance, model

1. Introduction – Uvod In times of a growing demand for renewable energy, forestry in Central Europe copes with an increasing demand for raw timber. On the one hand there are paper and sawmills that utilize timber, whereas on the other hand the request for biomass is increasing rapidly. This leads to an increasing demand particularly for low value timber. Market prices change and first thinnings even under unfavorable conditions generate a positive profit margin. Therefore, a highly mechanized harvesting in smaller stands is requested. The application of harvester and forwarder is beyond dispute under lowland conditions, which results in strong activities in timber mobilization from these areas. In low mountain range the discussion about the permanent conservation of trafficability sets limits to fully mechanized harvesting. The existing certification labels like FSC (Forest Stewardship Council) and PEFC (Program for the Endorsement of Forest Certification) require an appropriate machine usage so that the foresters are well advised to look for valid limitations.

Croat. j. for. eng. 34(2013)1

Trafficability of soils is endangered by the driving activities with forest machinery. Damages to soil result from soil compaction and erosion processes. Compaction is especially well investigated for agricultural soils (Blackwell et al. 1986; Bailey et al. 1996; Trautner and Arvidsson 2003, Arvidsson and Keller 2007; Ansorge and Godwin 2007, 2008; Horn and Fleige 2009) as well as for forest soils (Wästerlund 1983; Moffat 1991;Hutchings et al. 2002; Horn et al. 2004; Jun et al. 2004; von Wilpert and Schäfer 2006; Horn et al. 2007). Erosion on the other hand is often seen in conjunction with tillage operations and the opportunities through changing to conservation tillage or to a total abandonment of tillage (Tebrugge and During 1999). Besides tillage, traffic on soils leads to increased risk of erosion (Raper 2005).

Podsiadlowski (1988) and Mosimann et al. (2008) focused on the tracks of machines and detected that erosion increased. Bazoffi et al. (1998) showed that reduced tire inflation pressure and therefore better traction of the machines leads to reduced erosion on slopes. In low mountain ranges the problem of soil erosion is enforced by higher rain falls and needs for

71

Jörg Hittenbeck

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

traction to assure mobility of forest machines. Therefore, soil erosion caused by slip of the wheels is the main risk to a permanent trafficability in inclined terrain.

Results from Söhne (1952) indicate that the risk of serious harm done to the soil increases with increasing slip. High wheel spin leads to a complete shearing-off of the topmost soil structures. Loose soil material gets washed away with the next intense rain. Even for countries with lower requirements to soil protection the long term conservation of technical trafficability as well as the weight of soil as a fundamental factor of production are important. Objective limits for harvesting operations with land-based machinery are important for reducing the additional strain by higher wheel slip. Söhne (1952) found that especially slip levels up to 25% assure a minimum of continuity of the soil pores whereas higher slip leads to a shearing-off of the topmost soil. Particularly, grades are endangered by erosion because of these lost connections between the soil layers. Therefore, slip has to be limited when driving in grades in order to assure long term trafficability (ecological and technical).

The aim of this paper is to present a method to calculate maximum inclination angles for an accepted wheel slip level of 25% as this assures a minimum of continuity of the soil pores (Söhne 1952). Considerations about the downhill slope forces acting on the machines show a direct correlation between the inclination of a grade and the traction coefficient. The latter describes the relation between traction force and machine weight. Measurements of traction force in dependency of slip were taken for a forwarder under different soil conditions in order to oppose them to the downhill slope force at different inclination angles. In theory, it seems possible to calculate a trafficable grade for an accepted level of wheel slip. On the basis of a 25% slip limitation, a model for trafficable grades was developed and tested for typical low mountain range conditions in Germany. In addition, estimation for maximum inclinations with the meaning of safety against sliding down was calculated.

2. Theoretical Approach – Teorijski pristup Every self propelled machine has to produce a traction force (or »thrust«) which overpowers the motion resistance forces. The latter result from aerodynamic resistance (Ra), resistances due to the internal running gear (Rin) and resistance due to the interaction between drive line and terrain (Rt), commonly known as rolling resistance (Wong 2010). When driving on slopes, there is an additional grade resistance (Rg), which results

72

Fig. 1 Forces acting on machines while driving on slopes (Source: Jacke and Drewes 2004) Slika 1. Utjecaj sila na vozila na nagibu (Izvor: Jacke i Drewes 2004) from the downhill slope force (FT (N) in Fig. 1) acting on the machines.

Fig. 1 illustrates the forces appearing when driving on grades. The downhill slope force (FT) is derived from the product of weight of the loaded vehicle (G – gravity force (N)) and the sine of the inclination angle following the equation: FT = G × sina

(1)

In order to assure the mobility of a machine, a force at least equal to the downhill slope force has to act in the opposite direction. The force left when the motion resistances (at low speeds this is just the rolling resistance) are overpowered is the traction force (or net traction in Zoz and Grisso 2003). It can be calculated from the traction coefficient (μtr also called »net traction ratio«) multiplied with the normal force (FN (N)) acting between the vehicle and terrain.

Fd = FN × mtr

(2)

Under flat conditions, the normal force (FN) is equal to the weight (gravity force) of the machine. However, with increasing inclination angle of the slope, the forces differ substantially. The traction coefficient varies with soil properties, drive line of the vehicle and wheel slip. When soil and drive lines are defined, the traction coefficient changes depending on the slip. Typical progressions of the coefficient over slip were studied by numerous authors (Söhne 1952; Wanjii et al. 1997; Yoshida and Hamano 2002; Zoz and Grisso 2003; Hittenbeck 2009; Wong 2010).

G × sina = FN × mtr

(3)

The comparison between downhill slope force and drawbar pull in equation 3 reveals that the forces for the calculation differ. In order to solve the equation, the Croat. j. for. eng. 34(2013)1

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

normal force of the machine (FN) has to be expressed as a function of weight (G) and the inclination angle.

FN = G × cosa

(4)

Inserted into equation (3) this leads to: G × sina = G × (cosa) × mtr

Jörg Hittenbeck

optimum between a smooth increase of the required traction force and the distance travelled with different rates. 200 m rope and a pulley mounted to the load cell allow between three to seven traction measurements on a single pullout.

(5)

Equation 5 simplifies to: mtr = tana

(6)

From this point, the traction coefficient (μtr) leads to the slope which is barely accessible. The tangent value of the inclination can easily be transformed to slope inclination given in percent by multiplying with 100. As the traction coefficient is closely related to the wheel slip, it is possible to predict the occurring slip values. The general relation between the traction coefficient and the inclination trafficable is not new and it has already been shown by several authors (Pampel 1982; Schulz 1988; Kunze et al. 2002; Hoepke and Appel 2002).

3. Methods and Results – Metode i rezultati 3.1 Traction Measurements – Mjerenje vučnih značajki The determination of coefficients of traction was done by traction force vs. slip measurements carried out with a forwarder (Ponsse, model S10) under level conditions. This was done for different variants, which result from different setups of the final drive (worn tires, new tires, reduced tire pressure, combination of tracks and chains) as well as different soil types and varying soil moisture content. In addition to measurements with an empty forwarder, a few variants were tested for a loaded machine as well.

A special winch was constructed to apply different loads to the forwarder for the traction tests. The winch is linked to a breaking system from a heavy truck. Both were mounted to a platform in order to be able to install the winch on skidding lanes. Fig. 2 shows the winch fastened with ground anchors and tied to a tree.

The deceleration of the forwarder is controlled by a feed forward control. The forces acting on the brake disc are increased until the forwarder is not able to pull out the rope any more. Afterwards the pressure on the brake is reduced until the machine is able to drive without interference. Both for the pressure increase as for the decrease, a rate of 0.15 bar/s was chosen. This slew rate was the result of several pretests to get the Croat. j. for. eng. 34(2013)1

Fig. 2 Winch to apply the loads for traction measurements installed on a skidding lane Slika 2. Vitlo pomoću kojega se opterećuje stroj prilikom mjerenja trakcije During the measurements, the traction force of the forwarder as well as the speed of wheels and above ground were measured on the forwarder. Forces result from a load cell (Hottinger Baldwin Measurements (HBM), model U2B). The speed of the wheels and the effective velocity were determined using incremental rotary encoders (Kübler, model 5800). A modularly equipped measurement system MGCsplit by HBM served as data collector for all applied transducers.

The tests were driven on skidding lanes under different site conditions, which are characterized by high percentage of loess. Apart from soil moisture, the main difference between the site types is the stone admixture. The tests were carried out on bare soil because of the ambiguous effect of a brushwood layer on vehicle traction (Hittenbeck 2004; Jacke et al. 2004). Before the test drives started, all soil parameters like soil type, soil moisture, humus layer and inclination of the stand were determined. In total 57 different test series were taken with 22 different possible factors (Table 1) influencing the traction behavior of the forwarder. Every test run is characterized by the machine configuration (e.g. tracks and chains, unloaded) and stable soil and site conditions. For each of these series, the 22 factors concerning soil and machine parameters were determined and coded into numbers.

73

Jörg Hittenbeck

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

Table 1 22 factors considered before the test that can have an impact on traction performance Tablica 1. 22 čimbenika koji su promatrani prije mjerenja, a za koje se smatra da imaju utjecaja na vučne značajke Machine parameters

Site conditions

Značajke vozila

Uvjeti radilišta

Tires

Site

Water content

Gume

Radilište

Vlažnost tla

Inflation pressure

Slope

Soil type

Tlak u gumama

Nagib

Tip tla

Wheel chains

Cross slope

Skeletal admixture

Lanci

Poprečni nagib

Udio skeleta

Soil parameters Značajke tla

Bogie tracks

Tree species

Humus occurrence

Polugusjenice

Vrsta drva

Pojavnost humusa

Load

Skid trail

Humus type

Teret

Vlaka

Vrsta humusa Shear strength Posmična čvrstoća Sand content

Fig. 3 Traction coefficients over slip measured for the forwarder equipped with tracks and chains at 32.1% soil moisture content on loess soil Slika 3. Odnos faktora neto i klizanja forvardera opremljenoga polugusjenicama i lancima na prapornom tlu s 32,1 % vlage

Udio pijeska Silt content Udio praha Clay content Udio gline Bulk density (skid trail) Gustoća tla (u kolotrazima) Bulk density (between trails) Gustoća tla (između kolotraga) Bulk density (stand, untouched) Gustoća tla (negaženo tlo sastojine)

3.2 Traction – Trakcija The scatter plot in Fig. 3 is an example of the resulting coefficients of traction dependent on wheel slip. It displays the results of traction measurements with tracks and chains (chains were mounted on the front axle and tracks on the rear axle) on loess soil at a soil moisture content of 32.1%. The illustrated data are already revised by artifacts which for example result when the brake of the winch is released and the machine starts driving. In this situation higher slip values can occur although the force to pull the rope is rather low. More detailed information about the data pro-

74

cessing can be found in Hittenbeck (2009). With increasing wheel slip, the values of the traction coefficient rise immediately up to a level of about 30% slip.

Above that, the increase of the traction coefficients is still clear but less rapid. No maximum value has been specified for the coefficients, although it is often assumed for off road conditions (Moore 1975).

The scatter plot in Fig. 3 is typical for measurements at good traction conditions. The latter are characterized by dry soil conditions or tests with traction aids (tracks and/or chains). Beside this typical (for the conducted tests) shape of the scatter plot, there are other two different types. At higher soil moisture contents and without traction aids, the traction coefficients often culminated between 50% and 80% slip, whereas for poor traction conditions (wet soil, no traction aids, low soil strength) the scatter plots show a smooth increase of traction coefficients over slip which result in comparatively low maximum values.

3.3 Prognosis of Grade Ability – Procjena razreda kretnosti po nagibu For a given slip level (for example the 25% from Söhne /1952/), a wide range of coefficients of traction were measured (Fig. 3). Therefore, it is not plausible to identify an explicit limit resulting from the measureCroat. j. for. eng. 34(2013)1

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

Jörg Hittenbeck

Table 2 Regression coefficients for determination of soil ecological inclination limits Tablica 2. Regresijski koeficijenti za određivanje ekološke granice nagib Variable

Regression coefficient

Standard Error

p-value

Varijable

Regresijski koeficijent

Standardna pogreška

p-vrijednost

29.753

1.416

0.000

–0.831

0.123

0.000

6.185

1.115

0.000

7.579

2.663

0.007

6.035

3.283

0.072

Constant – Konstanta Soil water content (–30%) Vlažnost tla (–30 %) Skeletal rate (0; 1; 2) Kamenitost (0; 1; 2) Tracks (Dummy variable) Polugusjenice (dummy varijabla) Chains (Dummy variable) Lanci (dummy varijabla) Dummy variable: 1 – mounted; 0 – dismounted Dummy varijable: 1 – postavljene; 0 – nepostavljene

Table 3 Regression coefficients for determination of maximum inclination limit Tablica 3. Regresijski koeficijenti za određivanje maksimalnoga nagiba Variable

Regression coefficient

Standard Error

p-value

Varijable

Regresijski koeficijent

Standardna pogreška

p-vrijednost

47.530

1.180

0.000

–0.583

0.103

0.000

4.766

0.929

0.000

9.790

2.219

0.000

9.438

2.735

0.001

Constant – Konstanta Soil water content (–30%) Vlažnost tla (–30 %) Skeletal rate (0; 1; 2) Kamenitost (0; 1; 2) Tracks (Dummy variable) Polugusjenice (dummy varijabla) Chains (Dummy variable) Lanci (dummy varijabla) Dummy variable: 1 – mounted; 0 – dismounted Dummy varijable: 1 – postavljene; 0 – nepostavljene

ments. This can be resolved by adjusting a suitable regression model for the progression of the coefficients of traction above wheel slip. As there were different outlines of the scatter plots due to different traction conditions, which sometimes showed a maximum but most often missed one, a cubic regression model was applied. For the resulting 57 models, the traction coefficient (and therefore the resulting inclination through multiplication with 100) at 25% slip was calculated.

These values served as command variable for a stepwise linear regression with the possible influence factors on traction behavior. The test conditions are Croat. j. for. eng. 34(2013)1

coded by 22 different variables that describe the characteristics of the test stand, machine parameters and soil conditions. Table 1 gives a list of the variables related to the three main categories. From these factors (22 in total), soil water content and skeletal admixtures (grouped into 3 subclasses) as well as the application of traction aids (tracks and/ or chains) were identified by a stepwise linear regression as main parameters to traction behavior. The subclasses to skeletal admixtures in the topmost 20 cm were: free of skeleton (group 1), up to 7% skeletal admixture (group 2) and admixtures above 7% (group 3).

75

Jörg Hittenbeck

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

Table 4 Inclination limits resulting from regression models for loess soils Tablica 4. Granični nagib regresijskoga modela za praporna tla Soil-Water-Content (skeletal rate) Vlažnost tla (kamenitost)

Wheels – Kotači

Tracks and Chains – Polugusjenice i lanci

Inclination limit (ecological)

Inclination limit (max.)

Inclination limit (ecological)

Inclination limit (max.)

Granični nagib (ekološki)

Granični nagib (maksimalni)

Granični nagib (ekološki)

Granični nagib (maksimalni)

% 25 (0)

34

50

48

70

30 (0)

30

48

43

67

35 (0)

26

45

39

64

40 (0)

21

42

35

61

45 (0)

17

39

31

58

%

25 (<7)

40

55

54

74

30 (<7)

36

52

50

72

35 (<7)

32

49

45

69

40 (<7)

28

46

41

66

45 (<7)

23

44

37

63

%

25 (>7)

46

60

60

79

30 (>7)

42

57

56

76

35 (>7)

38

54

52

73

40 (>7)

34

51

47

71

45 (>7)

30

48

43

68

The resulting regression model is presented in Table 2. The constant describes the theoretically accessible grade (limit) with 25% wheel slip at a soil water content of 30% for the forwarder without tracks and chains. The expected negative influence of soil moisture on traction behavior and therefore the ability to climb slopes is indicated by the regression coefficient for the soil water content. An increase of soil water content of 1 Vol-% reduces the prognosis of trafficable grade by 0.83 grade percent. All the other factors enhance the prognosis of grade ability for the forwarder.

Changing the skeletal content of the top soil layers to one of the two superior subclasses (free of skeleton; up to 7%; over 7%) resulted, based on the measured data and the conducted linear regression, in an increase of the prediction (from the regression model, all other conditions remaining equal) of 6.2 grade percent. The skeletal content is divided into subclasses for practical

76

application of the whole model. The regression coefficients for the traction aids sum to 13.614 and point at the positive impact on traction ability.

Besides the grades accessible with an accepted level of wheel slip, the maximum inclinations that assure stability against sliding downhill are a question of practical relevance, and especially in cases where the mobility of machines in grades is assisted by a traction support winch (Forbrig et al. 2004; Nick 2005; Stuhlmann and Findeisen 2009). These techniques serve to extend the highly mechanized harvesting into slopes that cannot be traveled on according to the own drive line abilities or just with serious damages to the soil. In the light of safety for the machine operator and machine, it has to be assured that in case of a rope crack or technical problems the machine is able to stand without gliding. Limits depending on soil properties could be calculated similar to the already preCroat. j. for. eng. 34(2013)1

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

sented limits from the maximum values of the regression models. Since these models often lack a maximum or the values are not covered by measured data, another possibility was used. The measured data were assigned to one of ten groups of wheel slip, which are distributed evenly over the slip range from 0% to 100%. Data with < 10% slip were related to the first group, 10% < 20% to the second and so on. The average values of the traction coefficient were calculated for all of the 57 test series in every group. The maximum (average value) in each case (test series) served again as dependent variable of a linear regression. The resulting regression model is presented in Table 3.

On the basis of the presented regression models, trafficable grades can be estimated depending on soil properties and machine facilities. Table 4 presents the resulting ecological (max. 25% slip) and maximum (safety against sliding) inclination limits for a forwarder on loess soil for a wide range of soil water contents.

At the same time, distinction is made between the application of traction aids (tracks and chains) and the use of the forwarder just on wheels. For good weather conditions with low soil water contents even grades up to 34% seem trafficable uphill with a maximum of 25% wheel slip (ecological limit). Under the same conditions the model assumes stability against sliding up to longitudinal inclinations of 50%. When soil moisture increases, the trafficable grades are reduced for example to 21% inclination at 40% soil water content.

Notably higher slope angles can be climbed with the machine when tracks and chains are mounted as well as when the top soil layer offers an increased skeletal rate.

3.4 Validation – Provjera modela Table 4 presents inclination limits for loess soils calculated on the theoretical assumption of confrontation of traction force and downhill slope force. Although it works in theory, there is no proof that »off road« reality sticks to the theory. In order to answer that question, validation measurements in inclined terrain were taken parallel to the traction tests. The aim of these measurements was to verify the ecological inclination limits due to a maximum of 25% wheel slip.

Tests persist of trial drives up hill on steady slopes between 11% and 40%. The aim of the test was not to investigate the maximum limits assuming stability against gliding, as such investigation would be too dangerous.

During the tests, slip was measured and reduced to average and maximum values for further considerations. The aim was not to verify a precise prognosis of the wheel slip level but to make comparison with Croat. j. for. eng. 34(2013)1

Jörg Hittenbeck

the limits calculated from the model presented in Table 4. For every trial drive in inclined terrain, it was checked if mobility of the machine was possible with less than 25% slip from the model as well as from the observed average values. For 19 out of 31 validation tests, the inclinations seemed trafficable with no more than 25% wheel slip. Only one of these tests led to an average slip just above that limit. Driving at inclinations above the limit calculated from the model resulted in seven cases clearly exceeding the slip limit.

One test series led to an average slip level of well below 25%, whereas the maximum slip clearly exceeded that level. Four test series showed unexpected good traction for the forwarder so that driving uphill was possible without exceeding the slip limit of 25%.

4. Discussion – Rasprava The validation tests for the ecological limits showed that the prognosis (trafficable with ≤ 25% slip) matched the reality for over 80% of these tests. The maximum inclinations of the model were not tested. However, the validation test stated that the maximum inclinations from the model are not trafficable by driving uphill, which was expected. Anyhow the resulting maximum inclination values might be used for the question of limitations for machinery working downhill and for machines with traction support winches.

Both should only be done or used in areas where the stability of the machines against gliding is assured.

In the light of various factors that influence traction abilities and mobility in grades, the present approach of calculating limits for inclinations indicates a fairly good agreement between the model and validation tests. When evaluating the model, it has to be kept in mind that it was developed especially for loess soil.

These are for example widespread in the temperate zone in Europe (and here typical for the middle mountain ranges in Germany, where the tests were conducted) or the Midwestern United States (Haase et al.

2007). The focus on loess soils limits the scope of validity of the present models.

With increasing soil water content in loess soils, the interaction between soil and drive line of the machines changes. Under rather dry conditions, the frictional properties regarding the Mohr-Coulomb failure criterion (Wong 2010) of the soil dominate, while the cohesive part is marginal. For these conditions shear strength of the soil and hence also traction force depends on the internal shearing resistance of the soil material and the normal force acting on the soil. In wet loess soils the cohesive properties dominate so that the possible traction force depends on the contact area and

77

Jörg Hittenbeck

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

the cohesion of the terrain (Wong 2010). It is then independent of the normal force (under flat conditions the weight (in N)). This however conflicts with equation 2, where it is assumed that traction force is the result of normal force and coefficient of friction. For a wide range of trafficable soil, it can be expected that the frictional properties of the soil dominate and therefore equation 2 can be seen as a justifiable simplification. In addition, the traction force was measured at varying soil water contents in order to calculate the traction coefficient by division with weight of the machine. For solving the question of calculating the inclination limits, it could be assumed that the measurements compensate the main errors made by the theoretical traction equation (2) used.

Even though the present model gives appropriate inclination limits, it has to be kept in mind that working at the limits of machine trafficability is not just a matter of soil conservation and timber mobilization but also stress for the operator. Most of the machines used in steeper terrain are not equipped with tilt facilities, which allow a relaxed positioning of the operator. Harvesters are often provided with tiltable cabins but there are very few forwarders equipped with tiltable driver seats. This results in ergonomic stress (Lambert and Howard 1990; Heinimann 1999) for the operator. Apart from the physical strain, there are psychic stresses related to increased risks and higher requests for the operation because of the reduced handling opportunities. In order to ease the harvesting and forwarding operations in steep terrain, the machines that are regularly used for inclined areas should be equipped with tiltable cabins or driver seats and cranes. This would reduce the physical strain for the operator and in case of a tiltable boom the damages to the residual trees would be reduced.

5. Conclusion – Zaključak Even with the mentioned inaccuracy, the approach of calculating trafficable inclinations from measurements of traction force and slip leads to a good prognosis. The resulting model for grade ability on loess soils indicates that a soil conserving highly mechanized harvesting is possible in a wide range of inclination angles. However, it should be taken into consideration that the resulting limits from the model are based on the soil properties. Therefore, highly mechanized harvesting in inclined terrain should be linked to dry weather conditions or to the application of traction aids to allow the requested mobility on grades.

For rather unfavorable soil conditions (skeletal free, 40% soil water content) the present inclination model

78

results in an ecologically trafficable inclination of at least 20%. Under the same conditions, a machine equipped with tracks and chains is able to climb inclinations up to 35%. For both examples, however, it has to be noticed that the bearing capacity of the soil is already reduced. Considerable soil compaction effects could be expected. It should be emphasized that machine traffic has to be reduced to dry and favorable weather conditions, especially in inclined terrain, in order to reduce the damage to the soil and hence to ensure the long-term trafficability.

Harvesting operations in a terrain with more than 20% inclination should therefore be supported early enough by traction aids (tracks and/or chains). This improvement of soil protection (Kremer et al. 2007) and safety for the operator is the result of damages done to the root system of the trees at the skidding lane edge (Schardt et al. 2007) and of increased machine weight (Jacke 2007). In addition, there are different types of tracks that are optimized for special operation conditions and therefore differ in their ground pressure and traction abilities.

Acknowledgements – Zahvala The above project was financed by the German Ministry of Food, Agriculture and Consumer Protection (BMELV) under grant number 22027605.

6. References – Literatura Ansorge, D., Godwin, R. J., 2007: The effect of tyres and a rubber track at high axle loads on soil compaction. Part 1: Single axle-studies. Biosyst Engineering 98: 115–126.

Ansorge, D., Godwin, R. J., 2008: The effect of tyres and a rubber track at high axle loads on soil compaction. Part 2: Multi-axle machine studies. Biosyst. Engineering 99(3): 338– 347.

Arvidsson, J., Keller, T., 2007: Soil stress as affected by wheel load and tyre inflation pressure. Soil and Tillage Research 96(1–2): 284–291.

Bailey, A. C., Raper, R. L., Way, T. R., Burt, E. C., Johnson, C. E., 1996: Soil stresses under a tractor tire at various loads and inflation pressures. Journal of Terramechanics 33(1): 1–11.

Bazzoffi, P., Pellegrini, S., Rocchini, A., Morandi, M., Grasselli, O., 1998: The effect of urban refuse compost and different tractors tyres on soil physical properties, soil erosion and maize yield. Soil and Tillage Research 48(4): 275–286.

Blackwell, P. S., Graham, J. P., Armstrong, J. V., Ward, M. A., Howse, K. R., Dawson, C. J ., Butler, A. R., 1986: Compaction of a silt loam soil by wheeled agricultural vehicles. I. Effects upon soil conditions. Soil and Tillage Research 7(1–2): 97– 116.

Croat. j. for. eng. 34(2013)1

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

Jörg Hittenbeck

Forbrig, A., Lelek, S., Nick, L., 2004: Holzernte am Hang mit Seiltragschlepper – verfahrenstechnische Nische oder zukunftsweisender Quantensprung. (Timber Harvesting in Inclined Terrain with a Rope Forwarder – Process Niche or Trend Setting Quantum Leap). Forsttechnische Informationen 11–12: 153–155.

bewehrten Rückegassen und gewachsenem Boden bei variierenden Reifeninnendrucken. (Ecological and Economical Optimization Approaches in Forwarding from Spruce Stands. Traction Performance on Slash Armored Skid Trails and Bare Soil at Varying Tyre Inflation Pressures). IFA-Mitteilungen.

Haase, D., Fink, J., Haase, G., Ruske, R., Pécsi, M., Richter, H., 2007: Loess in Europe–its spatial distribution based on a European Loess Map, scale 1:2,500,000. Quaternary Science Reviews 26(9–10): 1301–1312.

Jun, H., Way, T. R., Lofgren, B., Landstrom, M., Bailey, A. C., Burt, E. C., McDonald, T. P., 2004: Dynamic load and inflation pressure effects on contact pressures of a forestry forwarder tire. Journal of Terramechanics 41(4): 209–222.

Heinimann, H., 1999: Ground-based Harvesting Technologies for Steep Slopes. Proceedings of the International Mountain Logging and 10th Pacific Northwest Skyline Symposium.

Kremer, J., Schardt, M., Borchert, H., Matthies, D., Ganter, C., 2007: Bogiebänder im Direktvergleich mit Reifen. (Tracks in Comparison with Tyres). Forst und Technik 1: 6–11.

Hittenbeck, J., 2004: Zugkraft- und Schlupfmessung an einem Forwarder unter besonderer Berücksichtigung von Aspekten der Bodenverträglichkeit in Fichtendurchforstungen, (Traction Force and Slip Measurements with a Forwarder with Special Regard to Aspects of Soil Tolerance in Thinnings of Spruce). Masterthesis Georg-August Universität.

Kunze, G., Göhring, H., Jacob, K., Scheffler, M., 2002: Baumaschinen. Erdbau und Tagebaumaschinen (Construction machines. Earthmoving and pit mining); mit 664 Abbildungen und 147 Tabellen. 1. Aufl. Braunschweig: Vieweg (Fördertechnik und Baumaschinen).

Hittenbeck, J., 2009: Entwicklung eines Grenzneigungsmodells für selbstfahrende Arbeitsmaschinen in der Forstwirtschaft. (Development of a Model for Trafficable Inclinations for Self Propelled Working Machines in Forestry). Dissertation Georg-August-Universität Göttingen, Cuvillier-Verlag.

Hoepke, E., Appel, W., 2002: Nutzfahrzeugtechnik. Grundlagen, Systeme, Komponenten. (Commercial vehicles technique. Background, systems, components). 2., überarb. Aufl.

Braunschweig: Vieweg (ATZ/MTZ-Fachbuch).

Horn, R., Fleige, H., 2009: Risk assessment of subsoil compaction for arable soils in Northwest Germany at farm scale, Soil Management for Sustainability. Soil and Tillage Research 102(2): 201–208.

Horn, R., Vossbrink, J., Becker, S., 2004: Modern forestry vehicles and their impacts on soil physical properties. Soil Physical Quality. Soil and Tillage Research 79(2): 207–219.

Horn, R., Vossbrink, J., Peth, S., Becker, S., 2007: Impact of modern forest vehicles on soil physical properties. Meeting the challenges of process-oriented management. Outcomes of the e-Conference »ForwardFORESTs«. Forest Ecology and Management 248(1–2): 56–63.

Hutchings, T. R., Moffat, A. J., French, C. J., 2002: Soil Compaction under timber harvesting machinery: a prelimary report on the role of brash mats in its prevention. Soil Use and Management 18(1): 34–38.

Jacke, H., 2007: Halbherzige »Bändermänner«. (Half-Hearted Tracks Men). Forst und Technik 8: 22–23.

Jacke, H., Drewes, D., 2004: Kräfte, Schlupf und Neigungen.

Ein Beitrag zur Terramechanik forstlicher Arbeitsmaschinen.

(Forces, Slip and Grades. A Contribution to Terramechanics in Forestry). Forst und Holz 59(6): 259–262.

Jacke, H., Drewes, D., Hittenbeck, J., 2004: Ökologische und ökonomische Optimierungsansätze des Tragschleppereinsatzes in Fichtenbeständen. Zugkraftentwicklung auf reisigCroat. j. for. eng. 34(2013)1

Lambert, M. B., Howard, J. O., 1990: Cost and productivity of new technology for harvesting and in-woods processing small-diameter trees. Pacific Nortwest Research Station, Research Paper PNW-RP-430.

Moffat, A. J., 1991: Forestry and soil protection in the UK. Soil Use and Management 7(3): 145–150.

Moore, D. F., 1975: The friction of pneumatic tyres. Elsevier, Amsterdam.

Mosimann, T., Sanders, S., Brunotte, J., 2008: Erosionsschutz in Fahrgassen. (Protection against Erosion in Machine Tracks). Landtechnik 63(1): 20–21.

Nick, L., 2005: Erster KWF-geprüfter Hangforwarder Forcar FC 200. (First KWF Tested Forwarder for Grades). Forsttechnische Informationen 8–9: 99–100.

Pampel, W., 1982: Holztransport. (Timber transport). Unter Mitarbeit von P. Hascke und Gaumitz. B. 1. Aufl. Berlin: VEB Deutscher Landwirtschaftsverlag.

Podsiadlowski, S. T., 1988: Wind erosion of light soil in the wheel tracks of a farm tractor. Journal of Agricultural Engineering Research 39(4): 231–243.

Raper, R. L., 2005: Agricultural traffic impacts on soil. Assessing the Impacts of Military Vehicular Traffic on Natural Areas. Journal of Terramechanics 42(3–4): 259–280.

Schardt, M., Kremer, J., Borchert, H., Matthies, D., 2007: Wurzelschutz beim Einsatz von Forwardern. (Protection of Roots at Forwarding Operations). Forst und Technik 2: 6–11.

Schulze, A., 1988: Theorie des Militärkraftfahrzeugs. (Theory of military vehicles). 1. Aufl. Berlin: Militärverl. d. Dt. Demokrat. Republik.

Söhne, W., 1952. Die Kraftübertragung zwischen Schlepperreifen und Ackerboden. (Power Transmission between Tractor Tyre and Farm Soils). Grundlagen der Landtechnik (3): 75–87.

79

Jörg Hittenbeck

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

Stuhlmann, C., Findeisen, E., 2009: Flexible und pflegliche Ernte im Hangübergangsgelände mittelsteiler Lagen. (Flexible and Protective Harvesting in the Change Over between Trafficable Slopes and Non-Trafficable Grades). Forst und Holz 64(3): 40–51.

Tebrugge, F., During, R.-A., 1999: Reducing tillage intensity – a review of results from a long-term study in Germany. Soil and Tillage Research 53(1): 15–28.

Trautner, A., Arvidsson, J., 2003: Subsoil compaction caused by machinery traffic on a Swedish Eutric Cambisol at different soil water contents. Experiences with the Impact and Prevention of Subsoil Compaction in the European Union.

Soil and Tillage Research 73(1–2): 107–118.

Von Wilpert, K., Schäfer, J., 2006. Ecological effects of soil compaction and initial recovery dynamics. A preliminary study. European Journal of Forest Research 125: 129–138.

Wanjii, S., Hiroma, T., Ota, Y., Kataoka, T., 1997: Prediction of wheel performance by analysis of normal and tengential stress distributions under the wheel-soil interface. Journal of Terramechanics 34(3): 165–186.

Wästerlund, I., 1983: Growth reduction of trees near strip roads resulting from soil compaction and damaged roots. A literature survey. Sveriges Skogsvaardsfoerbunds Tidskrift 81(2): 97–112.

Wong, J. Y., 2010: Terramechanics and off-road vehicle engineering. Terrain behaviour, off-road vehicle performance and design. 2nd ed., Elsevier, Amsterdam, pp. 206–351.

Yoshida, K., Hamano, H., 2002: Motion Dynamics and Control of a Planetary Rover With Slip-Based Traction Model.

Tohoku University, Sendai, Japan.

Zoz, F. M., Grisso, R. D., 2003: Traction and Tractor Performance. ASAE. St Joseph, ASAE Distinguished Lecture Series, pp. 1–48.

Sažetak

Procjena razreda kretnosti na nagibu na temelju vučne značajke forvardera Drvo kao obnovljivi resurs postaje sve zanimljiviji kao materijal, ali i kao izvor energije. Kako bi se udovoljilo potrebama za drvom, potrebno je da se dosad neiskorištene količine drva učine dostupne tržištu. Zbog toga je sve više zanimanja za nedovoljno iskorišteno drvo na nagnutim terenima. Nagnutost terena vrlo je važan ograničavajući čimbenik koji sprječava upotrebu potpuno visoko mehaniziranih sustava pridobivanja drva i koji smanjuje stabilnost vozila pa vozila čine štetu na tlu.

Utjecaj horizontalne sastavnice težine forvardera (FT na slici 1) upućuje na izravnu povezanost stupnja nagiba i faktora neto vuče. U teoriji se čini mogućim izračunavanje razreda kretnosti na nagibu za prihvatljivo proklizavanje kotača ili izračunavanje graničnoga nagiba pomoću najveće trakcije. Na osnovi odabrane granične vrijednosti klizanja od 25 % razvijen je i testiran model određivanja razreda kretnosti na nagibu.

Mjerenja faktora neto vuče obavljena su na prapornom tlu, na forvarderu Ponsse S 10. Za potrebe istraživanja konstruirano je dinamometarsko vitlo koje je služilo za mjerenje vučne sile pri kretanju neopterećenoga forvardera, pa sve do sila koje su prelazile mogućnosti istraživanoga vozila. Nadalje, kako bi se ispitala pouzdanost razvijenoga modela, provedena su istraživanja, bez upotrebe vitla, na nagnutom terenu uz ograničavanje klizanja na 25 %. Za poboljšanje kretnosti vozila na nagibu često se upotrebljavaju polugusjenice i lanci pa su zbog toga provedena istraživanja uz dva različita tlaka punjenja u gumama, pri različitim vrstama tovara te usporedbom novih i istrošenih guma. Tijekom istraživanja prikupljeni su podaci o samom mjestu istraživanja, kao što su vlažnost tla, kamenitost terena, gustoća tla, debljina humusnoga sloja i vrsta drveća (zbog različitih korijenskih sustava).

Mjerenja vučne sile nasuprot klizanju istaknula su vlažnost tla i kamenitost terena kao glavne značajke tla koje utječu na mogućnost kretanja strojeva uz nagib. Što se tiče značajki samoga stroja, razlika između guma i tlaka punjenja u gumama ima mali utjecaj, ali upotreba polugusjenica i lanaca značajno povećava trakciju te samim time i mogućnost kretanja forvardera uz nagib. Stoga je procjena razreda kretnosti obavljena na osnovi značajki tla (vlažnosti tla, kamenitosti) i upotrebe polugusjenica i lanaca.

Testne vožnje na nagnutom terenu pokazuju da se računanjem ograničenja razreda kretnosti na osnovi istraživanja vučne sile na ravnom terenu dobiju zadovoljavajući rezultati. Provjerom rezultata dokazano je da preko 80 % proračuna, za klizanje ≤ 25 %, odgovara stvarnim uvjetima. Najveći mogući nagib proračunat u modelu nije testiran.

80

Croat. j. for. eng. 34(2013)1

Estimation of Trafficable Grades from Traction Performance of a Forwarder (71–81)

Jörg Hittenbeck

No, kako je i očekivano, prekid je kretnosti dosegnut puno prije nego što predviđa model na osnovi maksimalnih faktora neto vuče. Čak i uz neke netočnosti, pristup računanju razreda kretnosti na nagibu preko mjerenja vučne sile i klizanja daje dobre rezultate. Razvijeni model pokazuje da je upotreba okolišno pogodnih mehaniziranih sustava pridobivanja drva moguća na različitim nagibima. Bitno je naglasiti da je osobito na nagnutim terenima kretanje šumskih strojeva dopušteno samo pri suhom i lijepom vremenu kako bi se smanjili negativni utjecaji strojeva na šumsko tlo te tako osigurala dugoročna prohodnost. Faze mehaniziranoga pridobivanja drva na nagibima većim od 20 % trebale bi se izvoditi uz upotrebu polugusjenica i lanaca.

Ključne riječi: mehanizirano pridobivanje drva, nagib, prapor, klizanje kotača, model

Author’s address – Autorova adresa:

Received (Primljeno): January 31, 2012 Accepted (Prihvaćeno): September 5, 2012 Croat. j. for. eng. 34(2013)1

Jörg Hittenbeck, PhD. e-mail: jhitten@gwdg.de Department of Forest Work Science and Engineering Faculty of Forest Science and Forest Ecology Georg-August-University Göttingen Büsgenweg 4 Göttingen GERMANY

81

Original scietific paper – Izvorni znanstveni rad

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System Miroslav Stanovský, Ján Schürger, Martin Jankovský, Valéria Messingerová, Richard Hnilica, Marian Kučera Abstract – Nacrtak Portable chainsaw is a basic operational instrument in forestry. To ensure long-lasting performance of a chainsaw, it is necessary to perform the necessary maintenance of its cutting system. The basis of this maintenance lies in the lubrication of the chain, which protects it from excessive wearing and deterioration. Various lubricating oils are used for lubricating the chain.

Biodegradable oils are promoted nowadays, with their main advantage being shorter decomposition time in the environment, and the same lubricating properties as conventional lubricants. This paper is aimed to evaluate and compare the ability to lubricate and cool the cutting system of a portable chainsaw filled with biodegradable oil and oils commonly used for its operation. Conditions of measurements were specified to simulate the actual operating conditions. During the evaluation no major differences between these oils were observed.

Keywords: biodegradable oil, portable chainsaw, temperature, thermograph

1. Introduction – Uvod At present, forestry machinery used on forest roads and in forest stands or other areas presents an ecological threat due to oil and petrol leakage. It is therefore logical that much effort is put into developing environmentally friendly fuels and lubricants, which are biodegradable or of biological origin. The decomposition time varies and depends on many factors, such as: temperature, humidity, quantity and type of bacteria, quantity of oxygen, etc. (Stanovský et al.

2012a). As Stanovský et al. (2012b) states, in the state forest enterprise of Slovakia (Lesy SR) alone, 1.8 million liters of oil were spilled into soil from chain saw operation (Lesy SR manages approximately 50% of the forest area in Slovakia). Finland forestry workers discharge about 2 million liters of oil per year, in Croatian forestry the number is 420 thousand liters (Oršanić et al. 2008). In Germany, the gap between oil consumption and oil purchase in forestry alone is 2 million liters, which implies the possibility of toxic oils being used in logging operations (Hartweg and Keilen 1989).

Based on many studies, it is safe to say that every lubricant is harmful to the environment. However, lubricants and fuels produced from biological matter are significantly less harmful than those of fossil origin Croat. j. for. eng. 34(2013)1

and are preferred over mineral or synthetic lubricants with similar characteristics in forest management (Schaffer and Buchschacher 2002). Based on empirical data, there are some apparent problems with using bio-oils. These are mainly shortened lifespan of the saw and thickness of the oil at lower temperatures.

Current legislation (Nature Protection Act 543/2002 Z.z., act 326/2005 Z.z., 360/2007 Z.z.) defines zones with different environmental protection intensity, which range from zones with almost no restrictions (1st level of protection) to zones with forbidden human intervention (5th level of protection). When logging in zones with higher levels of environmental protection, typically there is a requirement to use bio-oils in the operation. Forestry workers usually say that they avoid bio-oils because their lubricating characteristics are not as good as those of conventional lubricants. We tried to test this claim in conditions that simulate the conditions in practice.

2. Material and methods – Materijal i metode Two lubricants were selected for the evaluation – STIHL BioPlus (BP) and SHELL Helix Ultra VX 5W-30

83

M. Stanovský

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

(HU). BP oil is a lubricant designed for chain lubrication in chainsaws. It is a bio-oil and as such it is less harmful to the environment than conventional lubricants. The HU lubricant is synthetic motor oil and although when new it could be designed to protect the environment in case of spill in the nature, used motor oils are hazardous to the environment. The HU lubricant was drained from a combustion engine after the recommended mileage and then used in the measurements. This kind of lubricant (used motor oil) was selected because it is disturbingly commonly used in Slovak forestry. The basic characteristics of the lubricants can be seen in Table 1.

Two chainsaws were selected for measurements – Husqvarna 346XP and STIHL MS440. These chainsaws are commonly used in forestry for different purposes – the Husqvarna chainsaw is mainly used in thinning and the STIHL is most commonly used in regeneration felling. Their main characteristics are shown in Table 2.

2.1 Instrumentation – Mjerni uređaji The measurements were made with a FLUKE Ti25 thermal imager for temperature data collection, a stopwatch for time measurement of the operational cycles, a caliper for roundwood diameter measurements and a thermometer for outdoor temperature measurements. Specifications of the thermal imager are presented in Table 3 below.

Table 3 Thermal imager specifications Tablica 3. Tehnički podaci termalne kamere Specifications – Tehnički podaci Temperature measurement range, °C Raspon mjerenja, °C Temperature measurement accuracy, °C Preciznost mjerenja, °C Detector type

Table 1 Main characteristics of lubricants used in research Tablica 1. Glavne značajke ulja korištenih pri istraživanju Characteristic

BioPlus

Značajke ulja

Shell HU VX 5W-30

Melting point, °C Točka taljenja, °C

-2 – 10

-39

Flash point, °C Točka zapaljenja, °C

>230

230

0.92

0.848

37

67.1

Density (at 15°C), g/cm3 Gustoća (pri 15°C), g/cm3 Kinematic viscosity (at 40°C), mm2/s Kinematička viskoznost (pri 40°C), mm2/s

Table 2 Chainsaw characteristics Tablica 2. Značajke motornih pila korištenih pri istraživanju Characteristic

STIHL MS440

Husqvarna 346XP

70.7

50.1

4

2.7

Weight, kg Masa, kg

6.3

5

Oil tank volume, l Rezervoar za ulje, l

0.33

0.28

Idle speed, rpm Minimalni broj okretaja, min-1

2500

2700

Značajke Cylinder volume, cm3 Obujam, cm3 Performance, kW Snaga, kW

84

–20 to +350 ±2°C or 2% 160x120 uncooled microbolometer

Vrsta senzora Infrared spectral band, µm Raspon infracrvenoga spektra, µm Minimum span, °C

7.5 to 14 5

Minimalni raspon, °C Operating temperature, °C

–10 to +50

Radna temperatura, °C

The thermal data was subsequently evaluated by the SmartView 3.1 software. Statistical evaluations were carried out in Statistica 9 program and MS Excel 2010.

2.2 Measurements – Mjerenja The measurements took place on the grounds of the university campus and in a log conversion depot of the University forest enterprise in Lieskovec (approx. 5 km from the University). During the measurements, the following characteristics were measured: time of chainsaw operation during measurements, temperature of the cutting system (the center of the bar, chain on the upper side of the bar, the tip of the chainsaw bar), diameter of processed wood, number of cuts made with the chainsaws, outdoor temperature. Also the tree species was taken into consideration. The outdoor temperatures during individual measurements were taken into account. They were recorded in 2 hour intervals and then used during the evaluation. The SmartView program contains algorithms to compensate the temperature difference.

Croat. j. for. eng. 34(2013)1

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

M. Stanovský

The measurements were divided into two parts: A) Measurements without the cutting load, B) Measurements under cutting load.

2.2.1 Measurements without the cutting load Mjerenja bez prerezivanja drva One measurement cycle, during which thermal images were taken, consisted of the following intervals: 1) A snapshot of the cutting system without prior movement of the chain, in operational temperature of the chainsaw engine, 2) The second snapshot was taken after one minute of chainsaw movement in full engine load and maximum chain revolutions per minute (RPM), 3) The third interval lasted for two minutes. The chainsaw was in full engine load and maximum chain RPM, the snapshot was taken after finishing the interval, 4) The fourth interval lasted for three minutes. The chainsaw was in full engine load and maximum chain RPM, the snapshot was taken after finishing the interval.

There was a minimal pause (ranging from 30 to 40 seconds) between individual measurement intervals during which the snapshots were taken. Measurement cycles were carried out three times for both chainsaws and lubricants. Between the measurement cycles, the chainsaws were at rest to ensure their cooling down.

2.2.2 Measurements under cutting load – Mjerenja s prerezivanjem drva One measurement cycle, during which the thermal images were taken, consisted of the following intervals: 1) A snapshot of the cutting system without prior movement of the chain, in operational temperature of the chainsaw engine, 2) The second snapshot was taken after one minute of cutting in full engine load, 3) The third interval lasted for two minutes. The chainsaw was under cutting load and in full engine load, the snapshot was taken after finishing the interval, 4) The fourth interval lasted for three minutes. The chainsaw was under cutting load and in full engine load, the snapshot was taken after finishing the interval.

Measurements under cutting load were divided into two subgroups – hardwood and softwood cutting.

Hardwood used in the measurements consisted entirely of beech, softwood used during the measurements was mostly spruce, and on three occasions fir Croat. j. for. eng. 34(2013)1

Fig 1 The cutting process Slika 1. Prikaz prerezivanja drva wood was used. The cutting process is presented in Fig. 1.

Again there was a minimal pause (ranging from 30 to 45 seconds) between individual cutting intervals during which the snapshots were taken. Measurement cycles were carried out three times for both chainsaws, lubricants and wood type. Between the measurement cycles, the chainsaws were at rest to ensure their cooling down.

3. Results – Rezultati During the measurements, 144 snapshots of the cutting system were taken. The average diameter of processed wood was 45.6 cm and on average 5.49 cuts were made into the wood (average cut area was 0.9 m2) in one measurement interval.

The STIHL MS440 chainsaw reached lower temperatures while cutting with the BP oil, on average 0.27 °C lower. The average temperature difference was 3.57 °C in favor of the HU oil on the chain. The overall temperature difference of the bar and tip of the bar was 1.35 °C in favor of the BP oil, but the reason for that is that the BP oil rendered better results in measurements without cutting load (-10.50 °C on average).

As for measurements under cutting load, the chainsaw bar temperature was on average 3.23 °C lower when filled with the HU oil. The overall difference in temperatures of the bar tip was 3.03 °C, but again this was mainly due to better results of the bio-oil without the cutting load, where the difference was 14.37 °C in favor of the BP oil. In measurements under cutting load, the HU oil proved to be better reaching 2.64 °C lower temperatures than the BP oil.

85

M. Stanovský

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

Fig. 2 Infrared snapshots of the Husqvarna cutting system, left – 1st interval, right – 4th interval Slika 2. Infracrvene slike reznoga mehanizma Husqvarna, slika lijevo 1. interval, slika desno 4. interval The results of measurements carried out on the Husqvarna 346XP chainsaw had similar outcome. The BP oil performed worse with this chainsaw than with the STIHL chainsaw. The temperatures of PB oil were on average 2.68 °C lower than the temperatures of the HU oil. The chain temperature difference was 6.80 °C in favor of the BP oil. The only measurement group where the bio-oil tested worse than the synthetic oil was cutting softwood, where the HU oil reached 0.18 °C lower temperatures. The overall average temperature of the bar was also lower with the BP oil, but the difference was not as big as on the chain (0.71 °C). The temperatures of the bar, while operating with the BP oil, were higher in measurements without the cutting load (2.92 °C) and while cutting softwood (3.20 °C). This pattern was visible on the tip of the bar as well, where overall temperature difference was 0.51 °C in favor of the BP oil. On average the BP oil reached 8.26 °C higher

temperature without the cutting load and 5.48 °C when cutting softwood. In hardwood cutting, the tip reached 15.26 °C lower temperatures when filled with the BP oil.

The development of temperatures in the cutting systems of particular chainsaws is shown in Fig. 2 and 3. It is clear, that the highest temperatures are reached at the tip of the chainsaws and on the periphery of the bar, where there is friction between the chain, chain saw bar and material cut.

The average temperatures of the bar, tip and chain for both chainsaws and all measurement groups can be seen in Table 4 and 5.

Although more wood was processed with the STIHL MS440 chainsaw, it reached lower temperatures than the Husqvarna 346XP chainsaw, as can be seen in Fig. 4. Fig. 4 also shows variation of temperatures of the chainsaws.

Fig. 3 Infrared snapshots of the STIHL cutting system, left – 1st interval, right – 4th interval Slika 3. Infracrvene slike reznoga mehanizma STHIL, slika lijevo 1. interval, slika desno 4. interval

86

Croat. j. for. eng. 34(2013)1

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

M. Stanovský

Table 4 Results of measurements with the Husqvarna chainsaw, average temperature in °C Tablica 4. Rezultati mjerenja s motornom pilom Husqvarna, prosječne temperature u °C Husqvarna 346XP BioPlus Oil – Ulje

Shell Helix Ultra

Chain – Lanac

Bar – Vodilica

Tip – Vrh vodilice

Chain – Lanac

Bar – Vodilica

Tip – Vrh vodilice

1. interval

17.53

19.77

19.63

19.67

19.30

19.33

Without cutting

2. interval

49.33

53.77

83.10

54.73

49.83

72.87

Bez prerezivanja

3. interval

61.70

81.70

108.00

75.13

77.73

100.50

4. interval

65.47

92.40

119.30

82.57

91.53

112.27

1. interval

18.20

19.77

20.20

21.97

23.23

22.17

Beech cutting

2. interval

47.87

55.53

83.63

52.97

58.67

91.67

Prerezivanje bukve

3. interval

54.03

68.47

94.97

55.53

78.13

113.20

4. interval

53.87

72.67

102.90

73.10

84.67

122.40

1. interval

16.77

18.67

17.67

16.37

17.00

17.37

Spruce cutting

2. interval

46.83

52.87

81.83

49.43

51.50

77.33

Prerezivanje smreke

3. interval

61.13

74.50

104.57

58.47

74.57

100.57

4. interval

69.13

85.87

112.53

68.67

77.57

104.60

Table 5 Results of measurements with the STIHL chainsaw, average temperature in °C Tablica 5. Rezultati mjerenja s motornom pilom STHIL, prosječne temperature u °C STIHL MS440 BioPlus Oil – Ulje

Shell Helix Ultra

Chain – Lanac

Bar – Vodilica

Tip – Vrh vodilice

Chain – Lanac

Bar – Vodilica

Tip – Vrh vodilice

1. interval

23.10

21.97

22.47

27.40

30.17

28.80

Without cutting

2. interval

47.60

39.40

54.13

52.07

49.90

66.70

Bez prerezivanja

3. interval

60.47

58.97

76.70

66.97

69.03

90.17

4. interval

65.40

67.57

85.30

68.67

78.50

102.37

1. interval

20.67

22.90

22.43

17.87

18.87

17.27

Beech cutting

2. interval

44.60

42.50

52.90

40.03

40.37

50.53

Prerezivanje bukve

3. interval

52.93

54.33

69.27

49.87

53.73

70.80

4. interval

58.60

58.77

73.00

49.40

58.90

69.67

1. interval

12.07

12.73

12.13

16.67

18.33

17.07

Spruce cutting

2. interval

47.07

41.83

50.13

32.47

38.27

47.03

Prerezivanje smreke

3. interval

55.33

58.93

69.03

42.83

51.27

64.13

4. interval

48.97

63.20

75.90

46.50

57.67

72.20

Croat. j. for. eng. 34(2013)1

87

M. Stanovský

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

Fig. 4 Box charts showing temperature variability of cutting systems with both chainsaws and lubricant types Slika 4. Slikovni prikaz varijabilnosti temperature reznoga mehanizma obiju motornih pila i obiju vrsta ulja

4. Discussion – Rasprava The infrared snapshots, as seen in Fig. 2 and Fig. 3, show that the heat is mostly produced in the areas with most friction (tip of the bar and periphery of the bar) from where it spreads to the center of the bar. The highest temperatures were recorded on the tip of the

88

chainsaw, where the bearing that enables chain revolutions, is located. Typically, the temperature of the bar was higher than the temperature of the chain.

Several biodegradable oils with some characteristics similar or better than those of mineral oils are available in the market (Lauhanen et al. 2000). The data acquired show that STIHL BioPlus oil and used Croat. j. for. eng. 34(2013)1

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

Shell Helix Ultra synthetic motor oil can both lubricate the cutting system and reduce friction between individual components and between the cutting system and cut wood. The results also show that the bio oil performed worse than the used synthetic motor oil in softwood cutting. This is possibly due to characteristics specific to softwood (e.g. higher volume of resins, lower density, different structure, etc.). Furthermore as Klvač et al. (2002) states, thin layers of rapeseed oil easily oxidize when exposed to air. An oxidized rapeseed oil is sticky and may cause chain blocking. Residual oxidized oil can also restrict piston movement during oil pumping, resulting in damage to the drive wheels of the transmission system.

Both oils are suitable for use in chainsaws but only regarding friction reduction. From the environmental point of view, the use of used motor oils has vast negative effects on the soil and water in the ecosystems.

Skoupý (2004) states that 12–16% of oil sets on the surface and most of it is washed into underground or groundwater (Skoupý et al. 1990; Skoupý and Ulrich 1994). Used motor oils typically contain traces of metals (lead, iron, zinc, copper, etc.) and complex carbohydrates (Lengyel and Cvengroš 2008). It takes decades and/or a lot of financial resources to remediate soil from oil pollution.

5. Conclusion – Zaključak The data collected during measurements fail to show significant differences in temperature of the cutting system that would be dependent on the type of lubricating oil. Our research disproves the assumption of forestry workers that bio-oils have a highly negative impact on the wear of the chainsaw cutting system.

However, it is necessary to understand that the measurements carried out during our research were only short-term, meaning that some differences in lubricating characteristics were not able to show. The bio-oil studied does not provide significantly better lubrication for the chainsaw and in some cases it performs worse than the used synthetic motor oil. Its main advantage lies in its rapid degradation in the soil and very low ecotoxicity, which is a sufficient reason, in our opinion, to use it in practice.

Acknowledgements – Zahvala Ti25 thermal imager was provided by Apertis, Ltd. Zvolen.

This study is the result of implementation of the project VEGA 1/0525/09: Use of ecologically appropriate media in hydraulic and lubricating systems of forest machines.

Croat. j. for. eng. 34(2013)1

M. Stanovský

This contribution/publication is the result of implementation of the project: Centre of Excellence »Adaptive Forest Ecosystems«, ITMS: 26220120006, supported by the Research & Development Operational Programme funded by the ERDF.

6. References – Literatura Act on nature and environment protection, Zbierka zákonov, 543/2002.

Act on forests, Zbierka zákonov, 326/2005.

Act which changes law no 326/2005 on forests and law no 217/2004 on forest reproduction material, Zbierka zákonov, 360/2007.

Hartweg, A., Keilen, K., 1989: The environmental benevolence of biological oils. Scottish Forestry 43: 311–317.

Klvač, R., Kanali, C., Skoupy, A., 2002: Low temperature pumping characteristics of biodegradable chainsaw oils.

Research in Agricultural Engineering 48(2).

Lauhanen, R., Kolppanen, R., Takalo, S., Kuokkanen, T., Kola, H., Valimaki, I., 2000: Effects of biodegradable oils on forest machines and forest environment. Proc. of the Scientific Conference on Forest and Wood Technology vs the Environment, Brno: MZLU, pp. 203–206.

Lengyel, J., Cvengroš, J., 2008: Zložky neminerálnej povahy v mazacích olejoch (Non-mineral compounds in lubrication oils). Nakladanie s odpadovými olejmi s SR. Vyhne.

Oršanić, M., Horvat, D., Pernar, N., Šušnjar, M., Bakšić, D., Drvodelić, D., 2008: Growth of Pedunculate Oak Seedlings under Soil contamination by Mineral and Biodegradable Oils. Croatian jounal of forest engineering 29: 155–162.

Schaffer, H. P., Buchschacher, R., 2002: Setzen Sie umweltverträgliche Treib- und Schmierstoffe ein. Merkblatt Vollzug Umwelt.

Skoupý, A., 2004: Biologically degradable oils at working with power saws. Journal of Forest Science 50: 542–547.

Skoupý, A., Pechlak, B., Sejkora, B., 1990: A contribution of oil dispersion at the work with a chain saw by means of Radiotracer Method. Lesnictví 36: 937–946.

Skoupý, A., Ulrich, R., 1994: Dispersion der Ölabscheidung aus der Kettenschmierun von Einmanmotorsägen. KWF Forsttechnische Informationen, pp. 121–132.

Stanovský, M., Schürger, J., Jankovský, M., Allman, M., Messingerová, V., Hnilica, R., Kučera, M., 2012: Znečistenie pôdy olejmi a možnosti čistenia pôdy po znečistení olejmi.

Použitie ekologicky vhodných médií v hydraulických a mazacích systémoch lesných strojov, p. 8.

Stanovský, M., Schürger, J., Messingerová, V., 2012: Kvantifikácia množstva mazacích olejov uvoľnených do prostredia lesa pri použití prenosných reťazových píl. Použitie ekologicky vhodných médií v hydraulických a mazacích systémoch lesných strojov, Zvolen.

89

M. Stanovský

The Effect of Lubricating Oil on Temperature of Chainsaw Cutting System (83–90)

Sažetak

Utjecaj maziva na temperaturu rezne garniture motorne pile U slovačkim šumama godišnje se u okoliš prospe 1,8 milijuna litara ulja za podmazivanje lanaca motorne pile. Od ulja koje se upotrebljava za podmazivanje lanaca motornih pila najštetnije je sintetičko motorno ulje kojim se još uvijek većinom podmazuju lanci motornih pila. Okolišno prihvatljivije bilo bi ulje koje se, usprkos svim naporima, koristi u manjim količinama. Šumski radnici sjekači tvrde da bioulje ima negativan učinak na tehničku ispravnost motornih pila.

Cilj je ovoga istraživanja bio usporediti svojstva bioulja i sintetičkoga ulja kako bi se teze šumskih radnih sjekača potvrdile ili opovrgnule. Pri istraživanju je korišteno iskorišteno sintetičko motorno biorazgradivo ulje Shell Helix Ultra VX 5W-30 i bioulje STHIL BioPlus. Mjerenja su obavljena na dvjema motornim pilama – Husqvarna 346XP i STHIL MS 440.

Termalne su fotografije snimane pomoću termalne kamere FLUKE Ti25, a preciznost termalne kamere je +/- 2 °C s »uncooled moicrobolometer« senzorom. Mjerenje je provedeno tijekom rada motorne pile bez prerezivanja drva i tijekom prerezivanja bukova i smrekova drva.

Mjerilo se u četiri režima rada: 1. Snimanje temperature motorne pile prije upotrebe 2. Snimanje temperature motorne pile nakon 1 minute rada 3. Snimanje temperature motorne pile nakon 2 minute rada 4. Snimanje temperature motorne pile nakon 3 minute rada.

Ukupno su snimljene 144 fotografije. Prosječan je promjer prerezanoga drva bio 45,6 cm, u svakom režimu rada prosječno je napravljeno 5,5 rezova, a prosječna prerezana površina iznosila je 0,9 m2. Na temelju prikupljenih podataka sa sigurnošću možemo reći da obje vrste ulja, korištene u istraživanju, osiguravaju dostatno podmazivanje lanca motorne pile i smanjuju trenje između rezne garniture motorne pile i materijala koji se prerezuje. Pri prerezivanju mekoga drva (drva četinjača) sintetičko ulje pokazalo se boljim nego bioulje.

Istraživanje pokazuje da vrsta ulja nema značajniji utjecaj na trošenje rezne garniture. Bioulje nije se pokazalo boljim (ni lošijim) nego iskorišteno sintetičko motorno ulje, a njegova je glavna prednost smanjen negativan utjecaj na okoliš i brza razgradnja u šumskim ekosustavima, što je dovoljan razlog za povećanjem korištenja bioulja u šumarskoj praksi.

Ključne riječi: biorazgradivo ulje, motorna pila, temperatura, termograf

Authors’ address – Adresa autorâ:

Received (Primljeno): January 10, 2012 Accepted (Prihvaćeno): July 19, 2012

90

Miroslav Stanovský, PhD. e-mail: stanovsky@tuzvo.sk Ján Schürger, MSc. e-mail: jan.schurger@szm.sk Martin Jankovský, MSc.* e-mail: martin.jankovsky@apertis.eu Prof. Valéria Messingerová, PhD. e-mail: messingerova@tuzvo.sk Department of Forest Harvesting, Logistics and Amelioration; Technical University in Zvolen T. G. Masaryka 24, 960 53, Zvolen Richard Hnilica, PhD. Department of Manufacturing Technology and Materials Technical University in Zvolen Študentská 26, 960 53, Zvolen Asst. Prof. Marian Kučera, PhD. Department of Forest and Mobile Technology Technical University in Zvolen T. G. Masaryka 24, 960 53, Zvolen SLOVAKIA * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

Original scietific paper – Izvorni znanstveni rad

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia Mohammad Reza Ghaffariyan, John Sessions, Mark Brown Abstract – Nacrtak Roadside chipping is a common harvesting system to produce chips in Australian plantations.

This study investigated the productivity and cost of road-side chipping operation (chipping logs extracted by forwarder to the road side) using a Morbark chipper with flail delimber in a first thinning of Pinus radiata stands. An elemental time study method was used to collect the time working cycles. The regression approach was used to develop the productivity predicting model based on the log size in different wood piles. The statistical analysis yielded an average productivity of 59.4 GMt/PMH0 with the corresponding costs of 5.2 AU$/GMt for the Morbark chipper. The details on work time analysis, relocation time and fuel consumption of the machine are documented in this paper. The results provide basic information for planning roadside chipping operations in pine plantations.

Keywords: Morbark chipper, Loader, Truck, Work time, Productivity, Cost, Model

1. Introduction – Uvod Chipping can be done at the mill, at a storage yard, at forest roadside or in the stand (Kühmaier et al. 2007).

The most common option for many regions is chipping at the forest roadside. About 70% of the annual woody biomass production in Finland is chipped at roadside (Ranta and Rinne 2006; Junginger et al. 2005). Road side chipping is the most common due to the cost benefits (Ghaffariyan 2010). Spinelli et al. (2009, 2002) reported that the full-tree harvesting system with roadside chipping allows lower cost harvesting and transport than the CTL system for a range of conditions.

Roadside chipping is also common in Australian plan tations and utilises a mobile chipper to produce pulp chips in the forest. Road side chipping is preferred to the other harvesting systems due to minimizing materials processing (Lambert 2006). Stems must be debarked to produce quality pulp chips. Debarking for roadside chipping in Australia can be performed either by debarking the stems at the stump using a single-grip harvester, or alternatively, by debarking the stems with a chain flail delimber and debarker at the forest road prior to chipping. The flail and chipper are often integrated in one machine as in the Peterson Pacific flail chipper. The system of roadside chipping with debarking at the stump was developed by Eumeralla Pty Ltd Croat. j. for. eng. 34(2013)1

and AFM Pacific in Australia in 1998, for Timbercorp Limited (Lambert 2006). The system of roadside chipping with debarking at the forest road is currently used in the Green Triangle Region, Albany and Bunbury in Australia.

Two recent studies on roadside chipping in Western Australia have reported productivity of 33.90 GMt/PMH0 for the Peterson Pacific chipper (Wiedemann and Ghaffariyan 2010) and 51.7 GMt/PMH0 for the Husky precision chipper (Ghaffariyan et al.

2011) in Eucalypt plantations. The difference between productivity of both studies was due to tree size and machine power. Larger tree size and machine power resulted in higher productivity (Spinelli and Hartsough 2001). Both studies indicated that truck waiting time was the major operational delay.

Since the productivity and cost of chipping logs at roadside have not been documented in Australian pine plantations, this project investigated the productivity of roadside chipping operation in a first thinning of Pinus radiata. The objectives of this trial were to: Þ Determine the productivity and cost of roadside chipping operation, Þ Study the impact of log size on chipper productivity, Þ Determine the fuel consumption of the chipping system.

91

M. R. Ghaffariyan et al.

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

2. Study area and harvesting system – Mjesto istraživanja i sustav pridobivanja

product output (chip weight) was determined based on the delivered green metric tons (GMt) of chips (from truck weights).

The study area was a flat 19.7 ha pine plantation with an original stocking of 1561 stems per ha and an average tree size of 0.12 m3 near Mt Gambier, South Australia. The thinning system was performed by a cut-to-length (CTL) harvest system consisting of a harvester and forwarder, producing logs at roadside that were chipped by a Morbark B12 truck-based chipper (500 hp) directly into trailers for transport. The chipper was equipped with the debarking flail. A Hitachi ZAXIS 250L loader was used to feed the chipper (Fig. 1).

The same operator was used for loader and chipper.

Four B-double trucks (Fig. 2) were used for transport.

The logs were chipped straight into the trailer. The chips were transported to the Carter Holt Harvey MDF mill in Mt Gambier. The transportation distance varied from 22.0 to 27.5 km.

Fig. 1 Hitachi loader – Morbark chipper Slika 1. Utovarivač Hitachi – Iverač Morbark

3. Method – Metoda The study took place in September 2011. Elemental time study method was used to evaluate the machine productivity for chipping 7 piles. Element level measurement consists of splitting the work cycle into functional steps (elements) and then recording time consumption separately for each of them. This allows the work process to be described in more detail, which may contribute to a better understanding of process dynamics (Magagnotti and Spinelli 2012). In this case study, the working cycle was defined as the time required for loading each truck. The working cycle in this project was divided to six working elements (Table 2). Working delays (including personal, mechanical and operational delays) during the operation were also recorded by stopwatch (Table 3). The collected data at each pile consisted of work cycles per pile, average log diameter, and standard deviation of log diameter measurements within each pile (Table 4). The

Fig. 2 Morbark chipper and chip van Slika 2. Iverač Morbark i kamion za prijevoz drvne sječke

Table 1 Harvesting machines used at study site Tablica 1. Strojevi za proizvodnju drvne sječke korišteni tijekom istraživanja Type

Make

Model

Hours used

Operator experience, years

Tip stroja

Proizvođač

Model

Pogonskih sati

Iskustvo rukovatelja, god.

Hitachi

ZAXIS 250L

3000

7

Morbark

B12

4000

7

Tracked base loader Gusjenični utovarivač Chipper Iverač

92

Croat. j. for. eng. 34(2013)1

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

M. R. Ghaffariyan et al.

Table 2 Definition of working elements Tablica 2. Definicije radnih sastavnica Work element

Definition

Radna sastavnica

Definicija

Truck move to chipping place

Starts when truck moves to chipper and ends when chipping is started – Počinje kada se kamion pomiče prema iveraču, a završava s početkom iveranja

Pomicanje kamiona do mjesta iveranja

Starts when operator starts picking up the logs and feeding into chipper and ends when trailer is full and truck commences travelling – Počinje kada rukovatelj zahvati oblovinu i stavlja ju u iverač, a završava kada je prikolica puna i kamion se počinje kretati

Chipping Iveranje Moving chipper

Any time spent to move the chipper along the pile – Svako vrijeme utošeno za pomicanje iverača uz složaj

Premještanje iverača Debris clean up by loader

Any time spent by the operator to pick up the debris and clean up the chipping area – Svako vrijeme koje rukovatelj utroši za podizanje ostataka drvne sječke i čišćenje mjesta iveranja

Uklanjanje ostataka utovarivačem Planned fueling and knife change

Any time to fuel the loader/chipper and change the knives of the chipper – Svako vrijeme utrošeno za točenje goriva u utovarivač i iverač i za promjenu noževa iverača

Planirano točenje goriva i promjena noževa

Starts when chipper/loader starts moving to new pile in another place and ends when first truck starts moving into location to be loaded – Počinje kada se iverač i utovarivač počinju kretati do novoga složaja na drugom mjestu i završava kada se prvi kamion počinje kretati na mjesto utovara

Relocate to next pile Premještanje do sljedećega složaja

Table 3 Working delay Tablica 3. Zastoj rada Delay – Zastoj

Definition – Definicija Any interruption to previous elements (note cause of delay: operational, personal or mechanical) Svaki prekid prethodnih sastavnica (zabilježen razlog zastoja: organizacijski, osobni ili mehanički) Delay will be treated as follows: – Prekid će biti tretirani: Delays <0.1 minute (6 seconds) are included in the element in which they occur as the time interruption is considered too short to constitute a delay – Zastoji < 0,1 minute (6 sekundi) uključeni su u sastavnicu u kojoj su nastali jer se prekid rada smatra prekratkim da bi činio zastoj Delays <15 minutes are recorded as delays and included in productive time – Zastoji <15 minuta zabilježeni su kao zastoji i uključeni u proizvodno vrijeme Delays >15 minutes are considered non-productive time and excluded – Zastoji >15 minuta smatraju se neproizvodnim vremenom i isključeni su

Delay / Non-productive time Zastoj / Neproizvodno vrijeme

Table 4 Study lay-out and data collected at each pile Tablica 4. Dizajn istraživanja i podaci prikupljeni po pojedinom složaju Pile number

Collected work cycles

Average log diameter, cm

Standard deviation for log diameter, cm

Redni broj složaja

Broj snimljenih radnih ciklusa

Prosječni promjer oblovine, cm

Standardna devijacija promjera, cm

1

14

18.2

5.5

2

6

16.4

5.1

3

21

17.1

5.2

4

13

17.5

5.5

5

17

13.8

3.1

6

4

13.4

4.3

7

7

14.0

4.3

Croat. j. for. eng. 34(2013)1

93

M. R. Ghaffariyan et al.

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

The log diameter (at one end) was measured using sampling of 85 logs per pile. The pile length was divided to five sections. In each section, the log end diameter were randomly measured at three heights including bottom, middle and top to sample the logs from all places in the pile at both sides (front and back side of the pile to roadside). For each pile the average log diameter was calculated using the recorded samples. According to the analysis of variance using Tukey method, the average diameter of the piles 1, 2, 3 and 4 was not significantly different of each other but it was different from piles 5, 6 and 7. For the other piles, the average diameter of the piles 5, 6 and 7 were not significantly different.

The lengths of logs averaged 5 m based on the measurement records of 30 logs at the end of each pile. The average log volume was 0.103 m3, which was calculated based on Huber’s formula by multiplying the log length with the average sectional area of a log at its mid-point. In this trial, the average height and length of the piles was 4 m and 66 m, respectively. The distance between piles averaged 199 m (minimum distance of 20 m and maximum distance of 440 m).

3.1 Statistical analysis – Statistička analiza 3.1.1 Modelling – Modeliranje The working time and productivity were plotted depending on the average log diameter for the pile.

The productivity model was developed using the regression method in SPSS 18. The statistical procedure for modeling included: Þ plotting the working time depending on the parameter, Þ regression application to develop the model (considering outliers outside three standard deviations), Þ comparing different model types based on fit, error and plausibility, Þ checking model consistency,

Þ analyzing the variance to test significance of the model, Þ examining the residuals of the model and model evaluation, Þ sensitivity analysis to quantify the impact of the independent variable on chipper productivity.

3.1.2 Productivity – Proizvodnost Productivity was calculated from the delivered green metric tonnes (GMt) of chips (from truck weights) with the productive machine hours excluding all delays (PMH0) and productive machine hours excluding delays longer than 15 minutes (PMH15).

The fuel consumption of the loader and chipper was also recorded during the operation to estimate the consumption per produced unit of chips.

4. Results – Rezultati 4.1 Productivity model – Model za izračun proizvodnosti Average productivity of the chipper was 59.4 GMt/PMH 0 (56.6 GMt/PMH15). The confidence interval for the mean net productivity is 59.20±2.29 GMt/PMH0 at the significance level of 0.05.

A model was developed to predict chipping productivity.

Productivity (GMt/PHM0) = 18.79 + 2.505 × Average log end diameter of each pile, cm R2 = 19.2%, n = 79, df = 1,76, F = 18.07, p = 0.00059

4.2 Model evaluation – Ocjena modela From the collected work cycles, three samples were randomly taken out from the data and the model was developed without these witness samples. Then to verify the validity of the model, the confidence intervals of the coefficients were calculated in SPSS for the linear model (Table 5).

Table 5 Confidence intervals for coefficients of the model Tablica 5. Intervali pouzdanosti za koeficijente modela

Model Model

Unstandardized Coefficients

95.0% Confidence Interval for B

Nestandardizirani koeficijenti

95,0 %-tni interval pouzdanosti za B

B

Std. Error

Lower Bound

Upper Bound

Standardna pogreška

Donja granica

Gornja granica

Constant – Konstanta

18.79

9.56

–0.26

37.84

Log diameter – Promjer oblovine, cm

2.50

0.59

1.33

3.68

94

Croat. j. for. eng. 34(2013)1

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

M. R. Ghaffariyan et al.

Table 6 Random witness samples and validation test Tablica 6. Slučajni uzorci izostavljeni iz regresije i provjera valjanosti

Diameter, cm

Log length, m

Pay load, t

Promjer, cm

Duljina oblovine, m

Masa tovara, t

Productivity, GMt/PMH0

Chipping time, min/cycle Vrijeme iveranja, min/turnus

Proizvodnost, GMt/PMH0 Actual

Upper limit

Lower limit

Predicted

Stvarna

Gornja granica

Donja granica

Predviđena

16.40

5.41

25.50

19.45

42.3

98.2

21.6

59.9

17.10

4.95

25.22

22.41

42.9

100.8

22.5

61.6

13.80

4.60

25.72

27.09

57.0

88.6

18.1

53.4

Table 7 Descriptive statistics of productivity model at a = 0.05 Tablica 7. Opisna statistika modela za izračun proizvodnosti za a = 0,05 Minimum

Maximum

Mean

Std. Deviation

Najmanja vrijednost

Najveća vrijednost

Aritmetička sredina

Standardna devijacija

78

35.30

82.90

59.40

10.34

78

13.40

18.20

16.13

1.81

N Net productivity, GMt/PMH0 Efektivna proizvodnost, GMt/PMH0 Log diameter, cm Promjer oblovine, cm

The upper limit and lower limit for prediction for each witness sample are calculated at α = 0.05 (Table 6). Since each observed productivity is within the limits, the model is considered to be valid at a = 0.05.

The actual productivity of each witness sample (Table 6) was also compared with the predicted productivity by the developed model using the Paired Samples T Test (Spinelli and Magagnotti 2010). The significance level of this test was 0.272, which was higher than a = 0.05. This indicated that there was no significant difference between the actual and predicted values.

Mean productivity is 59.40 GMt/PMH0 at the mean log diameter of 16.13 cm (Table 7). The minimum and maximum value of net productivity and log diameter among the recorded cycle times of the chipper in this case study is included in Table 7.

Fig. 3 Impact of log diameter on chipping productivity (for average log length of 5 m). Data points are actual observations. Solid line is a straight line created by varying log diameter in the productivity model. Upper and lower confidence limits of the prediction are presented Slika 3. Utjecaj promjera oblovine na proizvodnost iveranja (za prosječnu duljinu komada oblovine od 5 m). Točke na grafu stvarna su opažanja. Puna je linija nacrtana variranjem promjera oblovine u modelu za izračun proizvodnosti. Prikazane su gornja i donja granica pouzdanosti procjene Croat. j. for. eng. 34(2013)1

Chipping larger diameter logs of the same length will increase chipping productivity and reduce the time to fill each trailer (Fig. 3).

4.3 Work element times – Vrijeme radnih sastavnica The operator spent most of the working time (83.99%) chipping (Fig. 4). The working delays were grouped into three categories; personal, mechanical and operational delays. The major operational delay

95

M. R. Ghaffariyan et al.

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

Fig. 4 Time breakdown (% of working time) for chipper Slika 4. Raščlamba vremena (u % od radnoga vremena) za iverač Table 8 Minimum, Maximum and Average values for the work element times Tablica 8. Najmanje, najveće i prosječne vrijednosti za utrošak vremena pojedinih radnih sastavnica Work element, minutes

Minimum

Maximum

Mean

Radna sastavnica, minute

Najmanja vrijednost

Najveća vrijednost

Aritmetička sredina

0.0

23.1

0.8

16.9

33.0

23.2

Move chipper – Premještanje iverača

0.5

4.9

1.5

Debris clean up by loader – Uklanjanje ostataka utovarivačem

0.4

5.1

1.7

Planned fuelling-knife change – Planirano točenje goriva i promjena noževa

6.4

15.6

10.7

Relocate to next pile – Premještanje do sljedećega složaja

4.5

16.2

9.4

Personal delay – Zastoj zbog osobnih potreba

0.1

5.5

2.3

Mechanical delay – Mehanički zastoj

6.6

6.6

6.6

Operational delay – Organizacijski zastoj

2.0

16.6

8.1

Truck move to chipping place – Pomicanje kamiona do mjesta iveranja Chipping – Iveranje

(4.37% of working time) occurred when the chipper was waiting for a truck. The mechanical delay occurred due to breakage in the loader grapple. The total work time observed (including delays) was about 2232.8 minutes in this case study.

96

The average work cycle time was about 27.2 minutes including delays shorter than 15 minutes. Minimum and maximum values for the work cycle time (including delays shorter than 15 minutes) were 17.3 and 43.5 minutes, respectively. The descriptive statistics of Croat. j. for. eng. 34(2013)1

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

M. R. Ghaffariyan et al.

Table 9 Time records for the major move of chipper-loader Tablica 9. Evidentirani utrošci vremena za glavno premještanje iverača i utovarivača Time, min. – Vrijeme, min.

1. Relocate chipper – Premještanje iverača 2. Relocate excavator – Premještanje utovarivača

21.15 117.29

2.1 Wait for truck for relocating – Čekanje na kamion koji će prevesti utovarivač

74.40

2.2 Loading into truck – Utovar utovarivača

18.34

2.3 Travel to next pile – Vožnja do sljedećega složaja

20.00

2.4 Unloading excavator from truck – Istovar utovarivača

4.55

the work element times recorded in this case study are presented in Table 8.

In this case study, the average payload of the trucks was about 24.7 GMt (with the minimum and maximum of 22.7 GMt to 26.9 GMt, respectively).

4.4 Major relocation – Glavno premještanje After completing the last pile in this study, the chipper and loader were moved to another forest area.

The chipper was truck-based and moved itself to the new location, whereas the excavator-based loader was moved using a float (lowboy) causing a much longer delay. The total time for the major move was 1.95 hours for a distance of 10.4 km (Table 9).

4.5 Fuel consumption – Utrošak goriva The fuel consumption for the chipper averaged 72.6 l/PMH15 and the loader consumed 25.1 l/PMH15.

As the loader only worked while the chipper was operating (except for some very short exceptions), the total consumption was about 97.7 l/PMH15 (1.72 l/GMt) or about 65.4 MJ/GMt. The fuel cost for the chipper and loader is estimated to about 122.0 A$/PMH15.

4.6 Cost of operation – Trošak iveranja The machine hourly cost was calculated based on operating, fixed and labour cost (Table 10, 11) using ALPACA: Australian Logging Productivity and Cost Appraisal Model (Murphy and Acuna 2009). The main inputs for cost estimating using ALPACA include equipment purchase price, machine life, salvage value, utilization rate, repair and maintenance, fuel consumption, operator wage and scheduled machine hours.

Increasing log diameter reduced chipping cost (Fig. 5) as the larger log diameter resulted in higher productivity of the chipper.

Croat. j. for. eng. 34(2013)1

Fig. 5 Impact of log diameter on chipping cost Slika 5. Utjecaj promjera oblovine na trošak iveranja

5. Discussion – Rasprava Using a separate loader with the Morbark chipper is one of the reasons for the high cost in this case study compared to European chippers (Stampfer and Kanzian 2009). The result of the productivity model is similar to the chipping productivity in Italy, where 85% of total working time was spent chipping and the productivity of chipper was a function of piece size and machine power (Spinelli and Hartsough 2001; Spinelli et al. 2011). However, the average productivity for their study was about 13.2 GMt/PMH0 due to smaller piece size of 0.07 m3 and lower power of the chipper. Spinelli and Maganotti (2010) developed a productivity-cost estimation tool that included the en-

97

M. R. Ghaffariyan et al.

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

Table 10 Machine cost calculations (based on Australian Dollar AU$) Tablica 10. Izračun troška strojnoga rada (iskazana u australskim dolarima AU$) Machine description/costs – Značajke stroja i troškovi

Morbark B12

Hitachi ZAXIS 250L

375000

228800

Machine life, years – Uporabni vijek stroja, godine

7.0

5.0

Salvage value, % – Ostatak vrijednosti, %

20

20

Utilization rate, % – Iskorištenost, %

75

75

Repair and maintenance, percent of depreciation, % – Popravci i održavanje, postotak od amortizacije, %

100

100

Interest rate, % – Kamatna stopa, %

7

7

Insurance and tax rate, % – Stopa osiguranja i poreza, %

4

4

Fuel consumption rate, l/h – Potrošnja goriva, l/h

72.6

25

Fuel cost, $/l – Cijena goriva, $/l

1.25

1.25

25

25

Operator wage and benefit rate, $/SH – Plaća i doprinosi rukovatelja, $/SMH

25.20

0.00

Scheduled machine hours, h – Planirani radni sati, h

2200

2500

Salvage value, $ – Ostatak vrijednost, $

75000

45760

Annual depreciation, $ – Godišnja amortizacija, $

42857

36608

Average yearly investment, $ – Prosječna godišnja investicija, $

246429

155584

Productive Machine Hours, PMH – Pogonskih sati godišnje, PMH

1650

1875

Interest cost, $/year – Trošak kamata, $/godina

17250

10891

Insurance and tax cost, $/year – Trošak osiguranja i poreza, $/godina

9857

6223

Yearly ownership cost, $/year – Godišnji trošak posjedovanja stroja, $/godina

69964

53722

Ownership cost per SMH, $ – Trošak posjedovanja stroja po planiranom radnom satu, $

31.80

21.49

Ownership cost per PMH, $ – Trošak posjedovanja stroja po pogonskom satu, $

42.40

28.65

Fuel cost, $/h – Trošak goriva, $/h

90.70

31.36

Lube cost, $/h – Trošak maziva, $/h

22.67

7.84

Repair and maintenance cost, $/PMH – Trošak popravaka i održavanja, $/PMH

25.97

19.52

Operator labor and benefit cost, $/PMH – Trošak rada rukovatelja i doprinosi, $/PMH

33.60

0.00

5.04

0.00

Operating cost per PMH, $/PMH – Trošak posjedovanja stroja po pogonskom satu, $/PMH

177.99

58.72

Operating cost per SMH, $/SMH – Trošak posjedovanja stroja po planiranom radnom satu, $/SMH

133.49

44.04

Total cost per SMH, $/SMH – Ukupni trošak po planiranom radnom satu, $/SMH

165.29

65.53

Total cost per PMH, $/PMH – Ukupni trošak po pogonskom satu, $/PMH

220.39

87.37

Purchase price, $ – Nabavna vrijednost, $

Lube and oil, percent of fuel cost, % – Mazivo i ulje, postotak od troška goriva, %

Ownership costs – Troškovi posjedovanja stroja

Operating costs – Troškovi korištenja stroja

Supervision, $/PMH – Nadzor, $/PMH

Total Costs – Ukupni troškovi

98

Croat. j. for. eng. 34(2013)1

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

M. R. Ghaffariyan et al.

Table 11 Summary of cost-production of the chipping system Tablica 11. Rekapitulacija troška proizvodnje sustavom iveranja Chipper cost, AU$/PMH

Loader cost, AU$/PMH

System cost, AU$/PMH

Productivity, GMt/PMH0

Trošak iverača, AU$/PMH

Trošak utovarivača, AU$/PMH

Trošak sustava, AU$/PMH

Proizvodnost, GMt/PMH0

220.39

87.37

307.76

59.40

gine power and piece size. Based on their results, larger piece size resulted in higher chipping productivity, which is similar to our study result (chipping productivity model, Fig. 3).

The productivity of the Morbark chipper in our case study is higher than the reported productivity of 8.12 GMt per scheduled hours (SH) for a Morbark Super Beaver Chipper (with integrated feeding loader) used to chip Eucalypt trees into trailers near Orland, California, USA. This low productivity was caused by small tree DBH of 7.5 cm in the California case study (Hartsough and Nakamura 1990). Watson et al. (1986) tested two types of Morbark chipper (Models 27 and 20) to chip the trees for bioenergy purposes in pine and hardwood plantations in Alabama, USA. The average productivity for chipping pine trees at DBH of 7.5 cm for Model 27 (650 hp) and Model 20 (350 hp) was 49.10 and 27.70 GMt/PMH, respectively. DBH was found to be a significant variable for the chipping productivity predicting model. Compared to our study, the Morbark 27 is more productive than the Morbark B12 due to its higher engine power. However, for the same DBH (about 18 cm), the Morbark 20 model in Alabama recorded a productivity of 37.0 GMt/PMH (average moisture content of 52.9%), which is lower than our case study results due to difference in engine power.

In another study in loblolly pine (Pinus taeda) in Louisiana (USA), the productivity for chipping bundles of stems (with DBH from 7.5 cm to 22.5 cm) by Peterson 5000 flail-chipper was about 13.5 GMt/PMH (Watson and Stokes 1994). Stokes and Watson (1990) mentioned that using flail delimbing and debarking allows economical processing and chipping to produce clean and acceptable chips in slash pine (Pinus elliotti) plantations in southern United States. In their case study the Peterson Pacific 4800 log debarker was combined with the Morbark 22 chipper for chipping pine logs, while the Morbark B12 in our case study was integrated with a flail debarker.

Waiting for trucks is the typical operational delay with roadside chipping. This has been previously investigated in Quindinup, Western Australia (Ghaffariyan et al. 2011). For whole tree chipping in Western Australia, 13% of total chipping time was delay due to Croat. j. for. eng. 34(2013)1

AU$/GMt 5.18

waiting for trucks, which is higher than the current case study. Three trucks worked with one Husky Precision chipper in that case study. In this trial four trucks transported the chips produced by the Morbark chipper. Spinelli and Visser (2009) investigated the working delay of 63 chipping productivity studies.

The overall average utilization of the chipper was 73.8%. Regardless of operation type, two-thirds of the total delay time was represented by organizational delays, which emphasizes the crucial role of operation management. The percentage of operational delay was also larger than other work delays in our case study.

However, its percentage was less than average delay of 63 case studies analyzed by Spinelli and Visser (2009).

6. Conclusions – Zaključci Increasing the diameter of the logs for the same length, and up to the maximum diameter accepted by the machine will increase the productivity of the chipping operation. This is probably only true when comparing diameter differences of more than 3 or 4 cm.

Future studies can investigate the chipping productivity model for larger tree sizes in a second thinning and final cuts. To decrease the cost of operation, a chipper with loader attachment (O’Neal and Gallagher 2007) can be also tested. Using a separate loader increases hourly cost based on the assumptions for the whole crew although ergonomically it is better for the operator to be far from vibrations and noise. Another possibility will be the application of a smaller loader for small pulp logs with a lower hourly cost to feed the chipper.

The most delay time derived from chipping occurred due to waiting for trucks. Proper planning and management of the truck fleet is critical to infield chipper efficiency. Making sure there are enough trucks so that the system is not delayed has a cost too, which should be considered in chipping operation planning.

A fairly simple operation like this can be managed manually, however, for more complex operations, truck scheduling systems like Fast Truck (Acuna et al.

2012), ASICAM (Weintraub et al. 1996), Asset Forestry

99

M. R. Ghaffariyan et al.

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

New Zealand transportation system (Robinson 2012) or Forestry Transportation Management System by Trimble (www.trimbleforestryautomation.com) could be used.

Robinson, P., 2012: A system for managing the execution of supply plans and logistics in the forestry environment. Precision Forestry in Advance Symposium. 27–28 March 2012, Mount Gambier, South Australia, 32 p.

Acknowledgement – Zahvala

Spinelli R., Hartsough, B., Owende P., Ward, S., 2002: Productivity and Cost of Mechanized Whole-tree Harvesting of Fast-growing Eucalypt Stands. International Journal of Forest Engineering 13(2): 49–60.

Forestry SA and K.C. & M.R. Boult Company supported this research by providing their plantation, equipment and resources. The authors would like to thank the reviewers who have provided valuable comments which helped improve this article.

7. References – Literatura Acuna, M, Mirowski, L., Ghaffariyan, M. R., Brown, M., 2012: Optimising transport efficiency and costs in Australian wood chipping operations. Biomass and Bioenergy 46: 291– 300.

Ghaffariyan, M. R., 2010: Review of European biomass harvesting technologies. Silva Balcanica 11(1): 5–20.

Ghaffariyan, M. R., Brown, M., Acuna, M., Sessions, J., Kuehmaier, M., Wiedemann, J., 2011: Biomass harvesting in Eucalyptus plantations in Western Australia. Southern Forests 73(3–4): 149–154.

Hartsough, B. R., Nakamura, G., 1990: Harvesting eucalyptus for fuel chips. California Agriculture 44(1): 8–9.

Junginger, M., Faaij, A., Bjorheden, R., Turkenburg, W. C., 2005: Technological learning and cost reductions in wood fuel supply chains in Sweden. Biomass and Bioenergy 29(6): 399–418.

Kühmaier, M., Kanzian, C., Holzleitner, F., Stampfer, K., 2007: Wertschöpfungskette Waldhackgut. Optimierung von Ernte, Transport und Logistik (Value chain of wood chips.

Optimization of harvesting, transport and logistics). Institut für Forsttechnik, Department für Wald und Bodenwissenschaften an der Universität für Bodenkultur Wien, 283 p.

Magagnotti, N., Spinelli, R. 2012: COST Action FP0902 – Good practice guideline for biomass production studies.

CNR IVALSA. Florence, Italy, 41 p.

Murphy, G., Acuna, M., 2009: Australian logging productivity and cost appraisal model (ALPACA). CRC Forestry internal toolbox: http://www.crcforestry.com.au O’Neal, B. S., Gallagher, T. V., 2007: Designing and testing a small-scale biomass harvesting system in the Eastern United States. Austro2007/FORMEC’07: Meeting for Tomorrows’ Forests- New Developments in Forest Engineering, October 7–11, Vienna and Heiligenkreuz, Austria.

Ranta, T., Rinne., S., 2006: The profitability of transporting uncomminuted raw materials in Finland. Biomass and Bioenergy 30(3): 231–237.

100

Spinelli R., Ward, S., Owende, P., 2009: A harvest and transport cost model for Eucalyptus spp. fast-growing short rotation plantations. Biomass and Bioenergy 33(9): 1265–1270.

Spinelli, R., Hartsough, B., 2001: A survey of Italian chipping operations. Biomass and Bioenergy 21(6): 433–444.

Spinelli, R., Magagnotti, N., 2010: A tool for productivity and cost forecasting of decentralised wood chipping. Forest Policy and Economics 12(3): 194–198.

Spinelli, R., Magagnotti, N., Paletto, G., Preti, Ch., 2001: Determining the impact of some wood characteristics on the performance of a mobile chipper. Silva Fennica 45(1): 85–95.

Spinelli, R., Visser., R., 2009: Analyzing and estimating delays in wood chipping operations. Biomass and Bioenergy 33(3): 429–433.

Stampfer, K., Kanzian, Ch., 2006: Current state and development possibilities of wood chip supply chains in Austria.

Croatian Journal of Forest Engineering 27(2): 135–145.

Stokes, B. J., Watson, W. F., 1990: Wood recovery with inwoods flailing and chipping. In: Proceedings of the 1990 Tappi Pulping Conf.; 1990 Oct. 14–17; Toronto, Canada. Atlanta, GA: Tappi Press, p. 851–854.

Watson, W .F., Stokes, B. J., 1994: Cost and utilization of above ground biomass in thinning systems. In: Proceedings of the meeting on Advanced Technology in Forest Operations: Applied Technology in Action; 1994 July 24–29; Portland/Corvallis, OR. Corvallis, Or: Oregon State University, p. 192–201.

Watson, W. F., Sabo, R .F., Stokes., B. J., 1986: Productivity of in-woods chippers processing understory biomass. In: A proceedings of the Council of Forest Engineering, Improving productivity through Forest Engineering; September 29 – October 2; Mobile, AL, USA, p. 69–72.

Wiedemann, J., Ghaffariyan., M. R., 2010: Preliminary results: volume recovery comparison of different harvesting systems in short-rotation hardwood plantations. CRC for Forestry, Bulletin 9: 4.

Weintraub, A., Epstein, R., Morales, R., Seron, J., Traverso, P., 1996: A truck scheduling system improves efficiency in the forest industries. Iinterfaces 26: 4 July – August 1996, p.

1–12. http://www.trimbleforestryautomation.com

Croat. j. for. eng. 34(2013)1

Roadside Chipping in a First Thinning Operation for Radiata Pine in South Australia (91–101)

M. R. Ghaffariyan et al.

Sažetak

Iveranje na pomoćnom stovarištu nakon prve prorede šumske kulture kalifornijskoga bora u Južnoj Australiji Iveranje je na pomoćnom stovarištu uobičajen sustav proizvodnje drvne sječke u australskim šumskim plantažama.

Ovom je studijom istražena proizvodnost i troškovi iverača Morbark pri iveranju oblovine (izvezene forvarderom na pomoćno stovarište) iz prve prorede šumske kulture kalifornijskoga bora. Pokus je proveden u borovoj šumskoj kulturi površine 19,7 ha koja je podignuta na ravnom terenu blizu grada Mount Gambier u Južnoj Australiji. Drvnu je zalihu činilo 1561 stablo po hektaru, prosječnoga obujma 0,12 m3. Proveden je studij vremena na razini radnoga ciklusa pri iveranju sedam složaja oblovine. Svaki je radni ciklus raščlanjen na šest radnih sastavnica: pomicanje kamiona do mjesta ivaranja, iveranje, premještanje iverača, uklanjanje ostataka utovarivačem, planirano točenje goriva i promjena noževa i premještanje do sljedećega složaja.

Model za izračun proizvodnosti iveranja razvijen je pomoću regresije na temelju veličine komada oblovine u različitim složajevima iskazane promjerom. Ocjena je modela obavljena pomoću tri slučajna uzorka izostavljena iz regresije. Testiranje je potvrdilo da je model valjan za razinu pouzdanosti 0,05. Statistička je analiza rezultirala prosječnom proizvodnošću od 59,4 GMt/PMH0 i odnosnim troškovima u iznosu 5,2 AU$/GMt. Povećanje promjera oblovine dalo je veću proizvodnost iverača i niži trošak iveranja. Rukovatelj je većinu radnoga vremena (83,99 %) utrošio na iveranje. Zastoji u radu objedinjeni su u tri kategorije: osobne, mehaničke i organizacijske. Glavni organizacijski zastoj (4,37 % radnoga vremena) nastao je prilikom čekanja kamiona. Ukupno vrijeme utrošeno za glavno premještanje na udaljenost od 10,4 km iznosilo je 1,95 h. Prosječni je utrošak goriva iznosio 72,6 l/PMH15 za iverač i 25,1 l/PMH15 za utovarivač.

Rezultati ovoga istraživanja pružaju osnovne informacije za planiranje iveranja na pomoćnom stovarištu u borovim šumskim kulturama.

Ključne riječi: iverač Morbark, utovarivač, kamion, radno vrijeme, proizvodnost, trošak, model

Authors’ address – Adresa autorâ:

Received (Primljeno): October 10, 2012 Accepted (Prihvaćeno): December 20, 2012 Croat. j. for. eng. 34(2013)1

Mohammad Reza Ghaffariyan, PhD.* e-mail: ghafari901@yahoo.com University of the Sunshine Coast Private Bag 12 7001 Hobart AUSTRALIA Prof. John Sessions, PhD. e-mail: john.sessions@oregonstate.edu Oregon State University Department of Forest Engineering, Resources and Management Peavy Hall 204 OR 97331-5706 Corvallis USA Prof. Mark Brown, PhD. e-mail: mbrown2@usc.edu.au University of the Sunshine Coast Locked Bag 4 4558 Maroochydore, Queensland AUSTRALIA * Corresponding author – Glavni autor

101

Original scietific paper – Izvorni znanstveni rad

Increasing Productivity and Controlling of Work Fatigue in Forest Operations by Using Prescribed Active Pauses: a Selective Review Christos Gallis Abstract – Nacrtak This paper presents and discusses a selective review of the current work/rest schedule bibliography, and provides a theoretical model to be used to create work-rest schedules for forest workers. Forest machine and manual motor tool operators expose themselves to a variety of factors that may cause work fatigue. Prolonged fatigue may produce a variety of musculoskeletal disorders, decrease performance, and cause health problems. Pauses at work may provide time for physical recovery and for reduction of health hazards. Prescribed active pauses at work may help forest machine and manual motor tool operators to increase their net working time and to decrease commonly existing spontaneous and disguised pauses. Ten-minute active pauses may relieve forest operators from fatigue effects, provide time for recovery and maintain adaptation to work. The use of electromyography (EMG) accompanied by the use of the Checklist of Individual Strength questionnaire (CIS) would be the methodology for determining the validity and assessing the results of the above mentioned work/rest schedules in mechanized and motor-manual forest operations.

Keywords: work fatigue, health problems, prevent accidents, electromyography, productivity, work pauses, forest operators, work/rest schedules

1. Introduction – Uvod In spite of increasing mechanization and automation of many forest operations, there is still a need for muscular power and activity to operate hand tools and machines (Gallis 1992). Muscular power cannot be maintained for prolonged work because fatigue sets in and muscular power is reduced. In practice, forest machine and manual motor tool operators work very intensively during a work day, with a limited number of pauses, and are exposed to a variety of working conditions, which affect their working capabilities and may cause work fatigue, health problems, musculoskeletal problems, lower performance, and increase working errors, injuries, and accidents (Herberts et al.

1980; Axelsson and Ponten 1990; Harstela 1990; Gallis 1992; Gellerstend 1993; Kirk 1998; Gallis 2006). Even if performance has not declined, the subjective sensations of fatigue are a very important aspect of human work. Fatigability as an objective inability to sustain power and performance can be measured electro Croat. j. for. eng. 34(2013)1

physiologically, but it is not necessarily related to the subjective sensation of fatigue (May and Kline 1988).

Work fatigue related musculoskeletal disorders of the neck and upper limb continue to be of interest to individual forest worker, forest organizations and researchers. This is due to significant disability, time lost from work, increasing worker’s compensation and increasing number of cases coming before the courts that can be associated with them (Ong 1992; Barker 1995; Stock et al. 1996). Fatigue at work is a normal everyday experience. However, in the case of severe fatigue, it may affect the person’s performance both at work and home. Moreover, severe long term fatigue may lead to sick leave and work disability (Beurskens et al. 2000).

Work fatigue would be categorized as acute and long term fatigue. Acute fatigue is characterized by reversibility, task specificity, and the functional use of compensation mechanisms (Lewis and Wessely 1992).

Acute fatigue is a normal phenomenon that disappears after a period of rest, when tasks are switched,

103

C. Gallis

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

or when particular strategies are used – for example, working at a slower pace. By contrast, long term fatigue is irreversible, not task specific, and the compensation mechanisms that were useful in reducing acute fatigue are no longer effective (Beurskens et al. 2000).

The factors that may affect fatigue and deteriorate work performance may be grouped into three categories: environmental factors (noise, vibration, temperature, humidity), human factors (age, sex, health, motivation, fitness, capabilities, training, emotional stability) and task factors (complexity, duration, skills, physical, mental or sensorimotor tasks). The impact of each of the environmental factors depends upon the value it deviates from acceptable levels and the duration of exposure. There is also a limitless number of possible combinations of all environmental factors, which may alter stressfulness at higher levels. The situation becomes more complex with the interaction of human, task, and environmental factors. Ergonomic factors that showed the highest association with neck and shoulder problems occurred when the forest operator was only sitting operating a forest machine and using small control levers, machine keyboards and display units. Arm and hand problems were associated with the use of hand control levers and forest machine keyboard position. Organizational factors showed that limited or extensive peer contacts were associated with work-related musculoskeletal disorders, and limited rest break opportunity appeared to be a major factor for several muscular problems. Time pressures at work and the work-rest schedules were also shown to be associated with work related musculoskeletal disorders in a Brazilian study of a group involved in telephone-computer tasks (Ferreira et al.

1997).

It is almost impossible to define the contribution of each of these factors to fatigue and deterioration of performance, especially in the »unmanaged« working conditions in the forest (Gallis 1992). To manage this complex situation, and to prevent fatigue, improvement in work organization and management of the overall working conditions with job redesign, breaks, and physical activation of workers is required. When doing physical work, humans need adequate resting periods to recover from the physiological consequences of exertion (Beyon et al. 2000).

Work pauses are needed to maintain the working performance and efficiency required by work and to avoid or decrease fatigue effects (Janaro and Bechtold 1985; Henning et al. 1997; Koparadeker and Mital 1994).

The aim of this paper is to propose and discuss as follows:

104

Þ review both past and current literature, Þ underline and discuss the importance of work pauses by introducing a properly designed work/rest schedule, which will help forest workers to recover from fatigue during the working day, to decrease the effects of fatigue, and to reduce time exposure to environmental factors, Þ present a theoretical model that can be used to create work-rest schedules for forest workers, i.e. a work/rest schedule for forest harvesting ma chine and chainsaw operators, by providing prescribed active ten (10) minute pauses after every fifty minutes of work.

2. Work pauses and their effect on fatigue and productivity – Prekidi rada i njihov utjecaj na umor i proizvodnost Forest machine operators usually work very intensively during a workday with a limited number of pauses (Staaf and Wiksten 1984; Gallis 1992; Byers 1997; Kirk 1998). Such a mode of work is unacceptable since physical exhaustion can be caused by extended periods of strain. The capacity of muscles to perform either severe or prolonged work is ultimately limited by a variety of physiological changes, which decrease performance by producing a variety of symptoms called fatigue (Chaffin 1973). Prolonged muscle fatigue may produce a variety of musculoskeletal disorders. Muscular fatigue can seriously affect the safety of workers, as well as their health and performance (Herberts et al. 1980).

Although most mechanized tasks in forest operations demand smaller oxygen intake and cabin environment is generally improved, muscular fatigue due to repetitive tasks and the static posture imposed by manipulative tasks results in localized tenderness (soreness in a specific muscle when it is touched), occupational hazards, and repetitive strain injuries (Ferguson 1976; Onishi et al. 1977; Axelsson and Porten 1990; Harstela 1990; Nakata et al. 1992; Veiersted 1996; Byrne 2000). Porten (1990) reported that 50% of forest machine operators in Sweden complained of injuries to the neck, arm and cervical spine. In his study on Norwegian forestry workers, Hagen (1997) found that 87.7% of the forest machine operators and 86.0% of manual motor tool forest workers reported musculoskeletal disorders in at least one body part during the previous 12 months. The lower-back, neck and shoulders were the three most common sites. Pauses (breaks) are needed for physical recovery and for reduction of health hazards (i.e. noise, vibration, and Croat. j. for. eng. 34(2013)1

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103â&#x20AC;&#x201C;112)

temperature) (Simonson 1971; Staaf and Wiksten 1984; Gallis 1997; Bovenzi et al. 1998; Giannini et al. 1999; Ising et al. 1999). The positive preventive effect of breaks on muscle restoration has been stated in the literature by several authors (Basmajan and DeLuca 1985; Sundelin and Hagberg 1992; Veiersted 1995; Westgaard and DeLuca 1999).

People at work are used to taking different kinds of pauses under various circumstances: prescribed pauses, which are rest breaks at work laid down by the management; spontaneous pauses, which are the obvious rest pauses that workers take on their own initiative; disguised pauses, which are times when workers occupy themselves with easier tasks; workconditioned pauses, which are interruptions arising either from the operation or the machines or the organization of work; and active pauses, which are short pauses during which workers are not passive but either use other muscle groups or the same muscle groups with or without light loads, or walk or perform gymnastic exercises (Grandjean 1988; Laporte 1966; Simonson 1971).

Disguised, active, and spontaneous pauses are often used by workers to avoid excessive fatigue and to relieve from fatigue symptoms. The use of these pauses explains why old workers, with small work capacities and little resistance to fatigue, can stay at work in heavy tasks in forestry and farming.

Rest pauses are very important for maintaining working performance and the required work efficiency, and for avoiding or decreasing fatigue effects. In general, the important reasons for having rest pauses during work are (Grandjean 1988; Simonson 1971; Karpovich and Sinning 1971): to restore the balance of oxygen and nutrients to the muscles and to remove waste products, particularly lactic acid, to avoid a low level of blood sugar by spreading out the intake of food and to pay oxygen debt. Additionally, according to the above-mentioned authors, the effect of pauses is rapid restoration of blood concentration in the legs and feet and decrease of blood pressure in veins. A replacement of water losses is regularly possible through a cooling down of the body. Pauses seem to limit exposure time of vibration, noise, temperature, and humidity. Additionally, pauses break down the monotony of work and provide time for social contacts. Prevention of musculoskeletal disorders as well as prevention of errors and accidents caused by fatigue is also made possible.

For most occupations, pauses of less than one or two minutes are not practical. Therefore, this information is concerned with pauses of several minutes to half on hour or one hour. Generally, the effect of pausCroat. j. for. eng. 34(2013)1

C. Gallis

es on work performance is favorable, more so for shorter (about 10 minutes) than for prolonged pauses (Rohmert 1965; Simonson 1971; Zwahlen et al. 1984; Janaro and Bechtold 1985; Swanson et al. 1989; Henning et al. 1997).

The energy requirement is the highest at the beginning of work and decreases until a steady state is reached. Occasionally it even continues to decrease with continuation of work, as a result of removal of part of the oxygen (O2) debt accumulated at the beginning of work (Simonson 1971). If the work is continued to the point of fatigue, the energy requirement increases again, largely because motor coordination deteriorates and biochemical processes change.

Physiological and psychological adaptation to work is a fundamental process involved in all occupations. The adaptation to work is the basis of the warmup period, as practiced in athletics. Previous exercise of the same or even of a different type increases performance capacity for subsequent maximum exercise.

The basis for adaptation is complex, involving cardiovascular, metabolic, muscular, and nervous functions (Rohmert 1973; Sundelin and Hagberg 1989). The adaptation to work is gradually lost on the recovery but outlasts the recovery of the O2 debt within ten minutes, and adaptation disappears within thirty minutes (Simonson 1971). Thus, both the exponential course of the recovery of O2 debt and related cardiovascular functions and the loss of adaptation favor short rest pauses. The favorable effect of short pauses may be the reason why a surprisingly high number of older workers, in spite of decreased capacity of oxygen intake, remain in physically heavy jobs such as forestry and farming (Astrand and Rodahl 1986).

The question of rest breaks and determination of their duration has been the subject of several studies.

In their study on telephone directory assistants, Koparadekar and Mital (1994) determine the preferable work-rest schedule. The schedules examined were: 30 minutes work followed by a 5 minute break; 60 minutes followed by a ten minute break; 120 minutes without any break reported that performance did deteriorate when the work duration was increased from 30 minutes to 60 minutes by 11% more errors. If no rest was provided performance deterioration became great, with errors increasing by almost 80% during 120 minutes of continuous work. The deterioration in performance was halted after a rest break.

Dul et al. (1991) developed a model to find the optimum work-rest schedule for static work. The model predicted that for a given total time and total rest time, many short-rest periods are better than fewer long work-rest cycles, and this model showed relatively

105

C. Gallis

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

good agreement with the actual measurements. In order to decrease fatigue of repetitive work, Kogi (1982) recommended brief intra-work pauses where the muscles are rested from static load, and a break after a period of continuous work.

Rohmert (1965) suggested that the reduction of fatigue by intermittent pauses is the most effective way of maintaining or increasing work performance. Graf (1954) reported that short breaks have shown an increase in net working time and a decrease in spontaneous and disguised pauses (Fig. 1). Rest breaks, taken by workers at their own initiative, decrease the net working time, while short breaks that are laid down by the management have favorable effects on net working time (Fig. 1). Henning et al. (1997) aimed to determine whether frequent, short rest breaks (three of 30-s and one of 3-min) from computer work each hour, in addition to conventional rest breaks, had a positive influence on workers’ productivity and wellbeing in processing of insurance claims at two sites.

They reported that there was an improvement in productivity, eye, leg and foot comfort when the short breaks included stretching exercises. These results provide evidence that frequent short breaks from continuous computer-mediated work can benefit worker productivity and well-being when the breaks integrate with task demands.

Dababneh et al. (2001) studied the impact of frequent short rest breaks on the productivity and well

being of a group of 30 workers in a meat-processing plant. Two rest break schedules were tested, both of which provided 36 min of extra break time over the regular break schedule (30-min lunch and two 15-min breaks). In the first experimental rest break schedule, workers were given 12 3-min breaks evenly distributed over the workday (3-min break for every 27 min of work). In the second schedule, workers were given four 9-min breaks evenly distributed over the workday (9-min break every 51 min of work). Results showed that neither of the two experimental rest break schedules had a negative effect on production, and the 9-min break schedule improved discomfort ratings for the lower extremities. The workers in the study mostly preferred the 9-min rest break schedule, indicating that workers in general might not as readily accept fragmentation of break time into short, frequent breaks. With respect to work breaks within a sift, it is generally accepted that additional short breaks (5–10 min each hour) are beneficial for physical as well as mental work (Graf 1954), resulting in improved physical and mental well-being and reduced discomfort together with unchanged or even improved performance (Galinsky et al. 2000; Debabneth et al. 2001; McLean et al.

2001; Faucett et al. 2007).

The past and current literature seems to support frequent and short rest pauses at work. It is worth mentioning here that the main problem with introducing these pauses is that there may be a disruption from tasks, resulting in loss of adaptation to work (Rohmert 1965; Henning et al. 1989). On the other hand, in his fundamental work on fatigue, Simonson (1971) argues that the adaptation to work is not disrupted if 10 minute pauses are introduced.

3. Rest breaks with active pauses – Prekidi rada s aktivnim odmorom

Fig. 1 Effects of short breaks on net working time (Modified from Graf 1954) Slika 1. Utjecaj kratkih odmora na efektivno radno vrijeme (prilagođeno prema Graf 1954)

106

Active pauses have been used, particularly by Russian workers, to reduce fatigue and to maintain work performance (Simonson 1971). The favorable results of pauses in occupational work are in large measure due to the effect of active pauses (Simonson 1971; Genaidy et al. 1995).

Marschak (1933), in his pioneer historical work, found that for restoring work capacity after fatigue of finger muscles (load 4–5 kg) or hand muscles (load 9–10 kg), the effect of active pauses of thirty seconds, with intermittent moderate work of other or the same muscle groups without load, was better than the effect of passive pauses ranging from 18.5 to 51%. Muratov (1967) reported that active pauses had a favorable effect on the reaction of pulse rate and blood pressure, Croat. j. for. eng. 34(2013)1

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

particularly in older people. Gisolfi et al. (1966) in their experiments, with the subjects running at 19.3–20.9 km/hour at 2.0–2.5 degrees to exhaustion on the treadmill, and with recovery from thirty-five to fifty minutes with running at 6.4–10.5 km/hour at 0 degree to 5 degrees grade, found that O2 debt was decreased by 1 to 2 litters. The rate of lactate removal was considerably faster when aerobic work was performed during recovery, as compared with values observed when the subjects rested during recovery. It appears that the favorable effect of active pauses on recovery may also be found in strenuous work at or near maximum VO2.

4. Prescribed work/rest schedules for forest operators and the application of Electromyogram – Propisani rasporedi rada i odmora rukovatelja šumskim strojevima i primjena elektromiograma Theoretical examples of managed work/rest scheduling for forest operations are illustrated in Figs. 2 and 3. Due to the advantages of having short breaks distributed through the day, a work/rest schedule involving more frequent short breaks might be proposed.

Also, a set policy of prescribed pauses, as supported

Fig. 2 Current work/rest schedule for forest operators Slika 2. Postojeći raspore rada i odmora za rukovatelje šumskim strojevima

Fig. 3 Proposed work/rest schedule for forest operators Slika 3. Predloženi raspored rada i odmora za rukovatelje šumskim strojevima Croat. j. for. eng. 34(2013)1

C. Gallis

by Janaro and Bechtold (1985), increases the work output and total rest time compared to when workers control their own rest-work schedule.

The proposed work/rest schedule would consist of an approximate ten-minute active rest pause after every fifty minutes of work. The ten-minute rest break will help the operators to recover from fatigue effects (Rohmert 1965; Simonson 1971), to maintain work capacity, performance and efficiency, and to maintain their adaptation to work (Rohmert 1973; Sundelin and Hagberg 1989; Gallis 1992).

Also, ten-minute rest pauses may increase the net working time because workers may avoid taking spontaneous and disguised pauses (Graf 1954; Grandjean 1988). It will also decrease exposure time to vibration, noise, and climate effects. Active pauses during working time, on the other hand, will improve the restoration of work capacity, and may have favorable effects on localized muscle fatigue (Genaidy et al. 1995; Gisolfi et al. 1966; Laporte 1966; Marschak 1933; Simonosn 1971).

With this proposed work/rest schedule, workers would have more meal breaks as well as the lunch breakdown into two ten-minute periods (Fig. 3). Several meal breaks would help the workers to avoid overloading their digestive organs from having only one heavy meal by spreading food intake during the day (Gallis 1992).

The effects of this proposed work/rest schedule on the physiological parameters of workers’ bodies, and on net working time and productivity would be evaluated by future research. Time studies are a valuable and useful tool to determine productivity and net working time (Magagnotti and Spinelli 2012). The use of an electromyogram (EMG) from the muscle group that is involved in a task may be a useful tool to indicate muscular fatigue and to determine appropriate design work/rest schedules (Gallis 1997; Kogi 1982; Sundelin 1993). Myosignals are the »picture« of the physiological status of muscles, both for fatigue and reco very. The electromyography could be used as a tool to determine the duration of work pauses for each in dividual operator working at a specific task. The shift of the distribution of EMG signal power from a high frequency band toward low frequencies can be used to monitor muscle fatigue (Lindstrom et al. 1977; Basmajan and DeLuca 1985). Since median and mean frequencies have been observed to increase toward their initial values during rest, the recovery time can be indicated by the EMG (Sabbahi et al. 1979; Merletti et al. 1984).

In mechanized forest operations, EMG has been used to define risks for musculoskeletal disorders

107

C. Gallis

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

among forest machine operators. Østensvik et al.

(2009) used surface electromyography to study the association between numbers of long periods of sustained low-level trapezius muscle activity and neck pain among forest machine operators by using EMG. The results of this study show that the exposure to several long periods (> 8 min) with sustained lowlevel muscle activity in the right upper trapezius muscle increase the risk of pain in the neck and shoulder.

The authors of the study argue for a possible causal relationship between these long periods and neck pain. In another study, Østensvik et al. 2008 recorded EMG of the right upper trapezius and extensor digitorum muscles during a whole working day (7.5–8 h per operator). The study was performed to analyze potential risk factors for upper extremity disorders in two groups of forest machine operators driving harvesting vehicles and performing equal tasks in France and Norway. In Finland, Asikainen and Harstela (1993) studied the Influence of Small Control Levers of forest grapple loader on muscle strain of forest operators. Electromyograms (EMG) of the trapezius muscle were analyzed using the root mean square (RMS – EMG) method. The trapezius muscle was chosen as the object of the study because it is the biggest muscle in the shoulder-arm system and the pain reported by operators focusing on this system is, generally, the most severe in this muscle. The EMG value for the maximum voluntary constriction was measured every morning and the results were calculated as relative values of these maximum values (%MVC). Working with mini levers, as compared to conventional levers, leads to lesser muscle constriction in the trapeziums muscle as measured by RMS – EMG. The use of EMG may be limited by the fact that muscles, other than those EMG is recording, may undergo great strain. Thus, there may be some practical limits on monitoring all the muscles involved in field forest tasks (Gallis 1997). EMG should, therefore, be accompanied by the use of the multidimensional Checklist Individual Strength questionnaire (CIS). The application of CIS has been proved to be able to discriminate between fatigued and non-fatigued employees. The CIS seems to be an appropriate instrument for measuring the overall fatigue of an operator (Beurskens et al. 2000). The CIS questionnaire 7–11 was used to measure chronic fatigue and it consists of four dimensions: the subjective experience of fatigue and reduction in motivation, reduction in activity, and reduction in concentration. This would be the methodology for determining the validity and assessing the results of the above mentioned work/rest schedules in mechanized and motor-manual forest operations.

108

5. Conclusions – Zaključci It is concluded that, despite some concerns about the manner of the integration of short pauses into the task, the literature supports the idea that short active pauses at work enhance work productivity, comfort, and health, if they are managed and well integrated into the task. Thus, managing work fatigue by using prescribed work/rest schedules would relieve operators from fatigue effects and increase their work performance. Net working time may be increased by avoiding worker disguised and spontaneous pauses. In forest operations, ten minute active pauses may provide time for recovery, maintain adaptation to work, and restore work capacity. Active pauses may be an effective method of workload reduction and prevention of musculoskeletal injuries and discomfort among forest operators.

The fact that it may not be at the expense of productivity is also an important issue for forest and logging managers, supervisors and crew bosses.

Further research should be focused on determining the type of active pauses (i.e. manner and rate of gymnastics, stretching, time for lunches) required for each specific task and workload in various types of forest work.

6. References – Literatura Asikainen, A., Harstela, P., 1993: Influence of Small Control Levers of Grapple Loader on Muscle Strain, Productivity and Control Errors. International Journal of Forest Engineering 5(1): 23–28.

Astrand, P-O., Rodahl, K., 1986: Textbook of work physiology. 3rd ed. McGraw-Hill.

Axelsson, S-A., Porten, B., 1990: New ergonomic problems in mechanized logging operations. International Journal of Industrial Ergonomics 5(3): 267–274.

Barker, K. L., 1995: Repetitive strain injury – a review of the legal issues. Physiotherapy 81(2): 103–106.

Basmajian, J. V., DeLuca, C. J., 1985: Muscles Alive: Their functions revealed by electromyography. Williams and Wilkins, Baltimore, 561 p.

Beurskens, A. J., Bultmann, U., Kant, I., Vercoulen, J. H., Bleijenberg, G., Swaen, G. M., 2000: Fatigue among working people: validity of a questionnaire measure. Occupational and Environmental Medicine 57(5): 353–357.

Beynon, C., Burke, J., Doran, D., Nevill, A., 2000: Effects of activity-rest schedules on physiological strain and spinal load in hospital-based porters. Ergonomics 43(10): 1063– 1770.

Bovenzi, M., Alessandrini, B., Mancini, R., Gannava, M. G., Centi, M., 1998: A prospective study of the cold response of digital vessels in forestry workers exposed to saw vibration.

Croat. j. for. eng. 34(2013)1

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112) International Archives of Occupational and Environmental Health 71(7): 493–498.

Byers, J., 1997: Operators Health-A survey of feller-buncher operators. LIRO report 22(17): 1–6.

Byrne, E., 2000: Aetiological considerations on some conditions in the borderlands of neurology: chronic fatigue syndrome, pan allergy syndrome and repetitive strain injury, a personal view. Journal of Cilnical Neuroscience 7(1): 9–12.

Chaffin, D. B., 1973: Localized muscle fatigue – definition and measurement. Journal of Occupational Medicine 15(4): 346–354.

Dababneh, A. J., Swanson, N., Shell, R. K., 2001: Impact of added rest breaks on the productivity and well being of workers. Ergonomics 44(2): 164–174.

Dul, J., Douwes, M., Smitt, P., 1991: A work-rest model for static postures. In: Y. Queinnec and F. Daniellou (eds.) Designing for Everyone: Proceedings of the 11th Congress of the International Ergonomic Association, vol. 1. Taylor and Francis. New York, p. 93–95.

Ferguson, D., 1976. Posture, aching and body build in telephonists. Journal Human Ergology 5(2): 183–186.

Fucett, J., Meyers, J., Miles, J., Janowitz, I., Fathallah, F., 2007: Rest break interventions in stoop labor tasks. Applied Ergonomics 38(2): 219–226.

Gallis, C., 1992: Work Fatigue in Forest Operations: Causes and Management. M.Sc. thesis. University of New Brunswick. Canada. 79 p.

Gallis, C., 1997: Determining work/rest schedules for forest workers using electromyography. In: Seminar and Workshop Proceedings on Safety and Health in Forestry are Feasible, Joint FAO/ECE/ILO Committee on Forest Technology, Management and Training, Konolfingen, 7–11 October, Switzerland, p. 326–331.

Gallis, C., 2006: Work-related prevalence of musculoskeletal symptoms among Greek forest workers. International Journal of Industrial Ergonomics 36(8): 731–736.

Gallinsky, T. L., Swanson, N. G., Sauter, S. L., Hurrell, J. J., Schleifer, L. M., 2000: A field study of supplementary rest breaks for data-entry operators. Ergonomics 43(5): 622–638.

Gellerstedt, S., 1993: Work and health in forest work – A multivariate study of cutters and machine operators. (English summary). Thesis, Department of Operational Effeciency, College of Forestry, Garpenberg, Sweden.

Genaidy, A. M., Delgado, E., Bustos, T., 1995: Active microbreak effects on musculoskeletal comfort ratings in meatpacking plants. Ergonomics 38(2): 326–336.

C. Gallis

Graf, O., 1954: Studien uber Fliebarbeitprobleme an einer praxisnahen Experimentieranlage (Studies on work problems with a practical experiment in a plant). Forschungsbericht d. Wirtschaft und Verkehrsministerium NordrheinWestfalen, Nos 114 and 115. Westdeutscher Verlag, Cologne.

Grandjean, E., 1988: Fitting the task to the man. 4th ed. Taylor and Francis. London.

Hagen, K. B., 1997: New troubles for old? Health and safety hazards in highly mechanized operations. In: Seminar and Workshop Proceedings on Safety and Health in Forestry are Feasible, Joint FAO/ECE/ILO Committee on Forest Technology, Management and Training, Konolfingen, 7–11 October, Switzerland, p. 153–171.

Harstela, P., 1990: Work posture and strain of workers in Nordic forest work: a selective review. International Journal of Industrial Ergonomics 5(3): 219–226.

Henning, R. A, Sauter, S. L., Salvendy, G., Krieg, E. E., 1989: Microbreak length, performance, and stress in a data entry task. Ergonomics 32(7): 855–864.

Henning, R. A., Jacques, P., Kissel, G. V., Sullivan, A. B., Alteras-Webb, S. M., 1997: Frequent short breaks from computer work: effects on productivity and well-being at two field sites. Ergonomics 40(1): 78–91.

Herberts, P., Kadefors, R., Broman, H., 1980: Arm positioning in manual tasks: an electromyographic study of localized fatigue. Ergonomics 23(7): 655–665.

Janaro, R. E., Bechtold, S. E., 1985: A study of the reduction of fatigue impact on productivity through optimal rest break scheduling. Human Factors 27(4): 459–466.

Ising, H., Babish, W., Kruppa, B., Lindthammer, A., Wiens, D., 1999: Subjective work noise: a major risk factor in myocardial infarction. Soz Praventivmed 42(4): 216–22.

Karpovich, V. P., Sinning, E. W., 1971: Physiology of muscular activity. 7th ed. W.B. Saunders Co. Philadelphia.

Kirk, P., 1998: The impact of shift length on processor operator fatigue. Liro report 23(18): 1–11.

Kogi, K., 1982: Finding appropriate work-rest rhythm for occupational strain on the basis of EMG and behaviour changes. In: Buser, P.A., Cobb, W.A., Okuma, T., (eds) Kyoto Symposium: Electroencephalography and clinical neurophysiology, Elsevier. Amsterdam. Biomedical Press pp.

738–749.

Koparadekar, P., Mital, A., 1994: The effect of different workrest schedules on fatigue and performance of a simulated directory assistance operator’s task. Ergonomics 37(10): 1697–1707.

Giannini, F., Rossi, S., Passero, S., Bovenzi, M., Mancini, R., Ciobi, R., Battistini, N., 1999: Multifocal neural conduction impairment in forestry workers exposed and not exposed to vibration. Clinical Neurophysiology 110(7): 1276–1283.

Laporte, W., 1966: The influence of a gymnastic pause upon recovery following Post Office work. Ergonomics 9(6): 501– 506.

Gisolfi, C., Robinson, S., Turrell, E. S., 1966: Effects of aerobic work performed during recovery from exhausting work.

Journal of Applied Physiology 21: 1767–1781.

Lewis, G., Wessely, S., 1992: The epidemiology of fatigue: more questions than answers. J Epidemiol Community Health 46(2): 92–97.

Croat. j. for. eng. 34(2013)1

109

C. Gallis

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

Magnatotti, N., Spinelli, R., (Ed.), 2012: COST Action FP0902 – Good practice quideline for biomass production studies.

CNR IVALSA, Florence, Italy. 41 p.

Sundelin, G., Hagberg, M., 1989: The effects of different pause types on neck and shoulder EMG activity during VDU work. Ergonomics 32(5): 527–537.

Marschak, M. E., 1933: Experimentelle Untersuchungen über den Einfluss der aktiven Erholung auf die Arbeitsfähigkeit des Menschen (Experimental studies on the infuence of active pauses on human work ability). Arbeitsphysiologie 6(6): 664–672.

Sundelin, G., Hagberg, M., 1992: Effects of exposure to exessive drafts on myoelectric activity in shoulder muscles. Journal of Electromygraphy and Kinesiology 2(1): 36–41.

May, J., Kline, P., 1988: Problems in using an adjective checklist to measure fatigue. Journal Personality and Individual Differences 9(4): 831–832.

McLean, L., Tingley, M., Scott, R. N., Rickards, J., 2001: Computer terminal work and the benefit of microbreaks. Applied Ergonomics 32(3): 225–237.

Muranov, I., 1967: Einfluss der aktiven Erholung auf die Blutkreislauffunktion bei Menschen in verschiedenem Alter (Infuence of active pauses on blood circulation on humans of various ages). Z. Alterforsch 20(2): 99–106.

Nakata, M., Hagner, I. M., Jonsson, B., 1992: Perceived musculoskeletal discomfort and electromyography during repetitive light work. Journal of Electromyography and Kinesiology 2(2): 103–111.

Ong, C.N., 1992: Musculo-skeletal disorder, visual fatigue and psychological stress of working with display units: current issues and research needs. In: H.Luczak, A.Cakir and G. Cakir (eds) Work with Display Units 92: Proceedings of the third International Scientific Conference on work with Display Units.Technische Universitat Berlin, p. 221–228.

Onishi N., Sakai, K., Itani, T., Shindo, H., 1977: Muscle load and fatigue of film rolling workers. Journal Human Ergology 6(2): 179–186.

Porten, B., 1990: New ergonomic problems in mechanized logging operations. Paper presented at the XIX IUFRO World Congress, Montreal. P.Q. Canada.

Rohmert, W., 1973: Problems of determination of rest allowances. Part 2: Determining rest allowances in different human tasks. Applied Ergonomics 4(3): 158–162.

Rohmert, W., 1965: Physiologische Grundlagender Erholungszeitbestimmung (Physiological basis of determination of recovery-time pauses in physical work). Zbi. Arbeit. Wiss.

19(1): 1–7.

Simonson, E., 1971: Physiology of work capacity and fatigue.

Charles C. Thomas, Publisher, Springfield. Illinois.

Staaf, K. A. G., Wiksten, N. A., 1984: Tree Harvesting Techniques. Forest Science, Ser. Martinus Nijhoff/Dr W. Junk Publishers.

110

Sundelin, G., 1993: Patterns of electromyographic shoulder muscle fatigue during MTM–paced repetitive arm work with and without pauses. International Archives of Occupational and Environmental Health 64(7): 485–493.

Stock, S. R., Cole, D. C., Tugwell, P., Streiner, D., 1996: Review of applicability of existing functional status measures to the study of workers with musculoskeletal disorders of the neck and upper limb. American Journal of Industrial Medicine 29(6): 679–688.

Swanson, N. G., Sauter, S. L., Chapman, L. J., 1989: The design of rest breaks for video terminal work: a review of the literature. In A. Motal (ed.), Advances in Industrial Ergonomics and Safety I: Proceedings of the Industrial Ergonomics and Safety Conference. Taylor and Francis. London, p. 895– 898.

Veiersted, K. B., 1995: Stereotyped light manual work, individual factors and trapezius myalgia. The National Institute of Occupational Health, vol Dr. Med. Ph. D. The Univeristity of Oslo, Oslo: 7703–7011.

Veiersted, K. B., 1996: Reliability of myoskeletal trapezius muscle activity in repetitive light work. Ergonomics 39(5): 797–807.

Westgaard, R. H., DeLuca, C. J., 1999: Motor unit substitution in long duration contractions of the human trapezius muscle. J Neurophysiol 82(1):501–504.

Zwahlen, H. T., Hartmann, A. L., Rangarajulu, S. L., 1984: Effects of rest breaks in continuous VDT work on visual and musculoskeletal comfort/discomfort and on performance.

In: G. Salvendy (eds) Human-Computer Interaction, Elsevier Science, Amsterdam. Østensvik, T., Veiersted, K. B., Nilsen, P., 2009: A method to quantify frequency and duration of sustained low-level muscle activity as a risk factor for musculoskeletal discomfort. Journal of Electromyography and Kinesiology 19(2): 283–294. Østensvik, T., Veiersted, K. B., Cuchet, E., Nilsen, P., Hanse, J. J., Carlzon, C., Winkel, J., 2008: A search for risk factors of upper extremity disorders among forest machine operators: A comparison between France and Norway. International Journal of Industrial Ergonomics 38(11–12): 1017–1027.

Croat. j. for. eng. 34(2013)1

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

C. Gallis

Sažetak

Povećanje proizvodnosti i kontrola radnoga umora pri šumskim radovima primjenom propisanih aktivnih odmora: odabrani pregled U članku se donosi odabrani pregled postojeće literature o rasporedu rada i odmora te raspravlja o teoretskom modelu koji se može primijeniti u određivanju odmora pri radu za šumske radnike. Rukovatelji šumskim strojevima i ručno-strojnim alatima izloženi su mnogim čimbenicima koji mogu uzrokovati umor pri radu. Unatoč povećanju mehaniziranosti i automatizacije u velikom broju šumskih poslova još uvijek je potrebna mišićna snaga i aktivnost za rukovanje mnogim ručnim alatima i strojevima. Mišićna se snaga ne može održati kod dugotrajnoga rada jer se pojavljuje umor i smanjuje snaga mišića. U praksi rukovatelji šumskim strojevima i ručno-strojnim alatima rade vrlo intenzivno tijekom radnoga dana, s ograničenim brojem odmora, i izloženi su raznovrsnim radnim uvjetima koji djeluju na njihovu radnu sposobnost i koji mogu uzrokovati umor, zdravstvene probleme, mišićno-skeletne probleme, slabu učinkovitost, učestalije pogreške pri radu, ozljede i nesreće. Čak ako se i ne smanjuje radni učinak, subjektivan je osjećaj umora važan aspekt ljudskoga rada. Umor je objektivna nemogućnost stjecanja snage i izvođenja rada, što se može izmjeriti elektrofiziološki. Umor je normalna svakodnevna pojava, no u slučaju ozbiljnoga i teškoga umora može utjecati na čovjekovu izvedbu i u profesionalnom i u privatnom okruženju. Štoviše, ozbiljan dugotrajan umor uzrok je bolovanja i radne nesposobnosti. S umorom povezani mišićno-skeletni poremećaji vrata i gornjih udova i dalje su u središtu interesa šumskih radnika, šumskih organizacija i istraživača. Tomu pridonosi značajna nesposobnost rada, gubitak radnoga vremena, povećanje naknada radnicima i sve veći broj s tim povezanih sudskih postupaka. Čimbenici koji djeluju na umor i umanjuju radni učinak mogu se svrstati u tri kategorije: okolišni čimbenici (buka, vibracije, temperatura, vlaga), ljudski čimbenici (dob, spol, zdravlje, motivacija, sposobnosti, utreniranost, emocionalna stabilnost) i čimbenici zadatka (složenost, trajanje, fizički, mentalni ili senzomotorički zadaci). Utjecaj svakoga okolišnoga čimbenika ovisi o stupnju u kojem odstupa od prihvatljive razine i trajanju njegova djelovanja.

Također su moguće bezbrojne kombinacije svih okolišnih čimbenika, a situacija postaje još složenija s interakcijom ljudskih, okolišnih i čimbenika zadatka.

Odmori pri radu mogu pružiti vrijeme potrebno za fizički oporavak i za smanjenje zdravstvenih rizika. Pri fizičkom radu čovjeku su nužna odgovarajuća razdoblja odmora radi oporavljanja od fizioloških posljedica radnoga napora.

Prekidi rada, tj. odmori, omogućuju zadržavanje traženoga stupnja izvršenosti i učinkovitosti u radu te izbjegavanje ili umanjenje utjecaja umora. S tim u vezi ciljevi su ovoga rada: Þ dati pregled prošle i sadašnje literature, Þ naglasiti i raspraviti važnost prekida rada uvođenjem ispravno oblikovanoga rasporeda rada i odmora, Þ prikazati teoretski model koji se može primijeniti u određivanju rasporeda rada i odmora za šumske radnike.

Obično se pri radu uzimaju različiti oblici odmora i pod različitim okolnostima: propisani odmori, koje je prekide rada utvrdilo vodstvo; spontani odmori, koji su očiti prekidi rada koje radnici uzimaju na vlastitu inicijativu; prikriveni odmori, koji predstavljaju trenutke kada se radnici bave lakšim zadacima; radom uvjetovani odmori, koji predstavljaju prekide rada zbog šumskih operacija, strojeva ili organizacije rada; aktivni odmori, koji su kratki prekidi rada tijekom kojih radnici nisu pasivni već koriste druge mišićne grupe ili iste mišićne grupe s lakšim opterećenjem ili bez opterećenja, ili hodaju ili obavljaju gimnastičke vježbe. Prikriveni, aktivni i spontani odmori često se koriste da bi radnici izbjegli pretjerano umaranje i oslobodili se simptoma umora. Glavni razlozi uzimanja odmora tijekom rada su: uspostava ravnoteže kisika i hraniva u mišićima, odstranjivanje otpadnih tvari, posebno mliječne kiseline, izbjegavanje niske razine šećera u krvi. Dodatno, odmori omogućuju brzi povratak koncentracije krvi u nogama i stopalima, smanjenje krvnoga tlaka u venama, nadoknadu gubitka vode hlađenjem tijela, ograničavanje izloženosti vibracijama, buci, vlazi, temperaturi, razbijaju monotoniju rada, omogućuju socijalne kontakte i dr.

U literaturi se podržavaju česti i kratki prekidi rada. Pri uvođenju takvih odmora kao glavni problem mogu se pojaviti poremećaji u zadatku koji rezultiraju gubitkom prilagodbe na rad. S druge strane, smatra se da se pri uvođenju kratkih 10-minutnih odmora ne narušava prilagodba na rad. U članku se predlaže takav raspored rada i odmora koji bi se sastojao od približno 10-minutnih aktivnih odmora nakon svakih 50 minuta rada. Odmor od 10 minuta pritom pomaže da se rukovatelj oporavi od djelovanja umora, da zadrži radni kapacitet, stupanj izvedbe i učinkovitost te da zadrži svoju prilagodbu na rad. Također, 10-minutni bi odmori mogli povećati efektivno radno vrijeme jer radnici mogu izostaviti spontane i prikrivene odmore. Isto se tako smanjuje trajanje izloženosti vibracijama, buci i klimatskim učincima. S druge strane, aktivni će odmori poboljšati obnavljanje radnoga kapaciteta i mogu imati povoljan učinak Croat. j. for. eng. 34(2013)1

111

C. Gallis

Increasing Productivity and Controlling of Work Fatigue in Forest Operations... (103–112)

na lokalizirani mišićni umor. S predloženim rasporedom rada i odmora radnici dobivaju više pauza za obroke te raspodjelu ručka u dvije 10-minutne pauze. Više prekida za obroke tijekom dana može pomoći radnicima da ne opterete svoj probavni sustav kao u slučaju jednoga teškoga obroka.

Učinke predloženoga rasporeda rada i odmora na fiziološke parametre radnika i na efektivno radno vrijeme i radnu učinkovotost potrebno je ispitati u budućim istraživanjima. Primjena elektromiografije (EMG) popraćena s »kontrolna lista individualne snage« (Checklist Individual Strength questionnaire) pritom se smatra odgovarajućom metodologijom za utvrđivanje valjanosti i procjenu rezultata navedenoga rasporeda rada i odmora pri mehaniziranim i ručno-strojnim šumskim radovima. Daljnja se istraživanja također trebaju usmjeriti na određivanje vrste aktivnih odmora (oblik i stupanj vježbi, istezanje, vrijeme obroka) za svaki specifični zadatak i radno opterećenje u različitim vrstama šumskoga rada.

Ključne riječi: umor pri radu, zdravstveni problemi i ozljede, elektromiografija, proizvodnost, prekidi rada, rukovatelji šumskim strojevima, raspored rada i odmora

Authors’ address – Autorova adresa:

Received (Primljeno): July 20, 2012 Accepted (Prihvaćeno): November 25, 2012

112

Christos Gallis, PhD. e-mail: cgalis@fri.gr Forest Research Institute GR-57 006, Vassilika Thessaloniki GREECE Croat. j. for. eng. 34(2013)1

Original scietific paper – Izvorni znanstveni rad

Research of Exhaust Emissions from a Harvester Diesel Engine with the Use of Portable Emission Measurement System Piotr Lijewski, Jerzy Merkisz, Paweł Fuć Abstract – Nacrtak This paper presents the test results of exhaust emissions from the engine of a wheeled harvester. On the basis of the present analysis, the ranges of most frequently used engine loads and speeds were determined. The obtained time density characteristics (distribution of engine speeds and loads in time) of the engines were referred to the measuring points of the exhaust emission homologation test. During the measurement of the exhaust emissions, the machine was in the forest engaged in tree cutting. Exhaust emission tests from non-road vehicle engines are currently performed on engine test beds in the NRSC and NRTC tests. The present methodology is a new solution that could be used in the future as a basis for the development of harvesting technology aimed at reducing exhaust emissions from engines. During the tests, the authors also measured the amount of timber harvested during the felling process in order to be able to relate the emission results to the amount of the generated product. Harvesters are more and more often used in forest operations but they still have to compete with handheld (gasoline powered) chainsaws and that is why in this paper the authors included a comparison of the exhaust emissions from a chainsaw with the emissions from the engine of a harvester.

The authors used a portable analyzers (SEMTECH DS and LAM manufactured by SENSORS) for the measurement of the exhaust emissions. The said analyzer can measure the concentrations of exhaust gas components in an on-line mode, while the engine is running under field conditions.

Keywords: exhaust emissions, harvester, chainsaw, on-board measurement

1. Introduction – Uvod Exhaust emissions are one of the most important issues related to transport and industry. For many years, the impact of the environment pollution (including that coming from transport) on climatic changes have been discussed. The negative impact of exhaust emissions on human health and human natural environment, however, is beyond doubt. When harvesting timber, despite the obvious influence of exhaust emissions on the environment (atmosphere, flora, etc.), there is a particular hazard that chainsaw operators are exposed to when using motor chainsaws. Chainsaw operators work very often in an environment with a high concentration of hazardous exhaust components. Chainsaw operators are exposed to the emission of PAH and nanoparticles (CzerwinCroat. j. for. eng. 34(2013)1

ski et al. 2001; Laanti et al. 2001; Jacke et al. 1996). Despite recently observed changes in the European Union member states (Central and Eastern Europe in particular) related to the mechanization of forest operations and use of harvesters, still most of the work is done with the use of motor chainsaws (equipped with gasoline engines).

The literature related to the problem of exhaust emissions generated during tree harvesting is chiefly based on test stand research performed in laboratories and estimates under actual operating conditions.

There are numerous works aimed at implementing technological solutions in order to reduce exhaust emissions and fuel consumption such as the use of catalytic aftertreatment (Schlossarczyk et al. 2004), improvement of the charge exchange (Rodenbeck et al.

113

P. Lijewski et al.

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

2006) or application of the stratified scavenging method (Ohtsuji and Kobayashi 2002). A wide variety of technical solutions for chainsaw engines was presented by Zahn (2000), who investigated fuel injection, engine lubrication, stratified scavenging and catalytic aftertreatment. Another ecological aspect related to tree harvesting is the emission of lubricating oil. According to Hartweg and Keilen (1988) while harvesting trees, the soil may absorb up to 0.2 dm3 of oil from the engine per 1 cubic meter of harvested timber. Similar values are given by Sonnleitner (1992). This is an indication of a serious ecological problem, and hence there are numerous papers on this issue. One of the proposals to solve the problem is the application of biodegradable oils and hydraulic fluids (Ahola 1998; Wightman et al. 1998). A comparison of the emission level and the operating parameters of the chainsaw engines using mineral and vegetable oils have been presented by Skoupy et al. (2010). There are also numerous publications related to the estimation and calculation of exhaust emissions during machine tree harvesting. In many cases, particular stress was put on the emission of GHG. In his publication, Athanassiadis (2000) presented the results of the estimation of energy requirements and the amount of emissions from chainsaw engines and heavy-duty diesel vehicles used for timber harvesting. In this paper, the emissions of CO2, CO, HC, NOx and PM were estimated related to 1 cubic meter of harvested timber. The tests were performed for diesel fuels and RME and for mineral and vegetable lubricating oil. Results of similar investigations were presented by Berg and Karjalainen (2003) and Karjalainen and Asikainen (1996). In their publications, they presented the emission of GHG and fuel consumption from machinery used in silviculture, forest improvement work, wood harvesting and timber transportation. In these works, the authors also presented a comparison of fuel consumption while harvesting timber with chainsaws and harvesters. The problem of fuel consumption and exhaust emissions was also addressed by Nordfjell et al. (2003) and Klvac et al. (2012). These works, however, are related to forwarders and cableway systems.

The investigations presented in this paper are slightly different. The majority of publications related to the exhaust emissions is based on calculated or estimated values (Athanassiadis 2000; Klvac and Hosseini 2010; Klvac et al. 2012; Nordfjell et al. 2003; Skoupy et al. 1996 and many others). In the present paper, the emission measurements were performed while the harvester was working. The fuel consumption was determined based on direct measurements using the carbon balance method – the most accurate method of fuel consumption measurement.

114

The investigations presented here have been carried out with a method that uses a PEMS analyzer. The tests under actual operating conditions are one of the latest technologies in exhaust emission measurement (Merkisz et al. 2010). Thanks to the above-mentioned method, the authors could determine the actual emissions generated during the operation of a harvester and a chainsaw. The tests related to the exhaust emissions are usually carried out in laboratories, on chassis or engine dynamometers. The results of these tests may determine the emission class of the engine or the vehicle but do not entirely reflect the actual emissions.

The measurements of the exhaust emissions under actual operating conditions are very valuable as they enable the determination of the emission depending on the existing engine operating conditions (a machine or a vehicle). Consequently, this methodology is now generally recognized as a very important complementing component of tests performed in laboratories.

Furthermore, scientific research centers or legislative bodies consider introducing the on-road exhaust emission testing of homologation procedures (Walsh 2011).

2. Testing methods and equipment Ispitne metode i oprema The tests were performed on a harvester equipped with a diesel engine. The technical specifications are Table 1 Basic technical specifications of the harvester Tablica 1. Osnovne tehničke značajke harvestera Capacity/number of cylinders

6.8 liters/6 cylinders in-line

Kapacitet/broj cilindara

6,8 litara/6 cilindara, linijski

Maximum power output

129 kW @ 2000 rpm

Najveća snaga

129 kW pri 2000 min-1

Maximum torque

780 Nm @ 1300–1400 rpm

Najveći zakretni moment

780 Nm pri 1300–1400 min-1

Cooling system

Coolant forced circulation

Rashladni sustav

Vodeno hlađenje

Compression ratio Omjer kompresije

17:1

Supercharging

Turbocharged

Prednabijanje

Prednabijanje plinskom turbinom

Injection system Sustav ubrizgavanja Engine emission regulations Norma ispušnih plinova

Common rail Ubrizgavanje sa stalnim tlakom Stage III A/Tier 3

Croat. j. for. eng. 34(2013)1

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

Table 2 European (Stage) and American (Tier) exhaust emission limits for the harvester engine Tablica 2. Europske (Stage) i američke (Tier) granične vrijednosti emisije ispušnih plinova za motor istraživanoga harvestera CO, g/kWh

HC, g/kWh

NOx, g/kWh

PM, g/kWh

Stage IIIA 5.0

NOx + HC – 4.0

0.3

5.0

0.19

3.3

0.025

Tier 3 5.0

NOx + HC – 4.0 3.54

Engine displacement Obujam motora Power Snaga Preporučena najveća duljina vodilice Sound power level

0.3

Tier Interim 4 5.0

Table 3 Basic technical specifications of the chainsaw Tablica 3. Osnovna tehničke značajke motorne pile lančanice

Recommended max. bar length

Stage IIIB

0.02

Razina buke Weight (without bar/chain) Masa (bez vodilice i lanca) Engine emission regulations

shown in Table 1. The engine of the tested harvester complies with the Stage IIIA (US Tier 3) standard. The testing methodology in both European and American exhaust emission regulations is the same. There are two types of tests, the stationary NRSC and the dynamic NRTC (Fig. 1). In section 3, the authors compared the exhaust emissions obtained under actual operating conditions with the Stage IIIA and Tier 3 limits because the tests should reflect most accurately the actual conditions of engine operation. The authors also compared the exhaust emissions from the harvester and chainsaw. The basic technical specifications

P. Lijewski et al.

Norma ispušnih plinova

50.2 cm3 2.5 kW 45 cm 112 dB(A) 5.2 kg Stage II

of the chainsaw are shown in Table 2. At the moment of testing, the chainsaw was used for approximately 6 months.

The measurements were done under actual operating conditions of the machine while felling trees, delimbing and cross-cutting tree trunks into 2 meter logs.

The harvester prepared for testing with the measurement equipment during forest operations is presented in Fig. 2. The works were performed in a pinewood

Fig. 1 Exhaust emission tests (harvester engine) Slika 1. Ispitivanja emisije ispušnih plinova (motor harvestera) Croat. j. for. eng. 34(2013)1

115

P. Lijewski et al.

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

Fig. 2 Forest harvester with the exhaust emission measuring equipment Slika 2. Izgled harvestera opremljenoga mjernom opremom za mjerenje emisije ispušnih plinova forest. The diameter of trees cut ranged approximately between 20 and 30 cm. The works were performed on a relatively small area of approximately 1 ha but during the tests the vehicle covered a distance of 2.3 km.

The location and the trajectory of the harvester are shown in Fig. 3. It took 4 hours to complete the whole testing cycle of the machine operation. A similar work was performed during the exhaust emission tests of a chainsaw (tree felling, delimbing and cross-cutting).

The authors made sure that the felled trees were pines.

In order to measure the concentration of exhaust emissions, a portable exhaust emissions analyzer SEMTECH DS by SENSORS was used (Fig. 4, Shahinian 2007). The analyzer measures the concentration of the exhaust components and simultaneously measures the

flow rate of the exhaust gases. The exhaust gases are introduced into the analyzer through a probe maintaining the temperature of 191oC. Then the particulate matter is filtered out (diesel engines) and the exhaust is directed to the flame-ionizing detector (FID), where HC concentration is measured. The exhaust gases are then cooled down to the temperature of 4oC and the measurement of concentration of NOx (NDUV analyzer), CO, CO2 (NDIR analyzer) and O2 follows in the listed order. It is possible to add data sent directly from the vehicle diagnostic system to the central unit of the analyzer and make use of the GPS signal (distance accuracy 0.05 % – <50 cm per km, velocity accuracy 0.2 km/h, averaged over 4 samples). The GPS signal is mainly used in the tests of on-road vehicles. In investigations

Fig. 3 Testing area and harvester route during testing (created at gpsvisualizer.com) Slika 3. Mjesto istraživanja i kretanje harvestera tijekom ispitivanja (gpsvisualizer.com)

116

Croat. j. for. eng. 34(2013)1

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

Fig. 4 A diagram of a portable analyzer SEMTECH DS; exhaust gas flow channels (arrow) and electrical connections circled (line) Slika 4. Dijagram prijenosnoga analizatora SEMTECH DS, protok ispušnih plinova (strelica) i električne veze (linija) Table 4 The specification of SMTECH-LAM Tablica 4. Specifikacija SMTECH-LAM-a Measurement Range

0–40 mg/m3

Raspon mjerenja

0–700 mg/m3

Dilution

Selective Ratio

Razrjeđivanje

Selektivni omjer

Particle Size Veličina čestica Resolution Preciznost

100–10 000 nm

<0.25mg/m3 over 6 hours

Odstupanje

<0.25mg/m3 tijekom 6 sati

Brzina protoka uzorka Sample rate Frekvencija uzorkovanja Output Izlaz

5 Hz (interno do 100 Hz) RS232 Analog 0 to 5 VDC option RS232 Analogni 0 do 5 VDC opcijski

Napajanje

12 do 24 VDC ili 110 do 240 VAC 0–40 °C

similar to the ones presented here, when the vehicle operates in a limited area, on-road emissions are not measured and GPS is of secondary importance. Measurements of emissions were made in test and, for the Croat. j. for. eng. 34(2013)1

The performed tests, apart from the possibility of determining ecological properties, enable an analysis of engine parameters under operation (Fig. 5). In the

5 Hz (internally to 100 Hz)

12 to 24 VDC or 110 to 240 VAC

Radna temperatura

3. Results of tests and discussion Rezultati ispitivanja i rasprava

1.5 dm3/min

Power supply Operating temperature

purpose of comparison, signals from the on-board diagnostic system were recorded, e.g. engine speed, load, vehicle speed, temperature of intake air. Some of these signals served to specify time density maps presenting participation of operating time of the vehicles in actual operation conditions.

For PM measurement, the SEMTECH Laser Aerosol Monitor (LAM) was used. The SEMTECH-LAM operates on the basis of laser light scattering and provides the concentration of fine particulate matter in the exhaust in real time. With two selectable ranges and variable dilution ratios, the analyzer is compatible with a variety of different motor types, vehicles and test conditions. It could be used as test cell equipment or for on road testing. The dual sample port enables testing of filter efficiency, for engines equipped with a diesel particulate filter. The SEMTECH-LAM uses three mass flow controllers that are automatically adjusted to maintain the desired dilution ratio. A pump pulls the diluted sample through the monitor. The monitor contains a laser light scattering analyzer that measures fine particulates from 100 nm to 10 000 nm. The specification of SEMTECH-LAM are shown in Table 4.

0.01 mg/m3

Drift Sample flow

P. Lijewski et al.

Fig. 5 Time density of harvester engine during the test under actual operating conditions Slika 5. Udjel brzina i opterećenja motora tijekom istraživanja harvestera

117

P. Lijewski et al.

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

Fig. 6 Exhaust emissions during harvester operation overlain on the map (created at gpsvisualizer.com) Slika 6. Kartografski prikaz emisija ispušnih plinova tijekom rada harvestera (gpsvisualizer.com) case of machinery tested, the engine operated in a very narrow range of engine speeds and namely 1700–1800 rpm. The most frequently used range of loads was 20–50% of the maximum engine load. The range of 50–70% had a slightly lower share. Besides the above shares of engine operation, the share of idle speed was also recorded (3%). The main engine operation range was outside the range of NRSC phases, as is in other non-road vehicles (Merkisz and Lijewski 2010; Noren and Pettersson 2001).

of pollution generated by harvesters operating on large forested areas. A similar analysis was performed for fuel consumption/CO2 emission. During harvester tests lasting over 4 hours the fuel consumption was 11.3 dm3/h, which was 0.69 dm3 related to 1 cubic meter of timber harvested. The emission of CO2 was 1810 g related to 1 cubic meter of timber harvested. Other authors have obtained slightly different values related to the evaluation of fuel consumption by harvesters.

Athanassiadis (2000) recorded fuel consumption of

Fig. 6 presents graphs showing the exhaust emissions overlain on a map with the marked vehicle route during operation. These graphs are only a 15-minute portion of the test. The authors decided to present only a fragment of the route because the presentation of the whole test would result in an excessive density of lines on the map, which would render the graphs illegible.

The graphs clearly show an increased emission of all the exhaust gas components in the spots where the harvester felled and processed trees as compared to vehicle idling. The increased emission probably results from the changes in the engine load while felling trees.

Fig. 7 presents the exhaust emissions from a harvester under actual operating conditions. The obtained results were compared to the limits of Stage IIIA and Tier 3. The emission of CO and HC is lower than the values set forth in the Stage IIIA and Tier 3 standards.

The emission of PM and NOx during the harvester operation is much higher though (almost three times higher than the limits set forth in Stage III).

The authors also performed an analysis of the exhaust emissions related to harvester work – the amount of timber harvested. Fig. 8 presents the exhaust emissions related to 1 cubic meter of timber harvested. Such an analysis enables a global assessment

118

Fig. 7 Exhaust emissions from a forest harvester under actual operating conditions compared to Stage IIIA and IIIB standards (based on AVL Current and Future Exhaust Emissions Legislation 2007) Slika 7. Usporedba emisije ispušnih plinova harvestera u radnim uvjetima s razinama propisanim normom Croat. j. for. eng. 34(2013)1

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

Fig. 8 Exhaust emissions from a forest harvester under actual operating conditions related to 1 cubic meter harvested Slika 8. Emisije ispušnih plinova harvestera u stvarnim radnim uvjetima po izrađenom kubnom metru 1.14 dm3 related to 1 cubic meter of timber harvested (this is approximately 3030 g of CO2 related to 1 cubic meter of timber harvested). In turn, Berg and Karjalainen (2003) stated that fuel consumption of the harvester in their study fell within 12–16.5 dm3/h. In contrast, Karjalainen and Asikainen (1996) reported the emission of CO2 from the harvester at the level of 1857 g related to 1 cubic meter of timber harvested, which is a value similar to the one obtained in the present paper. However, it is important to note that according to most publications, fuel consumption from harvesters is not measured with the use of the carbon balance method, which might be one of the reasons for different results obtained from different studies. Furthermore, fuel consumption of any harvester depends on a number of factors such as machine operator, terrain leveling, type of trees harvested, weather conditions, etc. and this makes direct comparisons of these studies even more difficult.

As far as the chainsaw is concerned, during the tests its fuel consumption was 0.38 dm3/h, and when related to 1 cubic meter of timber harvested it was 0.26 dm3. These values were lower than the ones found in other publications (e.g. Karjalainen and Asikainen 1996). As mentioned above, fuel consumption of chainsaws, as much as that of harvesters, depends on a number of factors.

In many European countries, forest operations are often carried out with chainsaws driven by gasoline Croat. j. for. eng. 34(2013)1

P. Lijewski et al.

Fig. 9 Comparison of exhaust emissions from a chainsaw and a harvester measured under actual operating conditions Slika 9. Usporedba emisija ispušnih plinova motorne pile lančanice i harvestera izmjerenih u radnim uvjetima

Fig. 10 Comparison of exhaust emissions from a chainsaw and a harvester measured under actual operating conditions related to 1 cubic meter of timber harvested Slika 10. Usporedba emisija ispušnih plinova motorne pile lančanice i harvestera izmjerenih u radnim uvjetima po izrađenom kubnom metru obloga drva engines. Harvesters are not widely used in Europe – Eastern Europe in particular (e.g. Poland). In Poland in 2008 there were 157 (Anon. 2011) harvesters i.e. one

119

P. Lijewski et al.

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

in 58 000 ha of forested land. For comparative purposes, the authors also tested exhaust emissions from a chainsaw under actual operating conditions. The comparison of exhaust emission from the chainsaw and harvester is presented in Fig. 9. The exhaust emissions from a chainsaw are much higher than the emissions from a harvester e.g. for HC the emission is several hundred times higher and for CO over 100 times higher. The smallest differences were recorded for the emission of NOx. To fully understand the differences in the ecological properties between the machines, their efficiency needs to be taken into account. It is estimated that the efficiency of the harvester equals that of 4 chainsaws. Besides, the occupational safety considerations are clearly against chainsaws.

Fig. 10 presents the exhaust emissions related to 1 cubic meter of timber harvested. The emission of CO and HC was much higher for the chainsaw while the emission of NOx was higher for the harvester.

4. Conclusions – Zaključci The presented results of exhaust emission tests performed under actual operating conditions of harvesters during tree harvesting diverge from the Stage IIIA and Tier 3 standards. One should note the engine operating conditions, time density characteristics (time density of the engine operation) in the aspect of the applicable NRSC 8 phase test. From the performed investigations it results that the engine of a harvester operated at a relatively constant speed and at partial loads. It seems that the present method can be successfully used to estimate the ecological burden while performing forest operations. In Poland it is a significant issue as forests cover 29% of the country area i.e.

9.1 million ha. The presented test results support the replacement of gasoline chainsaws with harvesters.

Such a solution is much better from the ecological point of view and in terms of occupational safety.

In further works related to exhaust emissions, the authors plan to carry out research related to other forest machinery (e.g. forwarders), which would allow a total estimation of the environmental burden generated by these machines. Moreover, in the future, the authors plan to conduct tests that will qualitatively evaluate the emissions of particulate matter generated by a harvester under actual operating conditions i.e. determine the PM number and the PM size. As other studies show, it is the PM number and the PM size rather than the PM mass emission that appear to be instrumental in the PM impact on human health (e.g.

Metz 2003). Moreover, it should be noted that harvesters are powered by diesel engines and the IARC report

120

published in June 2012 clearly classifies diesel exhaust as carcinogenic (IARC 2012). This report is based on years of research of the impact of diesel engines on human health (Attfield et al. 2012; Silverman et al.

2012). The above facts fully encourage the continuation of the research with a view to reducing the exhaust emissions from engines including those used in silviculture.

5. References – Literatura Ahola, P. M., 1998: Biodegradable Hydraulic Fluids in the Forest. SAE Paper 981517.

Anon., 2011: Forestry 2011 – Statistical information and elaboration. Polish Central Statistical Office, 2011, Warsaw.

Athanassiadis, D., 2000: Energy consumption and exhaust emissions in mechanized timber harvesting operations in Sweden. The Science of the Total Environment 255(1–3): 135– 143.

Attfield, M. D., Schleiff, P. L., Lubin, J. H., Blair, A., Stewart, P. A., Vermeulen, R., Coble, J. B., Silverman, D. T., 2012: The Diesel Exhaust in Miners Study: A Cohort Mortality Study With Emphasis on Lung Cancer. Journal of the National Cancer Institute 104(11): 869–883.

AVL, 2007: Current and Future Exhaust Emissions Legislation. AVL List GmbH, Graz 2007.

Berg, S., Karjalainen, T., 2003: Comparison of greenhouse gas emission from forest operations in Finland and Sweden. Forestry 76(3): 271–284.

Czerwinski, J., Wyser-Heusi, M., Mayer, A., 2001: Emissions of Small 2S-SI-Engine for Handheld Machinery-Nanoparticulates & Particulate Matter. SAE Paper 2001-01-1830.

Hartweg, A., Keilen, K., 1998: Der Umwelttvertraglichkeit von Bioolen. Allgemeine Forst Zeitschrift 7: 148–150.

IARC, 2012: Diesel Engine Exhaust Carcinogenic. Press Release N° 213, 12 June 2012.

Jacke, H., Hoss, C., Augusta, J., 1996: Gefahrstoffe beim Einsatz der Motorsage. Forst & Technik 10: 14–19.

Karjalainen, T., Asikainen, A., 1996: Greenhouse gas emissions from the use of primary energy in forest operations and longdistance transportation of timber in Finland. Forestry 69(3): 216–228.

Klvac, R., Fischer, R., Skoupy, A., 2012: Energy Use of and Emissions from the Operation Phase of a Medium Distance Cableway System. Croatian Journal of Forest Engineering 33(1): 79–88.

Laanti, S., Sorvari, J., Elonen, E., Pitkanen, M., 2001: Mutagenicity and PATHs of Particulate Emissions of Two-Stroke Chainsaw Engines, SAE Paper 2001-01-1825/4246.

Merkisz, J., Lijewski, P., Fuc, P., 2010: Exhaust Emission Tests from Agricultural Machinery under Real Operating Conditions. SAE Paper 2010-01-1949.

Croat. j. for. eng. 34(2013)1

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122) Merkisz, J., Lijewski, P., 2010: Time Density of Engine Operation in Non-road Vehicles in the Aspect of the Homologation Toxic Emission Test. SAE Paper 2010-01-1282.

Metz, N., 2003: Diesel particulate matter criteria for evaluation of health effects. In: 24 Internationales Wiener Motorensymposium, Wien 2003.

Noren, O., Pettersson, O., 2001: Development of Relevant Work-Cycles and Emission Factors for Off-Road Machines.

SAE Paper 2001-01-3637.

Nordfjell, T., Athanassiadis, D., Talbot, B., 2003: Fuel Consumption In Forwarders. International Journal of Forest Engineering 2(14): 11–20.

Ohtsuji T., Kobayashi, B., 2002: Investigation of low emission two-stroke cycle engine chainsaw. SAE Paper 2002-32-1841/ JSAE 20022434.

Rodenbeck, J., Auler, B., Lugger, J., Gorenflo, E., 2006: Development of a Valve Controlled Four-stroke Chainsaw to Meet Future Emission Regulations. SAE Paper 2006-320090/20066590.

Schlossarczyk, J., Maier, G., Roitsch, T., 2004: Conceptual Design Study for a Catalytic Muffler Chainsaw Application to Fulfill Emission Standards and Thermal Demands. SAE Paper SAE 2004-32-0060 / JSAE 20044347.

Shahinian, V. D., 2007: SENSOR tech-ct Update Application Software for SEMTECH Mobile Emission Analyzers. Sensors 4th Annual SUN (SEMTECH User Network) Conference, 22.

10. 2007.

Silverman, D. T., Samanic, C. M., Lubin, J. H., Blair, A. E., Stewart, P. A., Vermeulen, R., Coble, J. B., Rothman, N., Schleiff, P. L., Travis, W. D., Ziegler, R. G., Wacholder, S., Attfield, M. D., 2012: The Diesel Exhaust in Miners Study: A Nested CaseControl Study of Lung Cancer and Diesel Exhaust. Journal of the National Cancer Institute, (2012) 104(11): 855–868.

Skoupy, A., Klvac, R., Hosseini, S., 2010: Changes in the External Speed Characteristics of Chainsaw Engines with the Use of Mineral and Vegetable Oils. Croatian Journal of Forest Engineering 31(2): 149–155.

Sonnleitner, G., 1992: Bio-Motorsagenkettenole in der Forstwirtschaft. Osterreichische Forstzeitung 4, p. 39.

Walsh, M., 2011: Global trends in motor vehicle control: a 2011 update. Combustion engines 50(2): 106–117.

Croat. j. for. eng. 34(2013)1

P. Lijewski et al.

Wightman, P., Eavis, R., Walker, K., Batchelor, S., Carruthers, S., 1998: A comparative LCA of chainsaw bar lubricants made from mineral oil and rapeseed oil. Second International Conference, Ceuterick D. editor, LCA in Agriculture, Agro-Industry and Forestry, 3-4 December, Brussels, p. 35–38.

Zahn, W., 2000: Low Emission Technologies for High-Performance Handheld Outdoor Products. SAE Paper 2000-010896.

Abbreviations – Kratice CO – Carbon Monoxide – Ugljikov monoksid CO2 – Carbon Dioxide – Ugljikov dioksid EPE – Estimate of Position Error – Procjena greške položaja GHG – Greenhouse Gases – Staklenički plinovi FID – Flame Ionization Detector – Detektor ionizacije plamena GPS – Global Position System – Globalni pozicijski sustav HC – Hydrocarbons – Ugljikovodici IARC – International Agency for Research on Cancer Međunarodna agencija za istraživanje raka LAM – Laser Aerosol Monitor – Laserski monitor aerosola NDUV – Non-Dispersive Ultra-Violet – Neraspršujuća ultraljubičasta NOx – Nitric Oxides – Dušikovi oksidi NRSC – Non-road Stationary Cycle – Necestovni stacionarni ciklus NRTC – Non-road Transient Cycle – Necestovni ciklus kretanja OBD – On Board Diagnostics – Dijagnostika na vozilima PAH – Polycyclic Aromatic Hydrocarbons – Policiklički aromatski ugljikovodici PEMS – Portable Emissions Measurement System Prijenosni sustav za mjerenje emisija PM – Particulate Matter – Čestice čađe RME – Rapeseed Methyl Esters – Metilni esteri repičinoga ulja

121

P. Lijewski et al.

Research of Exhaust Emissions from a Harvester Diesel Engine ... (113–122)

Sažetak

Istraživanje emisija ispušnih plinova dizelskoga motora harvestera uporabom prijenosnoga sustava za mjerenje emisija U radu se istražuju emisije ispušnih plinova harvestera i motorne pile pri pridobivanju drva. Predstavljena je poredbena analiza glavnih spojeva u ispušnim plinovima (ugljikov monoksid, dušikovi oksidi, ugljikovodici i čestice čađi). Autori su proveli i analizu potrošnje goriva spomenutih strojeva pri sječi i izradbi drva. U šumskim se operacijama sve češće koriste harvesteri, no još se uvijek moraju natjecati s motornim pilama pogonjenima benzinom.

Uprvo su zbog toga autori uključili u rad i usporedbu emisija ispušnih plinova motorne pile s emisijama motora harvestera. Analizom dostupne literature u vezi s potrošnjom goriva i utjecaja na okoliš šumskih strojeva dolazi se do zaključka da je većina radova temeljena na laboratorijskim ispitivanjima i izračunima. U ovom je radu primijenjen drugačiji istraživački pristup. Istraživanje emisija ispušnih plinova i potrošnje goriva provedeno je u stvarnim radnim uvjetima tijekom pridobivanja drva (borovine). U tu su svrhu korišteni analizatori emisija ispušnih plinova PEMS (prijenosni sustav za mjerenje emisija). Za mjerenje plinova korišten je SEMTECH DS, a za mjerenje količine čestica čađe korišten je LAM (oba proizvođača SENSORS). Navedeni analizator može mjeriti koncentraciju komponenti ispušnih plinova u on-line modu tijekom rada motora u terenskim uvjetima. Predstavljena je metodologija novo rješenje koje bi u budućnosti moglo poslužiti kao osnova za rad na razvoju smanjenja emisija ispušnih plinova motora. Tijekom istraživanja autori su mjerili i količinu proizvedenoga drva, stoga su rezultati emisija mogli biti dovedeni u vezu s količinom proizvoda (jednoga kubnoga metra izrađene oblovine).

Ključne riječi: emisije ispušnih plinova, harvester, motorna pila, mjerenje na vozilima

Authors’ address – Adresa autorâ:

Received (Primljeno): December 9, 2011 Accepted (Prihvaćeno): September 4, 2012

122

Piotr Lijewski, MSc.* e-mail: piotr.lijewski@put.poznan.pl Prof. Jerzy Merkisz, PhD. Paweł Fuć, MSc. Poznan University of Technology Institute of Internal Combustion Engine and Transport Ul. Piotrowo 3 90-965 Poznan POLAND * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

Original scietific paper – Izvorni znanstveni rad

Effects of Rubber-Tired Cable Skidder on Soil Compaction in Hyrcanian Forest Baris Majnounian, Megdad Jourgholami Abstract – Nacrtak The use of skidders equipped with rubber tires is a well accepted practice for the extraction of timber from the forest, but the application also causes considerable environmental problems.

The aim of the study was to evaluate the effects of different slope gradient, number of machine passes on skid trails and soil depth on soil compaction. The study was designed as an experiment with the factors including slope gradient, soil moisture, and soil depth on various skid trails and with different number of machine passes. The effects of four slope classes (flat, 10%, -10% and -20%), three soil depth classes (5, 15 and 25 cm), and different compaction levels based on various number of machine passes (0, 1, 5, 8, 10, 15, 20, 25 and 30) were evaluated.

A Timberjack cable skidder was used and the study location was in the Kheyrud Educational and Research Forest located in the Hyrcanian forest in northern Iran. The increased number of machine passes increased soil bulk density, but the highest rate of compaction occurred after the initial few passes. Uphill skidding increases soil compaction more than downhill skidding.

The increases in bulk density were still significant at the maximum sampling depth of 20–30 cm. Soil bulk densities at 5, 15 and 25 cm depth averaged 35, 22 and 17% higher than densities of undisturbed soil.

Keywords: soil compaction, soil bulk density, rubber-tired cable skidder, Hyrcanian forest.

1. Introduction – Uvod Forest soils, in general, are susceptible to compaction as they are loose with high organic-matter, and are generally low in bulk density, high in porosity, and low in strength (Froehlich et al. 1985; Kolkaa and Smidt 2004). The impact of skidding operations on forest soils can be divided into three major categories: soil profile disturbance, soil compaction and soil puddling and rutting (Rab et al. 2005).

When a mechanical load is applied to the soil, soil particles are rearranged closer together resulting in increased bulk density (mass per unit volume) (Cullen 1991; Eliasson 2005; Grace et al. 2006), reduction of the total porosity associated with a reduction of macropores (Gayoso and Iroume 1991; Gomez et al. 2002; Ares et al. 2005), increase in soil strength; except for soil with low bearing capacities (Horn et al. 1994), decreased infiltration capacity (Horn et al. 1994, 2004), decreased gaseous exchange and soil aeration (Horn et al. 1994), an increase in resistance to penetration (Ampoorter et al. 2007), decrease in saturated hydraulic conductivity Croat. j. for. eng. 34(2013)1

(Greacen and Sands 1980; Horn et al. 1994; Grace et al.

2006), and increased micropore proportion (Kolkaa and Smidt 2004). One of the major impacts of harvesting operations is soil profile disturbance. Soil disturbance is usually defined in terms of mixing and/or removal of litter and soil, which may change the physical, chemical or biological properties of soil (Rab et al. 2005). Depending on the equipment used, the surface soils are variously mixed, buried or inverted.

During timber harvesting the degree of soil compaction depends on various factors including: site and soil characteristics (Adams and Froehlich 1984; Ampoorter et al. 2007) such as soil texture (Froese 2004; Rohand et al. 2004), soil moisture (Johnson et al. 2007), the number of machine passes (Eliasson 2005; Šušnjar et al. 2006; Ampoorter et al. 2007; Eliasson and Wasterlund 2007; Wang et al. 2007) and harvesting system. In addition, the machine characteristics affecting the degree of soil compaction include type of machine (Šušnjar et al. 2006; Wang et al. 2007), mass of vehicle and load (Rab 1996; Saarilahti 2002; Šušnjar et al. 2006; Horn et al. 2007), type, number of wheels and inflation pressure of the

123

B. Majnounian and M. Jourgholami

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

tires (Ziesak 2006), amount of logging slash (Wronski and Murphy 1994; Eliasson and Wasterlund 2007). A couple of studies reported by several researchers show that one of the critical factors affecting the degree of soil compaction is the number of machine passes over a specific point. These studies show that most compaction occurs during the first ten passes of a vehicle with the most occurring in the first three passes. Subsequent passes generally have little additional effect (Ampoorter et al. 2007). Most compaction occurred after the initial few passes (Matangaran and Kobayashi 1999), but bulk density also increased significantly after more than 3 passes (Gayoso and Iroume 1991; Eliasson 2005; Eliasson and Wasterlund 2007). Matangaran and Kobayashi (1999) found that the bulk density increased markedly by the first and second pass of the tractor, but did not change after the fifth pass.

Few studies have documented slope gradient of trail effects (longitudinal and transversal slope of trail) on the extent and degree of soil compaction and disturbance. Krag et al. (1986) showed that during timber harvesting, slope steepness had a stronger effect than season of logging on soil disturbance. Under steep terrain conditions, the machine slipped continuously and remained in a given place for a longer period of time, puddling and dragging the soil (Gayoso and Iroume 1991). Sidle and Drlica (1981) found that the slope did not significantly affect bulk density, but they concluded that it can be an important factor in the potential level of compaction. Jamshidi et al. (2008) found that there was no detectable difference in compaction between machine skidding on flat trails and trails with longitudinal gradient or transversal slope. Compacted layers are often found at different soil depths. However, the deeper layers of many soils are compacted further after a few passes. The values of the soil bulk density mostly depend on the quantity of organic matter. In the surface horizons, the soil bulk density is low, and as with the increase of depth, organic matter is rapidly decreased, the bulk density increases in subsoil (Froehlich and McNabb 1984). Increases in organic matter reduce soil compactibility (Kozlowski 1999).

A strong increase in bulk density is most distinct in the upper 20 cm of the soil, since the exerted pressure is maximal at the soil surface and declines with increasing depth as the total pressure is spread out over an enlarging area (Gent and Morris 1986; Ampoorter et al. 2007). Johnson et al. (2007) noted that soil compaction was generally limited to skid trails and top soil layers (<30 cm). Eliasson and Wasterlund (2007) reported that an increased number of machine passages increased soil dry density in the upper 20 cm.

The Hyrcanian mountainous forest in northern Iran is rich in biological diversity, with endemic and

124

endangered species, and a diverse range of economic and social conditions. In the Hyrcanian forest, a few studies have been carried out about the effects of forest operations on soil compaction and bulk density. Jamshidi et al. (2008) measured the changes in bulk density in the top 10 cm of soil following machine and animal skidding in the Hyrcanian forest. They found that the average soil bulk density in the tracks of machine skid trails was significantly greater than the soil density outside the tracks, but the increase in bulk density was not significant on the animal trails.

The extent of the severe disturbance from groundbased harvesting systems varies depending on slope and steep terrain, although the effects of slope on soil disturbance and bulk density have received less attention. The specific objectives were to: quantify the extent of trail area and winching line (disturbance area) throughout the harvest unit, to characterize and establish the threshold levels for the machine traffic with respect to bulk density and slope gradient or direction of skidding for three different soil depths.

2. Materials and methods – Materijal i metode 2.1 Study site – Područje istraživanja About 65% of the Hyrcanian forests are located in mountainous areas with terrain slope of more than 27% (Fig. 1), where forest lands are not readily acces-

Fig. 1 Proportion of the Hyrcanian forest area based on terrain slope Slika 1. Udio nagiba terena šumskih zemljišta Hirkanijske šume (Golestan) Croat. j. for. eng. 34(2013)1

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

B. Majnounian and M. Jourgholami

Fig. 2 Location of study site within the Hyrcanian Forest Slika 2. Područje istraživanja sible with ground-based logging equipments. The cable yarding technologies are still undeveloped in this forest area.

The research was carried out in compartment No. 220, which is located in Namkhaneh District within Kheyrud Educational and Research Forest in the Hyrcanian forest of northern Iran (Fig. 2). The altitude ranges from 1000 to 1135 m and the forest lies on a southwestern aspect. The average rainfall ranges from 1420 to 1530 mm/year, with the heaviest precipitation occurring in the summer and fall. The average daily temperature ranges from a few degrees below 0°C in December, January, and February, and up to +25°C during the summer. This area is dominated by natural forests containing native mixed deciduous tree species such as Fagus orientalis Lipsky, Carpinus betulus L., Acer velutinum Boiss. and Alnus subcordata. The management method is mixed un-even aged high forest with single and group selective cutting regime. The soil of the study site is classified as a brown forest soil (Alfi-

sols) and well-drained. The texture of the soil ranges from silt loam to loamy.

Trees to be removed were felled, limbed and topped motor-manually. Felled trees were bucked and processed with chainsaws into logs, sawn-lumber and pulpwood. The logs of 5–15 meter length were extracted by rubber-tired skidders to the roadside landings.

The fuel wood was extracted by mules. Also, in steep terrain that could not be reached by skidders, logs were processed to sawn-lumber and then hauled by mules.

An important strategy is to limit traffic on designated skid trails, hence, landings and skid trails were clearly flagged on the ground before harvesting. The intension was to require the skidder to stay on the skid trail and winch logs on the trail. Downhill and uphill skidding to the landing was planned without any excavation and the skidding operations were done on natural ground.

The extraction distances to the roadside landing was 780 m. The skid trail slope ranges from 0 to 35%. Table 1 presents some characteristics of the study site.

Table 1 Characteristics of the study site Tablica 1. Značajke istraživanoga područja Area, ha

Tree per ha

Volume, m3/ha

Total removed trees

Površina, ha

Broj stabala po ha

Obujam, m3/ha

Ukupno posječeno stabala

17

173

504

Croat. j. for. eng. 34(2013)1

270 (10 tree/ha – 10 stabala/ha)

Total volume of removed trees, m3

DBH of removed trees, cm

Ukupni drvni obujam, m3

Raspon prsnih promjera posječenih stabala, cm

872.3 (32 m3/ha)

20–135

125

B. Majnounian and M. Jourgholami

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

2.2 Experimental design and data collection Plan istraživanja Felling of marked trees was carried out in March and skidding operations were done in August 2008.

At the time of harvesting, weather conditions were very dry and warm for more than 5 weeks and these conditions remained constant during skidding operations. Fig. 3 presents the 4WD Timberjack 450C rubber-tired skidder used in the study. This machine is normally an articulated, four-wheel-drive vehicle weighing 10.3 ton (55% on the front and 45% on the rear axle) with engine power of 177 hp (132 kW) and engine model of 6BTA5.9. It is equipped with a blade for light pushing of obstacles and stacking of logs. The skidder was fitted with 24.5–32 tires inflated to 220 kPa on both front and rear axles, and it had a ground clearance of approximately 0.6 m with overall width of 3.1 m. Timber bunching was carried out by the winch installed in the rear part of the skidder from the stump to the skidder and one end of the round wood was dragged on the ground. In the study areas, the average logged volume in each pass was 3.5 cubic meters (1 and 3 logs, respectively).

Twelve sampling transects were selected at different slope gradients along the designated skid trail for bulk density measurements (Fig. 4). Organic horizons were removed from the soil surface prior to density measurements, so that depth readings were referenced to the mineral soil surface. In order to ensure that the measurements were made in the same place after a certain number of passes, we have put the painted sticks in the center of skid trails. The painted sticks indicated the centers of the experimental skid trails at the skid trial, so that the machines would follow the

Fig. 3 Timberjack skidder, equipped with rubber-tires while extracting timber Slika 3. Skider Timberjack prilikom privlačenja drva same tracks at subsequent passes. Before skidding, four slope gradients were established in the skid trail with 3 replications in disturbed areas at 0–10 cm soil profile depth, and the different levels of compaction were applied by varying the levels of machine traffic: 0 (undisturbed), 1, 5, 8, 10, 15, 20, 25 and 30 machine passes. A pass implies a drive back and forth the selected trail. Four slope gradients of skid trail were 0 (flat trail), 10%, –10% and –20%. Also, prior to any skidding operations and after 20 machine passes, bulk density was measured at this four slope gradient trail (flat trail, 10%, –10% and –20%) at the 5 cm, 15 cm and 25 cm soil profile depths in wheel rut (A and B) and control sample point (Fig. 4) on the adjacent skid trail

Fig. 4 Sketch of study layout for soil sampling Slika 4. Shema rada za uzimanje uzoraka tla

126

Croat. j. for. eng. 34(2013)1

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

(C). The soil sample cores were obtained from the layers of the mineral soil using a thin walled steel cylinder, 40 mm long and 56 mm in diameter, inserted into the soil by a hammer-driven device. After extracting the steel cylinder from the soil with minimal disturbance to the contents, the soil cores were trimmed flush with the cylinder end and extruded into a plastic bag for transporting it to the laboratory. Samples were weighed on the day they were collected and again after oven drying at 105 °C for 24 h to determine water content and bulk density.

In order to determine the extent of disturbance from skidder operations, disturbed widths were measured at 25 m intervals along the skid trails. Skid trail soil disturbance was classed as: A-horizon puddled and mixed with forest floor organic debris, and some Ahorizon removed and the rest mixed with B-horizon.

Also landing areas were measured in the compartment.

Log winching, however, cause the excavation of the line between stumps near the vehicle, hence, for determining this displacement, total length and width of winching lines in both sides of trails were measured.

2.3 Statistical analysis – Statistička obrada podataka The experimental design was a factorial arrangement of treatments conducted in a completely randomized design. General linear modeling (GLM) was applied to relate bulk density and rut depth to machine passes, slope gradient, and depth in relation to the skid trails. Post-hoc comparison of means was per-

B. Majnounian and M. Jourgholami

formed using Duncan’s multiple designs to meanbased grouping with a 95% confidence level. Analysis of variance of the data was conducted in SPSS (release 15.0) to identify differences between bulk density values of four slope gradients in skid trails. Treatment effects were considered significant if P<0.05. Soil bulk density before and after skidding operations was compared using independent samples t-test. Also, oneway ANOVA was performed.

3. Results – Rezultati 3.1 Soil disturbance – Oštećenje tla A detailed survey of the harvested unit following extraction with a cable skidder indicated that 5.8% of the total area (17 ha) was covered with skid trails and an additional 0.8% of the unit was occupied by the landing (Table 2).

In this study, ground-based winching of timber from the felling site to the skidder had substantial effect on soil displacement that occupied 0.9% of the total area. With the whole load lying on the ground, during winching it removed and pushed a layer of soil in front of itself. Finally, in this study 7.5% of the harvesting total area was disturbed and compacted. The disturbance width of the trail was significantly influenced by transversal slope gradients of trails. Therefore, the higher transversal slope, the wider is the trail width. There were three main skid trails in the harvest area with a total length of 1971 meters. The average

Table 2 Compartment disturbance area due to rubber-tired skidder operation Tablica 2. Oštećenje tla nastalo pri radu skidera Factor Mjesto Trail 1 Vlaka 1 Trail 2 Vlaka 2 Trail 3 Vlaka 3 Total area Uk. površina Area, % Površina, %

Trail length, m Duljina vlake, m

Trail average width, m

Disturbed area, m2

Prosječna širina Oštećena vlake, m površina tla, m2

Landing area, m2 Pomoćno stovarište, m2

Winching line, m

Line width, m

Winching disturbed area, m2

Total disturbed area, m2

Duljina Širina traga Površina tla oštećena Ukupno oštećena privitlavanja, m privitlavanja, m skupljanjem drva, m2 površina tla, m2

324

4.8

1555.2

1150

241

0.32

77.12

–

1387

5.1

7073.7

–

3752

0.34

1275.68

–

260

4.95

1287

180

380

0.29

110.2

–

1971

–

9916

1330

4373

–

1463

12709

–

–

5.8

0. 8

–

–

0.9

7.5

Croat. j. for. eng. 34(2013)1

127

B. Majnounian and M. Jourgholami

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

was 5 meters, and so occupied 5.8% of the total area.

4373 meters of winch line disturbance was recorded with an average width of 0.315 meters for a total of 0.9% of the area.

3.2 Soil compaction: influence of slopes and machine passes – Zbijanje tla: utjecaj nagiba terena i broja prolazaka vozila Table 3 shows the analysis of the soil bulk density data influenced by machine passes and slope gradient Table 3 Analysis of variance (ANOVA) for the effect of number of machine passes (NP) and slope gradient (SG) on bulk density in 0–10 cm soil depth Tablica 3. Analiza varijance za prolazak vozila (NP) i nagib terena (SG) na dubini tla do 10 cm Source Izvor podataka

Sum of Square Zbroj kvadrata

Mean Square

df

F-value P-value Srednja F vrijednost P vrijednost kvadratna vrijednost

NP

1.77

8

0.221

829.34

0.00

SG

0.17

3

0.058

216.36

0.00

NP × GS

0.06

24

0.003

9.67

0.00

Fig. 6 Relationship between the increase of bulk density (%) and machine passes Slika 6. Odnos povećanja u gustoći tla (%) ovisno o prolasku vozila for the cable skidder. The results showed that machine passes and slope gradient, and the interaction effects of machine passes × slope gradient were all significant variables (P<0.05).

The independent samples t-test indicated that skidding had a statistically significant effect on the bulk density of soil on trails before and after machine passes in each trail with different slope gradients and by independent samples t-test and Duncan’s test (b). The values are mean. Different letters within each slope treatment show significant differences (P < 0.05) (Fig. 5).

The results show that bulk density significantly increased as the number of machine passes increased (Fig. 6). Regardless of the slope gradient, the degree and level of compaction differed among trail slope using Duncan’s multiple range test (Fig. 7 and Table 4).

In Table 4, for each soil interval means are compared against each other after ANOVA using Duncan’s test.

Values are mean. The difference between values in a column followed by different superscripts is significant at P < 0.05.

Fig. 5 Average bulk density (a) and its relative changes before and after skidding in each slope gradient of the trail and by independent samples t-test and Duncan’s test (b) Slika 5. Prosječne vrijednosti (a) i relativne promjene gustoće tla nakon privlačenja drva (b)

128

In the other hand, generally, trails with four slopes show a similar trend of increasing soil bulk density with increasing amounts of machine passes. In flat trail, the bulk density in the top 0–10 cm of soil (1.06 g/cm3) increased by 5% after 1 pass, 19% after 5 passes, 25% after 8 passes, 31% after 15 machine passes. In trail with a 10% slope or uphill skidding, the soil bulk density inCroat. j. for. eng. 34(2013)1

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

B. Majnounian and M. Jourgholami

Fig. 7 Relationship between the increase of bulk density and machine passes on different slope gradients Slika 7. Odnos povećanja u gustoći tla s prolaskom vozila na vlakama različitih nagiba creased by 19% after 1 pass, 43% after 5 passes. Subsequent increase of the number of passes (up to 30 turns) did not increase the bulk density significantly. High level of increase in bulk density occurred after 5 machine passes and additional increase of passes did not increase the bulk density significantly. In the area with -10%, bulk density increased by 9% after 1 pass, 25% Croat. j. for. eng. 34(2013)1

after 5 passes, 34% after 8 turns and in trail with -20% slope, by 9% after 1 pass, 22% after 5 passes, 29% after 8 turns and 34% after 10 passes. In flat trail, the highest rate of compaction, as bulk density increased, took place during the first 15 passes by 1.37 g/cm3. In trail with 10% slope gradient, in contrast, high increase was observed in bulk density (1.44 g/cm3) and it occurred

129

B. Majnounian and M. Jourgholami

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

Table 4 Mean bulk density values (± standard deviation) as influenced by machine passes and slope gradient Tablica 4. Srednje vrijednosti gustoće tla (± vrijednost standardne devijacije) pod utjecajem prolazaka vozila i nagiba terena (SG) Number of machine passes – Broj prolazaka vozila

SG (%)

0

1

5

8

10

15

20

25

30

-10

1.08±0.01a*

1.17±0.01b

1.33±0.01b

1.42±0.02b

1.42±0.01b

1.43±0.01b

1.43±0.02b

1.44±0.02b

1.44±0.02b

-20

1.05±0.03ab

1.14±0.01c

1.27±0.02c

1.34±0.01c

1.39±0.03b

1.39±0.02c

1.4±0.01c

1.4±0.02c

1.41±0.01c

0

1.06±0.03ab

1.11±0.01d

1.25±0.02c

1.31±0.02c

1.34±0.02c

1.37±0.02c

1.38±0.01c

1.39±0.01c

1.39±0.01c

10

1.03±0.01b

1.22±0.01a

1.46±0.02a

1.46±0.02a

1.47±0.01a

1.47±0.03a

1.47±0.01a

1.48±0.01a

1.48±0.01a

after 5 machine passes. Also, in downhill skidding with –10% and –20%, bulk densities were increased significantly after 8 and 10 machine passes, respectively. Then, soil bulk density for –10% and –20% was 1.41 and 1.41 g/cm3, respectively. Bulk density in the 10% trail showed the highest value in comparison with other slope gradients of the trail (Fig. 5). Skidding operations along flat trail had the lowest compaction (Duncan’s).

3.3 Soil compaction: influence of slopes and soil depths – Zbijanje tla: utjecaj nagiba terena i dubine tla Table 5 shows the analyses of the soil bulk density data as influenced by position, slope gradient and depth after 20 machine passes. The results showed that position, depth, slope gradient, and the interaction effects of position × slope gradient and position × depth were all significant variables (P<0.05). General Linear

Model (GLM) indicated two significant interaction terms, and namely position × slope gradient (p<0.01) and position × depth (p<0.01). It can be noticed that bulk density values of control sample points (no pass) in four slope gradients are clearly less than compacted values. The interaction between position and depth was also significant.

Average pre-harvest bulk densities for three soil depth classes, 0–10 cm, 10–20 cm, and 20–30 cm were 1.06 g/cm3, 1.27 g/cm3, and 1.42 g/cm3, respectively.

After 20 machine passes, bulk density increased in depth under the skid trails in all slope gradients of trails, but the major increase occurred in the top of the soil profile at 0–10 cm. In flat trail, bulk density increased by 30% in 0–10 cm depth, by 20% in 10–20 cm, and by 17.5% in 20–30 cm, after 20 machines passes.

In trails with a 10% slope, the increase in bulk density for all depths was significantly higher as compared

Table 5 Analysis of variance (ANOVA) for the effect of sample position, slope gradient, and depth on bulk density in skid trials Tablica 5. Analiza varijance (ANOVA) podataka nagiba terena, dubine tla, gustoće tla i mjesta uzimanja uzoraka na traktorskim vlakama Source

Sum of Square

Izvor podataka

Zbroj kvadrata

df

Mean Square

F-value

P-value

Srednja kvadratna vrijednost

F vrijednost

P vrijednost

Slope – Nagib terena

0.012

3

0.004

3.78

0.02

Position – Položaj

1.611

1

1.611

1528

0.00

Depth – Dubina tla

1.035

2

0.518

491

0.00

0.021

3

0.007

6.59

0.00

0.003

6

0.001

0.48

0.82

0.046

2

0.023

21.7

0.00

0.001

6

0.00

0.21

0.97

Slope * Position Nagib terena * Položaj Slope * Depth Nagib terena * Dubina tla Position * Depth Položaj * Dubina tla Slope * Position * Depth Nagib terena * Položaj * Dubina tla

130

Croat. j. for. eng. 34(2013)1

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

with those observed in trails with –10%, –20% slope, and flat trail. Fig. 8 shows how the relative change in bulk density varied with slope trail in soil depth. In trail with 10% slope, bulk density increased by 42% in 0–10 cm depth, by 28% in 10–20 cm, and by 20% in 20–30 cm, after 20 machine passes. The independent samples t-test indicated that skidding had a statistically significant effect on the soil bulk density on trails before and after vehicle passes in soil depth (p<0.05).

Deeper in the soil profile, differences between control and the treatments in four slope gradient became smaller. The highest level of increase in bulk density was found in the trail with 10% slope gradient between control and the treatments.

For the soil bulk density samples in four slope gradient, one-way analysis of variance (ANOVA) and Duncan’s test were used to see if there were significant differences (P<0.05) between the soil depths. In flat trail before and after skidding, bulk density increased significantly in all depths (Fig. 9), but there were no significant differences between soil depths at 10–20 cm and 20–30 cm in trails with uphill (10%) and downhill (–10% and –20%) slope gradient. With respect to the bulk density values for different trails, smaller values were generally observed compared with uphill and downhill skidding for flat trail. However, a peak can also be seen in the depth interval of 20–30 cm. Also, 10% trail resulted in the highest bulk density values for all depths, while flat trail showed the smallest degree of compaction.

Fig. 8 Average decrease in bulk density with soil depth on skid trails of four slope gradients Slika 8. Prosječno smanjenje gustoće na različitoj dubini tla Croat. j. for. eng. 34(2013)1

B. Majnounian and M. Jourgholami

4. Discussion – Rasprava 4.1 Soil disturbance – Oštećenje tla Once the sampling method was established, the soil disturbance survey was both quick and easy to complete. In this forest where skidding is common, the skid trail pattern is distributed unevenly because of terrain steepness. Using a different sampling method in this situation did not result in an exactly accurate survey and had many overestimations. Compared to other studies (Froehlich and McNabb 1984; Rab et al.

2005; Šušnjar et al. 2006), soil disturbance in this study occupied less than 8 percents. This agreed with Eliasson (2005) who found that soil disturbance will be affected by several factors such as wheel slip, vibration and number of vehicle passes.

4.2 Soil compaction: influence of slopes and machine passes – Zbijanje tla: utjecaj nagiba terena i prolaska vozila The results show that the average bulk density significantly increased after the operation of rubber-tired skidders. However, in different slope gradient percentages, the increased bulk densities were statistically different. The results of most studies were consistent with our results (Sidle and Drlica 1981; Froehlich and McNabb 1984; Gayoso and Iroume 1991; Eliasson 2005; Šušnjar et al. 2006; Ampoorter et al.

2007; Eliasson and Wasterlund 2007; Horn et al. 2007; Jamshidi et al. 2008). Also, Horn et al. (2007) showed that each stress applied at the soil surface is always transmitted three-dimensionally and causes not only soil compaction but also shear effects.

The results show that bulk density significantly increased with the increase of vehicle passes. In general, trails with four slopes show a similar trend of increasing soil bulk density with the increasing number of vehicle passes. For most treatments, the highest rates of increase in bulk density were achieved in the first 5 to 15 vehicle passes. Beyond 15 vehicle passes, there was usually very little increase in bulk density (Matangaran and Kobayashi 1999). In the flat trail, the skidder operator used the whole width of the road instead of traveling in the same wheel tracks. Impacts of the frequency of vehicle passes on soil compaction showed similar results in many researches (Sidle and Drlica 1981; Gayoso and Iroume 1991; Ampoorter et al. 2007; Jamshidi et al. 2008). This agreed with Wang et al. (2007) and Horn et al. (2007) who both claimed that subsequent vehicle passes increased the soil compaction at a lesser extent until there is little or no more compaction associated with further vehicle passes.

131

B. Majnounian and M. Jourgholami

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123â&#x20AC;&#x201C;135)

Fig. 9 Mean bulk density depending on different sampling places Slika 9. Srednje vrijednosti gustoÄ&#x2021;e tla ovisno o mjestu uzimanja uzorka In this study, flat trails had the lowest bulk density, the trails with -10% and -20% slope gradient (downhill skidding) had intermediate bulk density and the trails with 10% slope gradient (uphill) had the highest compaction. This result can be explained based on the uneven load distribution between the downhill and uphill tires of the skidder (Jamshidi et al. 2008). Another reason for lower bulk density at the downslope track

132

might be the dragging of the logs on or close to this track. Dragged behind the skidder, the logs and especially the log heads might have ripped and loosened up the surface of the highly compacted downslope wheel track (Jamshidi et al. 2008). However, Gayoso and Iroume (1991) stated that this may be a consequence of the problem that the skidder might face when logging in steep terrains. Under these conditions Croat. j. for. eng. 34(2013)1

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

the vehicle slipped continuously and remained in a given place for a longer period of time, puddling and dragging the soil. In uphill skidding, rubber tires slipped on surface soil, then this wheel slippage, the vibration applied and shear strength caused the exposure of mineral subsoil, which has higher density than surface layer. Also, higher soil compaction in the uphill skidding can be explained by the higher load of the skidder rear axle. In the other hand, uphill skidding resulted in sever disturbance and compaction in initial vehicle pass, however bulk density in this condition had the highest amount in comparison to downhill skidding and flat skid trail. Also, Jamshidi et al.

(2008) found that there was no detectable difference in compaction between vehicle skidding on flat trails and trails with longitudinal gradient or transversal slope.

They concluded that the different site conditions and skidding frequencies might have affected the impacts of the different gradients/slopes.

4.3 Soil compaction: influence of slopes and soil depths – Zbijanje tla: utjecaj nagiba terena i dubine tla The results showed that average pre-harvest bulk densities significantly increased as soil depth increased for all slope gradients. The wheel or track slip directly affected the soil structure and altered physical soil properties down to deeper depths. In the other hand, the values of the undisturbed soil bulk density mostly depends on the quantity of organic matter, and increasing soil depth; the organic matter rapidly decreases, and the bulk density increases in subsoil. In the upper soil, biological activity (roots and animals) can act to reduce resistance and soil bulk density while at lower depths soil texture, gravel content and structure may increase soil resistance and soil bulk density (Greacen and Sands 1980; Adams and Froehlich 1984; Froese 2004; Johnson et al. 2007). In this study, with increasing soil depth the compaction level also increased, which is in agreement with the results of other researchers (Greacen and Sands 1980; Sidle and Drlica 1981; Gent and Morris 1986; Gayoso and Iroume 1991; Ares et al. 2005; Eliasson and Wasterlund 2007; Johnson et al. 2007). The results show that deeper in the soil profile, differences between control and the treatments in four slope gradient became smaller. The highest level of increase in bulk density was found in the trail with 10% slope gradient (uphill) between control and the treatments. In flat trail, bulk density increased significantly in all depths. This is related to the rather homogeneous weight distribution of the skidder on flat skid trails. Compaction effect is most distinct in the upper 20 cm of the soil, since the exerted pressure is maximal at the soil surface and declines Croat. j. for. eng. 34(2013)1

B. Majnounian and M. Jourgholami

with increasing depth as the total pressure is spread out over an enlarging area. Thus, an increase generally occurs especially in the upper soil layers (Greacen and Sands 1980). It could also be suggested that the bearing capacity of the soil (maximum load without soil structure failure) grows with increasing bulk density. In this way the surface layer is protected against further compaction when traffic is continued (Ampoorter et al. 2007).

5. Conclusion – Zaključak According to our findings, it may be concluded that the highest rate of compaction occurred after the initial few passes and reducing the number of trips made over the same trail had no effect in reducing soil compaction. In the other hand, subsequent vehicle passes will result in diminishing extra soil compaction.

Hence, even one pass is already sufficient to induce a strong increase in bulk density. So, skidding operations should be limited to pre-planned skid trails, because vehicle traffic away from skid trails can significantly affect the increasing of the soil bulk density. The results of this research confirmed that preplanning of skid trails and directional felling will reduce ground disturbance. Slope gradient has a significant effect on soil compaction. Based on the results, it can be concluded that uphill slope gradients on trails should be as low as possible, particularly when vehicles are traveling loaded. The study showed that the skid trail slope and vehicle passes had a significant effect on soil compaction. Severe compaction of soil adversely affects the growth of plants by a combination of physical soil changes and plant physiological dysfunctions.

Skidding operations should be planned when soil conditions are dry so as to minimize soil compaction, but if skidding must be done under wet conditions, the operations should be stopped when the vehicle traffic creates sever soil compaction. The distance between these trails must depend on the length of the felled tree and may range between 50 and 70 m distance in order to reduce the winching distance. The impact of felling of large trees is another source of compaction but this aspect has not been studied in this research.

Acknowledgements – Zahvala This paper is one of the results of a research project that was carried out in the period 2010–2012 in the Hyrcanian forest in northern Iran. The authors would like to acknowledge the financial support of Iranian National Science Foundation (INSF) for the research project No. 88001084.

133

B. Majnounian and M. Jourgholami

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135)

6. References – Literatura Adams, P. W., Froehlich, H. A., 1984: Compaction of forest soils. USDA Pacific Northwest Extension Publication. PNW 217, 13 p.

Ampoorter, E., Goris, R., Cornelis, W. M., Verheyen, K., 2007: Im pact of mechanized logging on compaction status of sandy forest soils. Forest Ecology and Management 241(1–3): 162– 174.

Ares, A., Terry, T. A., Miller, R. E., Anderson, H. W., Flaming, B. L., 2005: Ground-Based Forest Harvesting Effects on Soil Physical Properties and Douglas-Fir Growth. Soil Science Society of America Journal 69(6): 1822–1832.

Cullen, S. J., 1991: Timber Harvest Trafficking and Soil Compaction in Western Montana. Soil Science Society of America Journal 55(5): 1416–1421.

Eliasson, L., 2005: Effects of forwarder tire pressure on rut formation and soil compaction. Silva Fennica 39(4): 549–557.

Eliasson, L., Wasterlund, I., 2007: Effects of slash reinforcement of strip roads on rutting and soil compaction on a moist fine-grained soil. Forest Ecology and Management Journal 252(1–3): 118–123.

Froehlich, H. A., McNabb, D. H. 1984. Minimizing soil compaction in northwest forests. Proceedings 6th North American forest soils conference, June 19–23;1983, Knoxville, TN, USA. University of Tennessee, Department of Forestry, Wildlife and Fisheries: 159–192.

Froehlich, H. A., Miles, D. W. R., Robbins, R. W., 1985: Soil bulk density recovery on compacted skid trails in central Idaho.

Soil Science Society of America Journal 49(4): 1015–1017.

Froese, K., 2004: Bulk density, soil strength, and soil disturbance impacts from a cut-to-length harvest operation in north central Idaho. Master Thesis, University of Idaho, USA, 72 p.

Gayoso, J., Iroume, A., 1991: Compaction and soil disturbances from logging in Southern Chile. Annals of Forest Science 48(1): 63–71.

Gent, J. A., Morris, L. A., 1986: Soil Compaction from Harvesting and Site Preparation in the Upper Gulf Coastal Plain. Soil Science Society of America Journal 50(2): 443–446.

Gomez, A., Powers, R. F., Singer, M. J., Horwath, W. R., 2002: Soil Compaction Effects on Growth of Young Ponderosa Pine Following Litter Removal in California’s Sierra Nevada. Soil Science Society of America Journal 66(4): 1334–1343.

Grace, J. M., Skaggs, R. W., Cassel, D. K., 2006: Soil Physical Changes Associated with Forest Harvesting Operations on an Organic. Soil Science Society of America Journal 70(2): 503 – 509.

Greacen, E. L., Sands, R., 1980: Compaction of Forest Soil, A Review. Australian Journal of Soil Research 18(2): 163–189.

Horn, R., Vossbrink, J., Peth, S., Becker, S., 2007: Impact of modern forest vehicles on soil physical properties. Forest Ecology and Management 248(1–2): 56–63.

Jamshidi, R., Jaeger, D., Raafatnia, N., Tabari, M., 2008: Influence of Two Ground-Based Skidding Systems on Soil Compaction under Different Slope and Gradient Conditions. International Journal of Forest Engineering 19(1): 9–16.

Johnson, L. R., Page-Dumroese, D., Han, H. S., 2007: Effects of Machine Traffic on the Physical Properties of Ash-Cap Soils.

Proceedings of U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station: Volcanic-Ash-Derived Forest Soils of the Inland Northwest: Properties and Implications for Management and Restoration. March 2005, Coeur d’Alene, Fort Collins, CO, USA, p. 69–83.

Kolkaa, R. K., Smidt, M. F., 2004: Effects of forest road amelioration techniques on soil bulk density, surface runoff, sediment transport, soil moisture and seedling growth. Forest Ecology and Management 202(1–3): 313–323.

Kozlowski, T. T., 1999: Soil Compaction and Growth of Woody Plants. Scandinavian Journal of Forest Research 14(6): 596– 619.

Krag, R., Higgingbotham, K., Rothwell, R., 1986: Logging and soil disturbance in southeast British Columbia. Canadian Journal of Forest Research 16(6): 1345–1354.

Matangaran, J. R., Kobayashi, H., 1999: The Effect of Tractor Logging on Forest Soil Compaction and Growth of Shorea selanica Seedlings in Indonesia. Journal of Forest Research 4(1): 13–15.

Rab, M. A., 1996: Soil physical and hydrological properties following logging and slash burning in the Eucalyptus regnans forest of southeastern Australia. Forest Ecology and Management 84(1–3): 159–176.

Rab, M. A., Bradshaw, F. J., Campbell, R. G., Murphy, S., 2005: Review of factors affecting disturbance, compaction and trafficability of soils with particular reference to timber harvesting in the forests of south-west Western Australia, Consultants Report to Department of Conservation and Land Management, Western Australia, Sustainable Forest Management Series, SFM Technical Report No. 2, 146 p.

Rohand, K., Kalb, A. A., Herbauts, J., Verbrugge, J. C., 2004: Changes in some mechanical properties of a loamy soil under the influence of mechanized forest exploitation in a beech forest of central Belgium. Journal of Terramechanics 40(4): 235–253.

Saarilahti, M., 2002 Soil Interaction Model. Development of a protocol for ecoefficient wood harvesting on sensitive sites (ECOWOOD). Project deliverable D2 (Work Package No. 1), University of Helsinki, Department of Forest Resource Management, December 2002, 1–87.

Horn, R., Taubner, H., Wuttke, M., Baumgartl, T., 1994: Soil physical properties related to soil structure. Soil and Tillage Research 30(2–4): 187–216.

Sidle, R. C., Drlica, D. M., 1981: Soil Compaction from Logging with a Low-Ground Pressure Skidder in the Oregon Coast Ranges. Soil Science Society of America Journal 45(6): 1219– 1224.

Horn, R., Vossbrink, J., Becker, S., 2004: Modern forestry vehicles and their impacts on soil physical properties. Soil and Tillage Researc 79(2): 207–219.

Šušnjar, M., Horvat, D., Šešelj, J., 2006: Soil compaction in timber skidding in winter conditions. Croatian Journal of Forest Engineering 27(1): 3–15.

134

Croat. j. for. eng. 34(2013)1

Effects of Rubber-Tired Cable Skidder on Soil Compaction... (123–135) Wang, J., LeDoux, C. B., Edwards, P., 2007: Changes in Soil Bulk Density Resulting from Construction and Conventional Cable Skidding Using Preplanned Skid Trails. Northern Journal of Applied Forestry 24(1): 5–8.

Wronski, E. B., Murphy, G., 1994: Responses of forest crops to soil compaction. Soil Compaction in Crop Production. Elsevier Science, Amsterdam, p: 317–342.

B. Majnounian and M. Jourgholami

Ziesak, M., 2006: Avoiding soil damages, caused by forest machines. IUFRO Proceedings: International Precision Forestry Symposium »Precision Forestry in Plantations, Seminatural and Natural Forests«. 5–10 March 2006, Stellenbosch University, South Africa, p. 9.

Sažetak

Utjecaj broja prolazaka kotačnoga skidera na zbijanje tla Upotreba je kotačnih skidera vrlo česta za privlačenje drva iz šume, ali pritom vozilo negativno utječe na okoliš.

Cilj je istraživanja bio ispitati kako različiti nagibi terena (odnosno traktorskih vlaka), broj prolazaka vozila po traktorskoj vlaci i dubina tla utječu na zbijanje tla. Istraživano je zbijanje tla na četiri nagiba traktorskih vlaka: ravan teren, nagibi od +10 %, -10 %, -20 %; na tri različite dubine tla: 5, 15 i 25 cm te s obzirom na broj prolazaka vozla po vlaci: 0, 1, 5, 8, 10, 15, 20, 25 i 30 (slika 7). Uzorci su tla uzimani unutar traktorske vlake i izvan nje (slika 4) svakih 25 m duž vlake kako bi se vidio utjecaj natovarenoga vozila na tlo. Istraživanje je provedeno u Nastavno-pokusnom šumskom objektu Kheyrud, koji se nalazi unutar Hirkanijske šume u sjevernom Iranu, a drvo je privučeno skiderom Timberjack. Traktorske su vlake zauzimale 5,8 % ukupne površine istraživanoga područja (17 ha) uz dodatnih 0,8 % površine potrebne za pomoćno stovarište. S povećanjem broja prolazaka vozila povećala se i gustoća tla, ali je ipak najveće zbijanje tla ustanovljeno u prvih nekoliko prolazaka vozila. Privlačenje drva uzbrdo (+10 % nagiba terena) više je zbijalo tlo (slike 5 i 6) nego privlačenje drva nizbrdo (nagibi terena -10 % i -20 %). Povećanje je gustoće tla i na dubini od 20 do 30 cm bilo značajno (slika 8). Gustoća je tla na dubini od 5, 15 i 25 cm bila veća za 35, 22 i 17 % od gustoće tla na netaknutom tlu (slika 9). Kako bi se smanjilo zbijanje tla, potrebno je privlačiti drvo po unaprijed planiranim i za to predviđenim traktorskim vlakama te usmjereno obarati stabla kako bi se smanjilo kretanje vozila po šumskom bespuću jer već i nakon prvoga prolaska vozila dolazi do povećanja gustoće tla i njegova zbijanja. Također, privlačenje se drva treba odvijati u uvjetima suhoga tla kako bi se smanjila oštećenja na šumskom tlu.

Ključne riječi: zbijanje tla, gustoća tla, kotačni skider, Hirkanijske šume

Authors’ address – Adresa autorâ:

Received (Primljeno): February 12, 2012 Accepted (Prihvaćeno): Jaqnuary 11, 2013 Croat. j. for. eng. 34(2013)1

Prof. Baris Majnounian, PhD. e-mail: bmajnoni@ut.ac.ir Asst. Prof. Meghdad Jourgholami, PhD.* e-mail: mjgholami@ut.ac.ir University of Tehran Faculty of Natural Resources Department of Forestry and Forest Economics P.O.Box: 31585–431Karaj IRAN * Corresponding author – Glavni autor

135

Original scietific paper – Izvorni znanstveni rad

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia Aleš Kadunc Abstract – Nacrtak Rot is one of the most important defects in timber assortments of Norway spruce (Picea abies).

With the aim to investigate this phenomenon, we analyzed 1,334 spruce trees from 65 locations across Slovenia. The study showed that the incidence of rot was higher in even-aged stands compared to uneven-aged stands. Lower-trunk rot (for the purpose of this study, lower-trunk rot is rot that affects the stem up to 5 m above ground) is significantly connected with the diameter of the tree, whereas rot in the upper part of the trunk, i.e. upper-trunk rot, does not show such dependence. Rot is most commonly associated with dolomite bedrock. Its incidence increases with tree age, site productivity, and site altitude. Mechanical stem injury is another contributing factor. On the other hand, slender trees, sun-exposed sites, and uneven-aged stands carry a lower risk for the studied defect. The incidence of rot may decrease the value of usable timber by as much as 19 €/m3, with the decrease being highest in 50 to 70 cm tree diameter.

Keywords: rot, Norway spruce, Slovenia.

1. Introduction – Uvod In North and Central Europe, Norway spruce (Picea abies (L.) Karst.) is generally considered as the leading tree species in terms of economic value (Kenk and Guehne 2001; Jöbstl 2011). However, the species’ distribution outside its natural sites and the unstable structure of its stands (monocultural stands, evenaged stands) have prompted forestry to look for suitable methods to convert these forests into more stable and more natural structures (Knoke and Plusczyk 2001). Despite the high risks associated with artificial spruce stands and their questionable value in terms of ecology, contemporary studies have confirmed that a significant addition of spruce in broadleaved stands is economically justifiable (Knoke et al. 2008).

The incidence of rot is one of the most important problems in spruce management (Kohnle and Kändler 2007). The organisms responsible for the decay process are various fungi, including Heterobasidion annosum s.l., Armillaria spp., and Stereum sanguinolentum (Graber 1996; Čermák et al. 2004, Korhonen and Holdenrieder 2005; Arhipova et al. 2011). The decay normally starts through the infection of the roots, which then spreads to the trunk. Alternatively, a tree may become infected through a wound in its bark or stem (Kohnle in Kändler 2007), which either occurs during mechanical harvesting (Ivanek 1976; Vasiliauskas 2001) or is caused, in certain cases, by animals fraying against the tree or consuming its bark (Čermák et al.

2004).

The high commercial value of spruce, as well as numerous dilemmas and issues related to the management of forest stands, where spruce holds a significant or even predominating share, have become the driving force following the analyses of the quality and value characteristics of Norway spruce and its stands.

Assessments of timber value losses clearly demonstrate that spruce rot is a serious economic issue. For EU countries, value loss due to rot caused by the fungus Heterobasidion annosum s. l. (Woodward et al. 1998) was assessed at EUR 790 million for 1998, but may be well above that mark (Seifert 2007).

Croat. j. for. eng. 34(2013)1

137

A. Kadunc

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

The negative effects of spruce rot, however, reach beyond timber assortment structure (Gonthier et al.

2012); they affect volume increment growth (Oliva et al. 2010) and mechanical stability (Dimitri in Tomiczek 1998) of stands, and increase the cost of harvesting operations (additional log shaping).

The occurrence of the disease is associated with a variety of factors, including past land use, stand characteristics, soil factors, and silvicultural measures (Woodward et al. 1998; Jurc 2001). Nevertheless, the results of the studies are often inconclusive (Rönnberg and Jørgensen 2000) and the disease has not been thoroughly studied in all its complexity (Seifert 2007; Arhipova et al. 2011). Furthermore, a comprehensive study of spruce rot would need to take into account considerable differences in the studied rot that exist between climatic regions and soil conditions (Seifert 2007; Gonthier et al. 2012).

Spruce rot can be caused by a variety of causal agents, which the studies find difficult to distinguish reliably (Kohnle and Kändler 2007). Also, fungal taxonomy is a complex task requiring extensive laboratory work (Arhipova et al. 2011). In the light of the above, the present study only focuses on determining the presence of rot in a tree, without studying its cause and decay rate.

The aim of this study was to (1) determine the incidence of rot in spruce trees by tree diameter, stand type, and site conditions; (2) illustrate the effect of rot on loss in timber value; and (3) outline the stand characteristics which may lead to the occurrence of rot.

2. Matherials and Methods – Materijal i metode 2.1 Study area – Područje istraživanja The aim of the research was to cover the maximum range of growth site conditions under which Norway spruce becomes an important element for forest stands. The sample included even-aged and unevenaged stands, with the main focus on mature stands.

Aware of the fact that only felled trees would be analyzed and that the choice of trees to be harvested was often rather subjective (Matić et al. 2003; Prka 2006), the study area was expanded to include a higher number of locations, thereby covering a large part of Slovenian territory where spruce is a common element of commercial forests. The first focal area covered the forests of North Slovenia in the Alpine arc (Gorenjska, Koroška, Pohorje), and the second encompassed the Dinaric part of South Slovenia (Dolenjska, Kočevska);

138

the study plots in Central Slovenia are relatively dispersed.

In total, the study included spruce trees from 65 sites, i.e. 20 different plant associations, which allowed us to cover all the major site strata. In the course of the study, a total of 1,334 spruce trees were felled, of which 1,015 came from even-aged stands and 319 from uneven-aged stands. The average diameter at breast height (hereafter: DBH) amounted to 47.5 cm (maximum DBH was 98.5 cm) and the average height was 30.8 m (maximum height 53.8 m). The average age of the analyzed trees was 112.8 years, and the oldest analyzed tree was 224 years old.

2.2 Research methods – Metode istraživanja Prior to the felling, all sample trees were measured for diameter at breast height (DBH) and checked for other special characteristics, e.g. forked tops, stem or butt log damage (mechanical injury), and broken tops.

During harvesting operations, stem analyses of sample trees were carried out (Pranjić and Lukić 1997; Husch et al. 2003; Kahle et al. 2008). Each section of the stem was classified into a quality class according to the JUS standard for coniferous logs (D.B4.029 1979). This standard was used because it continues to be the prevalent standard used in timber trade in Slovenia. The length from the stump to the crown base was also measured. The obtained data were used to calculate the relative crown length (RCL) with the following formula: Relative crown length: (tree height – crown base)/ tree height Stem analysis data were used to calculate the mean ring width and slenderness ratio (tree height /diameter at breast height).

Each stem cross-section was checked for rot, and the presence or absence of rot was recorded. However, as the extension of rot was not consistently measured in all the sampled trees, these data were not included in the analysis. The stage of wood decay was also not consistently assessed. For the above reasons, the occurrence of rot was only studied as a binary variable (present/not present). However, since all stem sections were checked for signs of the disease, we were able to distinguish between the rot that affected the lower part of the trunk up to 5 m from the ground (hereafter: lower-trunk rot), and rot in the upper sections of the trunk, above the 5 m mark (hereafter: upper-trunk rot).

Clearly, in certain spruce trees, rot was present in the lower as well as upper sections of the tree. However, as spruce rot is more common in the lower parts of the trunk and lower-trunk rot has the highest effect on Croat. j. for. eng. 34(2013)1

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

timber value, most analyses were only carried out for this form of disease.

At each harvest site an appropriate number of dominant (thickest) trees were also harvested, in addition to other trees, to provide the data required for the calculation of the site index (Kotar 2005; Charru et al. 2010). Using standardized methodology (Kotar 2005; Van Laar and Akça 2007), the site index for reference age of 100 years (hereafter: SI100 ) was determined by measuring the tree height and age.

The value of wood at the roadside was calculated by multiplying the volume of each assortment class by the price for that quality class. To calculate the prices for timber assortments at the roadside, we first calculated the current average prices offered by 10 Slovenian forestry companies dealing in timber trade and relevant for the studied tree species.

The following price list was used for the calculation of prices at the roadside: Þ veneer logs 205.68 €/m3, Þ sawlogs, class I, 92.00 €/m3, Þ sawlogs, class II, 66.58 €/m3, Þ sawlogs, class III, 54.31 €/m3, Þ pulpwood, 24.23 €/m3.

With a view to calculating the differences in value between rot-infected trees and healthy trees by individual strata, we first used a second order parabola, commonly applied in similar analyses (Prka 2003), to determine the dependence of the price at the roadside on DBH. In the second step, the values for rot-infected trees (by strata) were deducted from the values for trees without rot.

On account of the high number of forest associations or site units included in the study, the results of these analyses were divided by stand type (even-aged and uneven-aged stands), and type of bedrock (limestone, dolomite, silicate).

After studying the scatter graph data, we decided to use linear regression to calculate the dependence of the share of rot-affected trees on diameter class.

Then, the study focused on determining the factors associated with the incidence of rot (in the lower and upper sections of the stem). The following parameters were tested as predictors: DBH, height, age, slenderness ratio, average ring width, relative crown length, SI100 (in parabolic form), altitude, sun exposure (a dichotomous variable: trees from south facing locations are coded 1, other trees 0), trunk damage (a dichotomous variable: trees with wounded stems are coded 1, other trees 0), top breakage (a dichotomous variable: trees with broken tops are coded 1, other trees 0), Croat. j. for. eng. 34(2013)1

A. Kadunc

forked tops (a dichotomous variable: trees with forked tops are coded 1, other trees 0), stand type (a dichotomous variable: trees from even-aged stands are coded 1, other trees 0) and bedrock (two dummy variables were formed: silicate, where silicate bedrock is coded 1, and limestone and dolomite are coded 0, and dolomite, where dolomite bedrock is coded 1, and the other two are coded 0).

The probability of rot presence (lower-trunk rot or upper-trunk rot) was modeled using a binary logistic regression. Rot-affected trees were coded 1 and healthy trees were coded 0. The model was estimated in PASW SPSS Statistics 18 using the backward stepwise procedure where removal testing was based on the probability of the likelihood-ratio statistic based on the maximum partial likelihood estimates (Kleinbaum and Klein 2002). The Nagelkerke pseudo-R2 was used to evaluate goodness-of-fit for the model.

All statistical analyses were conducted in PASW SPSS Statistics 18.

3. Results – Rezultati 3.1 Occurrence of rot with regard to stem part, stand type and bedrock – Pojava truleži s obzirom na dio stabala, tip sastojine i matičnu podlogu Signs of rot were detected in more than a third of all spruce trees in the analyzed sample (Table 1). The data show that the occurrence of lower-trunk rot and upper-trunk rot is relatively more frequent in evenaged stands. Considerable differences exist, however, in the frequency of lower-trunk rot and upper-trunk rot: lower-trunk rot was found in 34% and 26% of all trees in the sample in even-aged and uneven-aged stands, respectively, whereas in only 14% and 11% of trees rot was found in the upper part of the stem.

In even-aged stands, the share of trees that have succumbed to lower-trunk rot is substantial already in relatively small trees. In trees from uneven-aged stands, the share of trees with lower-trunk rot is initially very low, but rises markedly with the diameter.

In both stand types, the share of trees with lower-trunk rot varies considerably between diameter classes.

The share of trees with upper-trunk rot is moderate in relatively small trees, but shows no clear trend when the diameter increases.

The increase in DBH has no effect on the increase in the share of upper-trunk rot (Fig. 1). The grey straight lines are not statistically characteristic (Appendix 1); however, they indicate that the disease is slightly less common in uneven-aged stands. The in-

139

A. Kadunc

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

Table 1 Share of rot-infected trees by diameter class, with regard to stand type and rot location Tablica 1. Udjel natrulih stabala po debljinskim razredima s obzirom na vrstu sastojine i mjesto javljanja truleži

Lower and upper-trunk rot

Broj stabala

Lower and upper-trunk rot

Udjel natrulih stabala, %

Number of analyzed trees

Udjel natrulih stabala, %

Share of rot-infected trees, % Broj stabala

Share of rot-infected trees, %

Uneven-aged stand – Raznodobna sastojina Number of analyzed trees

Diameter class

Debljinski razred

Even-aged stand – Jednodobna sastojina

Lower-trunk rot

Upper-trunk rot

Trulež u donjem dijelu debla

Trulež u gornjem dijelu debla

–

0.0

0.0

0.0

5

–

–

0.0

0.0

0.0

5

2.7

18.9

37

0.0

12.5

12.5

8

15.2

6.1

18.2

33

0.0

0.0

0.0

5

22.5

41.4

17.2

48.3

29

7.7

30.8

38.5

13

27.5

26.7

22.2

37.8

45

16.7

0.0

16.7

12

32.5

19.7

14.5

31.6

76

5.6

11.1

16.7

18

37.5

38.6

17.1

50.0

140

0.0

14.3

14.3

14

42.5

40.8

16.7

46.7

120

26.5

26.5

47.1

34

47.5

31.3

15.3

43.5

131

31.3

6.3

34.4

32

52.5

36.1

12.6

45.4

119

44.0

4.0

44.0

25

57.5

37.5

15.0

46.3

80

36.7

10.0

40.0

30

62.5

31.3

10.0

35.0

80

35.7

3.6

35.7

28

67.5

36.5

19.2

42.3

52

27.0

5.4

32.4

37

72.5

42.9

2.9

42.9

35

40.0

6.7

40.0

15

77.5

57.9

21.1

68.4

19

35.7

7.1

35.7

14

82.5

41.7

25.0

50.0

12

30.8

30.8

53.8

13

87.5

50.0

16.7

50.0

6

25.0

25.0

25.0

8

92.5

100.0

0.0

100.0

1

–

–

–

–

97.5

–

–

–

–

33.3

0.0

33.3

3

34.2

14.5

42.4

1015

26.0

11.0

33.2

319

Lower-trunk rot

Upper-trunk rot

Trulež u donjem dijelu debla

Trulež u gornjem dijelu debla

2.5

–

–

–

7.5

–

–

12.5

18.9

17.5

cm

Total Ukupno

Trulež na oba kraja debla

cidence of lower-trunk rot rises with DBH, reaching higher shares in even-aged stands. In these stands, the disease is quite common in small trees. On the other hand, the incidence of rot rises more markedly in uneven-aged stands, but even in large-diameter trees the value stands below the figures for even-aged stands.

A similar upward trend is also observed in the share of trees where rot is present in the lower or upper part

140

Trulež na oba kraja debla

of the trunk, mainly due to the increase in the share of trees showing signs of lower-trunk rot. In trees with DBH greater than 50 cm, the share of lower-trunk rot stands at 30%, and pushes towards 40% or higher in large-diameter trees. Regression parameters are given in Appendix 1. The regression analysis only explains a small part of the variability, which confirms its stochastic nature.

Croat. j. for. eng. 34(2013)1

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

A. Kadunc

Fig. 1 Share of rot-infected trees with regard to stand type and rot area Slika 1. Udjel natrulih stabala s obzirom na vrstu sastojine i mjesto javljanja truleži Table 2 Relationship between lower-trunk rot and upper-trunk rot (contingency test) Tablica 2. Povezanost pojave truleži u donjem i gornjem dijelu debla (test slučaja) Upper-trunk rot – Trulež u gornjem dijelu debla Not present – Nije prisutna c2 = 8.752 P = 0.003

Lower-trunk rot Trulež u donjem dijelu debla

Not present Nije prisutna

Total – Ukupno

Present Prisutna

Total – Ukupno

Observed frequency

Expected frequency

Observed frequency

Expected frequency

Observed frequency

Expected frequency

Izmjerena frekvencija

Očekivana frekvencija

Izmjerena frekvencija

Očekivana frekvencija

Izmjerena frekvencija

Očekivana frekvencija

798

780.7

106

123.3

904

904

354

371.3

76

58.7

430

430

1152

1152

182

182

1334

1334

In the next step of the study, we decided to verify whether the presence of lower-trunk rot is connected with the occurrence of rot above the 5 m mark. To this aim, a contingency test was conducted (Table 2), which showed that the relationship between the phenomena is statistically significant. If rot has affected the lower part of the stem, the probability that it will also affect Croat. j. for. eng. 34(2013)1

Present – Prisutna

the upper sections of the trunk is higher, and vice versa.

Taking into account the higher incidence of lowertrunk rot compared to upper-trunk rot and close relationship between the two phenomena, we decided to carry out a more detailed analysis into the dependence of lower-trunk rot upon stand type, bedrock and di-

141

A. Kadunc

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

Fig. 2 Share of trees with lower-trunk rot with regard to stand and bedrock type Slika 2. Udjel natrulih stabala (trulež u donjem dijelu debla) s obzirom na vrstu sastojine i matičnu podlogu ameter (Fig. 2). The differences between bedrock types are substantial. Even for small trees, the share of rot is high in even-aged stands on limestone and dolomite, and obviously increases with diameter. On limestone sites (even-aged stands), the share of rot-infected trees rises slowly. Large-diameter trees growing on lime-

stone bedrock show no difference between even-aged and uneven-aged stands. On silicate bedrock, the share of rot is low in small trees, particularly in uneven-aged stands. It increases with the diameter in both stand types, but the difference between stand types is preserved. In large-diameter trees, rot inci-

Fig. 3 Differences in average usable timber values by stand and bedrock type Slika 3. Razlike u vrijednosti drva s obzirom na vrstu sastojine i matičnu podlogu

142

Croat. j. for. eng. 34(2013)1

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

A. Kadunc

Table 3 Parameters for the binary logistic regression (for the left part the outcome variable is lower-trunk rot and for the right part the outcome variable is upper-trunk rot) Tablica 3. Parametri binarne logističke regresije (za lijevi dio ishodna je varijabla: trulež u donjem dijelu debla, a za desni dio ishodna je varijabla: trulež u gornjem dijelu debla Predictor – Prediktor

Lower-trunk rot – Trulež u donjem dijelu debla

Upper-trunk rot – Trulež u gornjem dijelu debla

b

Exp(b)

P

b

Exp(b)

P

Constant – Stalnica

-3.961

0.019

0.0000

-5.715

0.003

0.0000

Stand type – Vrsta sastojine

0.275

1.317

0.0976

–

–

–

SI100 – Bonitet staništa

0.068

1.070

0.0000

0.055

1.057

0.0134

Dolomite – Dolomit

0.841

2.318

0.0000

1.282

3.603

0.0000

–

–

–

1.849

6.354

0.0000

Altitude – Nadmorska visina

0.001

1.001

0.0006

–

–

–

South-facing site – Stanište s južnom ekspozicijom

-0.285

0.752

0.0936

–

–

–

Age – Dob

0.007

1.007

0.0105

–

–

–

Slenderness ratio – Mjera vitkosti

-0.019

0.981

0.0005

–

–

–

Trunk damage – Mehaničke ozljede debla

1.696

5.450

0.0000

1.101

3.009

0.0000

Top breakage – Slomljen vrh

–

–

–

3.195

24.409

0.0000

Forked tops – Rašljavo stablo

–

–

–

2.424

11.296

0.0000

Silicate – Silikat

dence is highest on dolomite bedrock (even-aged stands), followed by even-aged stands on silicate ground, and uneven-aged stands on the same bedrock. In general, rot is less frequent in uneven-aged stands. As regards tree diameter, rot affects about 45% of trees with DBH about 50 cm on dolomite ground, between 25% and 35% of such trees on silicate bedrock, and between 25% and 30% of such trees on limestone. In large-diameter trees, the share of rot on dolomite pushes above 50% and above 40% on silicate.

The regression parameters are given in Appendix 1.

The regression analysis explains a relatively small part of variability (Appendix 1).

3.2 Value loss due to rot − Gubitak vrijednosti drva zbog pojave truleži Value loss, i.e. the difference in the value of spruce timber due to rot, was determined for all strata for which sufficient data was obtained, in relation to the tree diameter (Fig. 3). The loss in €/m3 is highest on silicate ground, in even-aged stands, and on limestone, in uneven-aged stands. It is lowest, however, in even-aged stands on limestone bedrock. The loss is relatively low in small trees (less than 8 €/m3), and considerably higher in trees with DBH between 50 and 70 cm (15–19 €/m3).

In larger diameter trees, the loss drops again. Regression parameters are given in Appendix 1.

Croat. j. for. eng. 34(2013)1

3.3 Value factors affecting the incidence of rot Čimbenici koji utječu na pojavu truleži A binary logistic regression method was used to test both types of disease, lower-trunk rot and uppertrunk rot, for factors affecting its occurrence. Lowertrunk rot is more likely to occur on high-productivity sites, on dolomite bedrock, at higher altitude, in older and wounded trees (Table 3, left). Tree slenderness reduces the probability of lower-trunk rot. The results also indicated that the disease is more likely to occur in even-aged stands and less on south-facing sites. The model explained a total of 17.8% pseudovariance (Nagelkerke’s R2 = 0.178).

Upper-trunk rot is also more likely to develop on high-productivity sites, on dolomite and silicate bedrock, in wounded trees, trees with broken tops, and trees with forked stems (Table 3, right). The model explained a total of 30.0% of pseudovariance. (Nagelkerke’s R2 = 0.300).

4. Dicussion and conclusion – Rasprava i zaključci Rot infection in trees can be assessed using a variety of methods. These can be either destructive or invasive (tree felling, extracting increment cores) or non-

143

A. Kadunc

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

destructive (vibro-acoustic diagnostics, electromagnetic radiation methods) (Oliva et al. 2011).

According to the existing study results, non-destructive methods lack accuracy (Oliva et al. 2011) and, surprisingly, the results of extracting cores have also shown to be relatively unreliable (Stenlid and Wästerlund 1986).

The present study has shown that the frequency of rot mainly increases with the diameter at breast height.

However, the relationship between the two is not tight, in particular on limestone bedrock and on high-productivity sites. Besides, the incidence of rot in the upper sections of the stem is not related to tree diameter.

The analyses conducted by Kotar (2006) also confirmed the high stochastic value of rot in Norway spruce, pointing to a high local variability of the phenomenon. This supports the finding that rot occurs in clusters of trees (Piri et al. 1990). Most studies state that rot is more likely to be found in large stumps (Arhipova et al. 2011; Gonthier et al. 2012). There are two possible explanations for that: first, infected trees try to compensate for the loss of their decaying tissue through faster increment growth, or, secondly, large trees are more at risk because their growth is faster and they are quick to get in contact with infected stumps (Arhipova et al. 2011).

Our results show that rot is less frequent in uneven-aged stands, which confirms the findings of other studies (Gorše 2009) and, indirectly, supports the observation that the quality of usable timber is higher in selection forests (Knoke 1998; Hanewinkel 2001).

Mechanical injury to trees (trunk and roots) resulting from timber harvesting is closely related to the occurrence of rot (Ivanek 1976; Vasiliauskas 2001). A German study showed that practically all spruce trees with bark injury were infected with rot (Kohnle and Kändler 2007). The latter points to a high importance of careful planning and implementation of forestry operations in spruce stands, since the share of wounded trees in these stands increases greatly with the number of thinning operations (Košir 1998, 2008), whereas the extent and type of wound damage depend on the selected technology and management intensity (Fjeld and Granhus 1998; Košir 2008).

The occurrence of rot is associated with considerable losses in usable timber value, in particular in Norway spruce (Piškur 2001). The study showed that the loss amounted to 20 €/m3, which is 5–25% of timber value. In Italy the direct financial loss caused by rot was assessed at 18–34% of timber value (Gonthier et al. 2012). Piškur (2001) stated a 23% decrease in the timber value of spruce trees which were subject to a mechanical injury that normally leads to the decay of

144

the wounded part. This drop in price was considerably higher than observed in fir, which is less vulnerable to mechanical injury (Kohnle and Kändler 2007). In Scandinavian sites, rot is responsible for a 2–9% loss in timber values on the stump, depending on the degree of infection, (Möykkynen et al. 1998), but in most cases the infected trees were low in diameter. According to this study, the value loss is highest with regard to the trees with DBH between 50 and 70 cm. In largediameter trees, loss due to rot is slightly lower, primarily because these trees are normally of lower quality, largely on account of their high branch quantity and conical shape.

The share of rot-infected trees could be considerable already in young spruce trees or in trees of low diameter. In Latvia, for example, the incidence of rot in 30-year-old trees ranged from a few specimens to 30% of trees (Arhipova et al. 2011). As for Germany, every second tree in spruce stands older than 20 years has been found to be affected by an agent causing decay (Dimitri and Tomiczek 1998).

In the event of rot occurring as a result of bark removal by deer, the share of rot-infected trees is highest in polewood stands, but drops in higher-diameter trees due to salvage cutting (Čermák et al. 2004). In trees with mechanical injury the damage reaches 2–4 m up the stem, and rot develops in more than a half of injured trees. Injuries which make the tree particularly vulnerable to rot are those where the injury affects the wood tissue below the removed bark (Pawsey and Stankovicova 1974).

The rot, spreading from the roots or butt log, normally reaches from 2.2 to 6.5 m up the stem in Germany (Zycha et al. 1970), 4.3 m in Southern Finland (Tamminen 1985), and between 2.3 and 2.7 m in France (Perrin and Delatour 1976). Based on these findings, we decided to distinguish between the incidence of rot in the lower five meters of the stem (lower-trunk rot) and above that mark (upper-trunk rot), as we predicted that the incidence of rot in the higher sections of the stem would be largely driven by a different set of factors.

The study showed that rot in the higher sections of the trunk mainly resulted from a mechanical stem injury, broken tree top or forked top. The latter is caused by a dying off or breaking of the terminal shoot in the past. Furthermore, upper-trunk rot is more common on high-productivity sites and on silicate and dolomite bedrock. High-productivity sites tend to be more intensively managed, which increases the risk of injury to the stand. Higher rot incidence on silicate and dolomite sites can be explained through a higher content of sand and lower content of organic matter compared Croat. j. for. eng. 34(2013)1

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

to limestone sites (Kralj 2008), which increases the probability of rot incidence (Stenlid and Redfern 1998).

On silicate and dolomite sites, where bedrock outcrops are rare, agricultural use was frequently predominating in the past, although silicate bedrock is also characterized by a long tradition of spruce management in many places. Both causes, past agricultural use and succession of (artificial) spruce stands, increase the incidence of rot (Woodward et al. 1998; Jurc 2001).

The occurrence of rot in the lower-trunk rot is more likely on dolomite bedrock, on high-productivity sites, at higher altitude, and in older stands, for less slender trees and mechanically injured trees. Rot is more frequent in even-aged stands and less frequent on sunfacing sites.

In the past, dolomite sites were often intended for agricultural use, which contributed to higher rot incidence. On the other hand, limestone is more favorable for the occurrence of rot because of lower sand share and higher content of organic matter (Kralj 2008). High sand content, low organic matter content, and high pH all contribute to the occurrence of rot (Stenlid and Redfern 1998).

Increased rot incidence on high-productivity sites has already been established in previous studies (Korhonen and Stenlid 1998; Thor et al. 2005; Mattila and Nuutinen 2007), although it was not confirmed by all (Nilsen 1983). Site productivity may have an indirect effect on the incidence of rot, in particular because highly productive sites tend to be more intensively managed, which increases the probability of injury and, consequently, the incidence of rot.

This study confirmed the positive effect of altitude on the incidence of rot, which is in contradiction to several other studies (Korhonen in Stenlid 1998; Jurc 2001), although certain more recent studies point to a high share of rot-infected trees at high elevations (Gonthier et al. 2003).

The age-related increase in rot incidence corresponds with the findings of previous studies (Thor et al. 2005, Arhipova et al. 2011).

In our study sample, slender trees were more rarely infected with rot, which may be due to the fact that these trees are normally younger, thinner, and grow more slowly (narrower rings), all of these negatively affecting the occurrence of rot. A recent study conducted in Switzerland has shown that slowly-growing spruce trees reach an older age. Evidently, a hypothesis was developed that fast-growing trees are more vulnerable to infections with the fungi Heterobasidion annosum and Armillaria sp. (Rötheli et al. 2011).

Croat. j. for. eng. 34(2013)1

A. Kadunc

The well-established connection between mechanical injury and rot was confirmed a long time ago and has been explained in detail (Vasiliauskas 2001; Kohnle and Kändler 2007).

According to this study, the incidence of rot is lower on sun-facing sites, which corresponds to the findings of most recent research (Korhonen and Stenlid 1998).

When a considerable number of spruce trees in a stand are infected with rot, the recommended measures are as follows: a shorter production period (Bachmann 1968; Korhonen et al. 1998; Möykkynen et al. 2000), thinning when the temperatures fall below zero (Jurc 2001), well considered thinning concept (Möykkynen in Miina 2002), replacement of Norway spruce with other tree species (Korhonen et al. 1998; Kohnle and Kändler 2007), measures to minimize mechanical injury during timber harvesting (Korhonen et al. 1998), and removal or coating of infected stumps (Korhonen et al. 1998; Möykkynen et al. 2000). Several researchers propose that economically over-mature stands with already high rot incidence should be assigned the status of protected forests (Arhipova et al. 2011).

The present study has clearly highlighted the complexity of any research aimed at explaining the occurrence of rot. It is a stochastic issue, associated with a number of inter-related factors. To a certain extent, the results of the study could be improved through consistent recording of the rot type, rot rate, and extent of rot infection. The occurrence of rot remains one of the crucial issues in Norway spruce management, but its extent and frequency can be reduced through the formation of more appropriate stand structures, well considered tending, and careful planning and implementation of timber harvesting.

5. References – Literatura Arhipova, N., Gaitnieks, T., Donis, J., Stenlid, J., Vasaitis, R., 2011: Butt rot incidence, causal fungi, and related yield loss in Picea abies stands in Latvia. Can. J. For. Res. 41(12): 2337– 2345.

Bachmann, R. P., 1968: Untersuchungen zur Wahl des Verjüngungszeitpunktes im Waldbau. Dissertation, ETH Zürich, 112 p.

Charru, M., Seynave, I., Morneau, F., Bontemps, J. D., 2010: Recent changes in forest productivity: An analysis of national forest inventory data for common beech (Fagus sylvatica L.) in north-eastern France. Forest Ecology and Management 260(5): 864–874. Čermák, P., Jankovský, L., Glogar, J., 2004: Progress of spreading Stereum sanquinolentum (Alb. et Schw.: Fr.) Fr. wound rot and its impact on the stability of spruce stands.

Journal of Forest Science 50(8): 360–365.

145

A. Kadunc

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

Dimitri, L., Tomiczek, C., 1998: Germany and Austria. In: Heterobasidion annosum− Biology, Ecology, Impact and Control (Woodward, S., Stenlid, J., Karjalainen, R., Hüttermann, A., Eds.), CAB International, Wallingford, UK, p. 355–368.

Fjeld, D., Granhus, A., 1998: Injuries After Selection Harvesting in Multi-Stored Spruce stands – The Influence of Operating Systems and Harvest Intensity. Journal of Forest Engineering 9(2): 33–40.

Gonthier, P., Garbelotto, M., Nicolotti, G., 2003: Swiss stone pine trees and spruce stumps represent an important habitat for Heterobasidion annosum spp. in subalpine forests. Forest Pathology 33(3): 191–203.

Gonthier, P., Brun, F., Lione, G., Nicolotti, G., 2012: Modelling the incidence of Heterobasidion annosum butt rots and related economic losses in alpine mixed naturally regenerated forests of northern Italy. Forest Pathology 42(1): 57–68.

Gorše, G., 2009: Rast in struktura raznomernih sestojev na rastišču dinarskega jelovo-bukovega gozda v GGE Poljane.

Diplomsko delo, Univerza v Ljubljani, Biotehniška fakulteta, Odelek za gozdarstvo in obnovljive gozdne vire, 61 p.

Graber, D., 1996. Die Kernfäuleschäden an Fichte (Picea abies Karst.) in der Schweiz nördlich der Alpen. Beiheft Schweiz.

Zeitschr. F. Forstwesen 79: 1–283.

Hanewinkel, M., 2001: Financial returns of selection forest enterprises with high proportions of valuable timber – results of an empirical study and their application. Swiss Forestry Journal 152(8): 343–349.

Husch, B., Beers, T. W., Kershaw, J. A., 2003: Forest mensuration. John Wiley & Sons, Inc., Fourth edition, New Jersey, USA, p. 368–371.

Ivanek, F., 1976: Vrednotenje poškodb pri spravilu lesa v gozdovih na Pohorju. Inštitut za gozdno in lesno gospodarstvo pri Biotehniški fakulteti v Ljubljani, Strokovna in znanstvena dela, 51, 195 p.

Jöbstl, H. A., 2011: Do altered prices and logging costs for logs of small diameters affect the optimum rotation of Norway spruce in mountainous regions? Forest Policy and Economics 13(4): 266–272.

JUS D.B4.029, 1979. Trupci četinara za rezanje. Pravilnik br.

31 – 11078/1 od 22. VI. 1979, Službeni list SFRJ, br. 32/79.

Jurc, D., 2001: Rdeča trohnoba. Povzročitelji, opis bolezni in ukrepi proti njej. Strokovna monografija, Gozdarski inštitut Slovenije, Ljubljana, 36 p.

Kahle, H. P., Spiecker, H., Unseld, R., Pérez-Martínez, P.-J., Prietzel, J., Mellert, K. H., Straussberger, R., Rehfuess, K. E., 2008: Sampling, measurements and analysis methods. In: Causes and Consequences of Forest Growth Trends in Europe (Kahle, H. P., Karjalainen, T., Schuck, A., Ågren, G., Kellomäki, S., Mellert, K., Prietzel, J., Rehfuees, K. J., Spiecker, H., Eds.), European Forest Institute Research Report 21, Brill, Leiden-Boston, p. 35–45.

146

Kenk, G., Guehne, S., 2001: Management of transformation in central Europe. Forest Ecology and Management 151(1–3): 107–119.

Kleinbaum, D. G., Klein, M., 2002: Logistic Regression: A Self-Learning Text, Second Edition. Springer Atlanta, USA, 513 p.

Knoke, T., 1998: Analyse und Optimierung der Holzproduktion in einem Plenterwald. Dissertation, Ludwig-Maximilians-Universitaet Muenchen, Forstliche Forschungsberichte, 182 p.

Knoke, T., Ammer, C., Bernd, S., Mosandl, R., 2008: Admixing broadleaved to coniferous tree species: a review on yield, ecological stability and economics. European Journal of Forest Research 127: 89–101 Knoke, T., Plusczyk, N., 2001: On economic consequences of transformation of a spruce (Picea abies (L.) Karst.) dominated stand from regular into irregular age structure. Forest Ecology and Management 151(1–3): 163–179.

Kohnle, U., Kändler, G., 2007: Is Silver fir (Abies alba) less vulnerable to extraction damage than Norway spruce (Picea abies)? European Journal of Forest Research 126: 121–129.

Korhonen, K., Holdenrieder, O., 2005: Neue Erkenntnisse über den Wurzelschwamm (Heterobasidion annosum s. l.) - Eine Literaturübersicht. Forst Holz 60: 206–210.

Korhonen, K., Delatour, C., Greig, B. J. W., Schönhar, S., 1998: Silvicultural control. In: Heterobasidion annosum − Biology, Ecology, Impact and Control (Woodward, S., Stenlid, J., Karjalainen, R., Hüttermann, A., Eds.), CAB International, Wallingford, UK, p. 283–314.

Korhonen, K., Stenlid, J., 1998: Biology of Heterobasidion annosum. In: Heterobasidion annosum− Biology, Ecology, Impact and Control (Woodward, S., Stenlid, J., Karjalainen, R., Hüttermann, A., Eds.), CAB International, Wallingford, UK, p.

43–70.

Košir, B., 1998: Presoja koncepta zgodnjih redčenj z vidika porabe energije in poškodb sestojev. Zbornik gozdarstva in lesarstva 56: 55–71.

Košir, B., 2008: Modelling Stand Damages and Comparison of Two Harvesting Methods. Croat. j. for. eng. 29(1): 5–14.

Kotar, M., 2006: Kakovost debel v prebiralnih in enomernih gozdovih jelke in smreke. Gozdarski vestnik 64(9): 409–427.

Kotar, M., 2005: Zgradba, rast in donos gozda na ekoloških in fizioloških osnovah. Zveza gozdarskih društev Slovenije in Zavod za gozdove Slovenije, p. 354–362.

Kralj, T., 2008: Primerjava sistemov za razvrščanje tal na izbranih tleh v Sloveniji. Doktorska disertacija, Univerza v Ljubljani, Biotehniška fakulteta, 151 p.

Matić, S., Anić, I., Oršanić, H., 2003: Uzgojni postupci u bukovim šumama. In: Obična bukva u Hrvatskoj (Matić, S., Prpić, B., Gračan, J., Anić, I., Dundović, J., Eds.), Akademija šumarskih znanosti, Zagreb, p. 340–369.

Croat. j. for. eng. 34(2013)1

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

A. Kadunc

Mattila, U., Nuutinen, T., 2007: Assessing the Incidence of Butt Rot in Norway spruce in Southern Finland. Silva Fennica 41(1): 29–43.

Prka, M., 2003: Vrijednosne značajke bukovih stabala prema vrsti sijeka u sječinama Bjelovarske Bilogore. Šumarski list 127(1–2): 35−44.

Möykkynen, T., Miina, J., 2002: Optimizing the Management of a Butt-rotted Picea abies Stand Infected by Heterobasidion annosum fro the Previous Rotation. Scandavian Journal of Forest Research 17(1): 47–52.

Prka, M., 2006: Značajke doznačenih bukovih stabala po vrsti sijeka u sječinama Bjelovarske Bilogore i utjecaj na strukturu sortimenata. Šumarski list 130(7–8): 319−329.

Möykkynen, T., Miina, J., Pukkala, T., 2000: Optimizing the management of a Picea abies stand under risk of butt rot.

Forest Pathology 30(2): 65–76.

Möykkynen, T., Miina, J., Pukkala, T., von Weissenberg, K., 1998: Modelling the spread of butt rot in a Picea abies stand in Finland to evaluate the profitability of stump protection against Heterobasidion annosum. Forest Ecology and Management 106(2–3): 247–257.

Nilsen, P., 1983: The occurrence of rot in old stands of Norway spruce (Picea abies (L.) Karst.), related to tree data and stand properties. Rapp. Norsk Inst. Skogforsk 83(2): 1–34.

Oliva, J., Thor, M., Stenlid, J., 2010: Reaction zone and periodic increment decrease in Picea abies trees infected by Heterobasidion annosum s. l.. Forest Ecology and Management 260(5): 692–698.

Oliva, J., Romeralo, C., Stenlid, J., 2011: Accuracy of the Rotfinder instrument in detecting decay on Norway spruce (Picea abies) trees. Forest Ecology and Management 262(8): 1378–1386.

Pawsey, R. G., Stankovicova, L., 1974: Studies of extraction damage decay in crops of Picea abies in southern England. II. The dveleopment of Stereum sanguinoletum following experimental wounding and inoculation. European Journal of Forest Pathology 4(4): 203–214.

Perrin, R., Delatour, C., 1976: Estimating the height of decay in standing Norway spruce attacked by Fomes annosus. European Journal of Forest Pathology 6(4): 193–203.

Piri, T., Korhonen, K., Sairanen, A., 1990: Occurrence of Heterobasidion annosum in pure and mixed stands in southern Finland. Scand. J. For. Res. 5(1–4): 113–125.

Piškur, M., 2001: Vpliv mehanskih poškodb drevja na vrednost in strukturo gozdnih lesnih sortimentov. Gozdarski vestnik 59(3): 128–138.

Pranjić, A., Lukić, N., 1997: Izmjera šuma. Udžbenici Sveučilišta u Zagrebu, Šumarski fakultet, p. 142−150.

Croat. j. for. eng. 34(2013)1

Rönnberg, J., Jørgensen, B. B., 2000: Incidence of Root and Butt Rot in Consecutive Rotations of Picea abies. Scandinavian Journal of Forest Research 15(2): 210−217.

Rötheli, E., Heiri, C., Bigler, C., 2011: Effects of growth rates, tree morphology and site conditions on longevity of Norway spruce in the northern Swiss Alps. European Journal of Forest Research 131: 117–1125.

Seifert, T., 2007: Simulating the extent of decay caused by Heterobasidion annosum s. l. in stems of Norway spruce. Forest Ecology and Management 248(1–2): 95−106.

Stenlid, J., Redfern, D. B., 1998: Spread within the Tree and Stand. In: Heterobasidion annosum − Biology, Ecology, Impact and Control (Woodward, S., Stenlid, J., Karjalainen, R., Hüttermann, A., Eds.), CAB International, Wallingford, UK, p.

125−141.

Stenlid, J., Wästerlund, I., 1986: Estimating the frequency of stem rot in Picea abies using an icrement borer. Scandinavian Journal of Forest Research 1(1–4): 303–308.

Tamminen, P., 1985: Butt-rot in Norway spruce in southern Finland. Communicationes Instituti Forestalis Fenniae 127: 1−52 Thor, M., Ståhl, G., Stenlid, J., 2005: Modelling root rot incidence in Sweden using tree, site and stand variable. Scandinavian Journal of Forest Research 20(2): 165–176.

Van Laar, A., Akça, A., 2007: Forest Mensuration. Springer, Verlag, Netherlands, 383 p.

Vasiliauskas, R., 2001: Damage to trees due to forestry operations and its pathological significance in temperate forests: a literature review. Forestry 74(4): 319−336Woodward, S., Stenlid, J., Karjalainen, R., Hüttermann, A., 1998: Heterobasidion annosum − Biology, Ecology, Impact and Control.

CAB International, Wallingford, UK, 589 p.

Zycha, H., Dimitri, L., Kliefoth, R., 1970: Ergebnisse objektiver Messungen der durch Fomes annosus verursachten Rotfäule in Fichtenbeständen. Allgemeine Forst und Jagdzeitung 141: 66–73.

147

A. Kadunc

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

Appendix: Regression parameters – Dodatak: parametri regresije

y = udjel natrulih stabala%); x = prsni promjer stabla (cm)

y = udjel stabala zaraženih s truleži na donjem dijelu debla (%); x = prsni promjer stabla (cm) y = razlika u vrijednosti drva (€/m3); x = Prsni promjer stabla (cm)

1 2

y = share of trees with lower-trunk rot (%); x = DBH (cm) y = difference in value (€/m3); x = DBH (cm)

3

148

y = the share of rot-infected trees (%); x = DBH (cm)

Figure – Slika

Curve form – Oblik krivulje

R2

P

y = 19.067+1.551x

0.098

0.000

y = 0.050+2.400x

0.257

0.000

Not significant – Nije značajno

0.001

0.273

Not significant – Nije značajno

0.000

0.868

y = 28.781+1.393x

0.071

0.000

y = 13.919+1.785x

0.156

0.000

y = 27.232+0.340x

0.009

0.045

y = 20.686+2.355x

0.137

0.000

y = –1.544+3.533x

0.431

0.000

y = 6.331+1.718x

0.321

0.000

0.044

0.199

y = –14.390+3.881x

0.522

0.000

y = (4.820+1.924x-0.014x2) – (5.282+1.608x0.012x2)

0.484

0.000

0.305

0.000

y = (-4.816+2.447x-0.020x2) – (2.760+1.777x0.015x2)

0.759

0.000

0.498

0.000

y = (-11.762+2.966x-0.025x2) – (9.851+1.574x0.013x2)

0.538

0.000

0.258

0.000

y = (-11.675+2.746x-0.021x2) – (1.200+1.749x0.013x2)

0.396

0.000

0.208

0.006

Uneven-aged stand, dolomite

Not significant, not enough data

0.044

0.665

Raznodobna sastojina na dolomitu

Nije značajno, premalo podataka

0.236

0.133

Uneven-aged stand, silicate

Not significant, not enough data for trees with rot

0.596

0.000

Raznodobna sastojina na silikatu

Nije značajno, premalo podataka o stablima s truleži

0.035

0.765

Stratum – Izvor podataka Even-aged stand, lower-trunk ro Jednodobna sastojina, trulež na donjem dijelu debla Uneven-aged stand, lower-trunk rot Raznodobna sastojina, trulež na donjem dijelu debla Even-aged stand, upper-trunk rot Jednodobna sastojina, trulež na gornjem dijelu debla Uneven-aged stand, upper-trunk rot Raznodobna sastojina, trulež na gornjem dijelu debla Even-aged stand, rot-together Jednodobna sastojina, trulež na oba dijela debla Uneven-aged stand, rot-together Raznodobna sastojina, trulež na oba dijela debla Even-aged stand, limestone Jednodobna sastojina na vapnencu Even-aged stand, dolomite Jednodobna sastojina na dolomitu Even-aged stand, silicate Jednodobna sastojina na silikatu Uneven-aged stand, limestone Raznodobna sastojina na vapnencu Uneven-aged stand, dolomite

Not significant, not enough data

Raznodobna sastojina na dolomitu

Nije značajno, premalo podataka

Uneven-aged stand, silicate Raznodobna sastojina na silikatu Even-aged stand, limestone Jednodobna sastojina na vapnencu Even-aged stand, dolomite Jednodobna sastojina na dolomitu Even-aged stand, silicate Jednodobna sastojina na silikatu Uneven-aged stand, limestone Raznodobna sastojina na vapnencu

Croat. j. for. eng. 34(2013)1

The Incidence of Rot in Norway Spruce and its Influence on the Value of Trees in Slovenia (137–149)

A. Kadunc

Sažetak

Pojava truleži kod obične smreke i njezin utjecaj na vrijednost stabala u Sloveniji U sjevernoj i srednoj Europi obična smreka (Picea abies /L./ Karst.) slovi za vodeću vrstu po gospodarskom značenju (Kenk i Guehne 2001; Jöbstl 2011). Velik je problem pri gospodarenju smrekom pojava truleži (Kohnle i Kändler 2007).

Stoga je cilj istraživanja bio: 1) uočiti frekvenciju pojavljivanja truleži na smrekovim stablima s obzirom na prsni promjer stabla, vrstu sastojine i sastojinske prilike, 2) prikazati utjecaj truleži na gubitak vrijednosti drva i 3) ustanoviti koji parametri utječu na pojavu truleži.

Istraživana su smrekova stabla na 65 mjesta u Republici Sloveniji, odnosno na 20 različitih biljnih asocijacija.

Posječena su 1334 smrekova stabla, od toga 1015 u jednodobnim i 319 u raznodobnim sastojinama.

Prije sječe svakomu je stablu izmjeren prsni promjer i zabilježena pojava više vrhova stabala, mehaničkih oštećenja stabla odnosno žilišta te odlomljenoga vrha stabla. Svaki je dio debla razvrstan u razred kakvoće s obzirom na norme prema JUS-u (D.B4.029 1979) koje se i dalje primjenjuju u Sloveniji. Daljnja je analiza značajki oborenih stabala omogućila izračunavanje prosječne debljine godova i mjere vitkosti (visina stabla/prsni promjer stabla). Na svakom je prerezu stabla zabilježena prisutnost truleži. Trulež je provjeravana na svim poprečnim presjecima te je ustanovljena prisutnost truleži na donjem dijelu debla (do visine 5 m od tla) i na gornjem dijelu debla (na visini većoj od 5 m od tla).

Analiza pokazuje da je pojava truleži u donjem i gornjem dijelu debla relativno češća u jednodobnim sastojinama.

Između frekvencije pojavljivanja truleži u donjem i gornjem dijelu debla razlike su velike. Trulež se javila u donjem dijelu debla 34 % (jednodobne sastojine) odnosno 26 % (raznodobne sastojine) stabala, a trulež u gornjem dijelu debla samo kod 14 % (jednodobne sastojine) odnosno 11 % (raznodobne sastojine) stabala. S povećavanjem prsnoga promjera udio se truleži u gornjem dijelu debla nije povećao. Trulež je u donjem dijelu debla veća ako se povećava prsni promjer te je češća u jednodobnim sastojinama.

Razlike su između matičnih podloga pojedinih sastojina osjetne. Već kod tanjih stabala udio truleži u donjem dijelu debla veći je na vapnencu i dolomitu u jednodobnim sastojinama. Gubitak vrijednosti drva zbog pojave truleži najveći je na silikatnoj podlozi u jednodobnim sastojinama i na vapnencu u raznodobnim sastojinama. Najniži je gubitak vrijednosti drva u jednodobnim sastojinama na vapnencu. Najveći je novčani gubitak zbog pojave truleži kod stabala s prsnim promjerom od 50 do 70 cm (15 – 19 €/m3).

Za trulež u donjem i gornjem dijelu debla uz pomoć binarne logičke regresije provjereno je koje varijable utječu na spomenute pojave. Pojava truleži u donjem dijelu debla vjerojatnija je na produktivnijim staništima, na dolomitnoj matičnoj podlozi, na višoj nadmorskoj visini, kod starijega i ozlijeđenoga drveća. Vitkost stabla pridonosi manjoj vjerojatnosti pojave truleži u donjem dijelu debla. Također se vidi da je pojava više vjerojatna u jednodobnim sastojinama, a manje na južnim ekspozicijama. Pojava truleži u gornjem dijelu debla vjerojatnija je na produktivnijim staništima, na dolomitu i silikatu i kod ozlijeđenih stabala, na stablima sa slomljenim vrhom te stablima s više vrhova.

Istraživanje je pokazalo da je teško objasniti pojavu truleži zato što mnogi čimbenici utječu na njezin nastanak.

Rezultate bi istraživanja trebalo dopuniti podacima o tipovima truleži, stupnju natrulosti i obujmu pojave. Trulež kod smreke ostaje jedan od većih problema gospodarenja tom vrstom.

Ključne riječi: trulež drva, obična smreka, Slovenija

Authors’ address – Autorova adresa:

Received (Primljeno): April 11, 2012 Accepted (Prihvaćeno): October 16, 2012 Croat. j. for. eng. 34(2013)1

Asst. Prof. Aleš Kadunc, PhD. e-mail: ales.kadunc@bf.uni-lj.si University of Ljubljana, Biotechnical Faculty Department of Forestry and Renewable Forest Resources Večna pot 83 1000 Ljubljana SLOVENIA

149

Original scietific paper – Izvorni znanstveni rad

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees in the Karavanke Region Aleš Poljanec, Aleš Kadunc Abstract – Nacrtak The paper studies the quality of beech trees in the Karavanke mountain range. The data obtained for this study came from the permanent sample plots (here in after PSPs) of the studied area. The timber assortment structure of mature beech trees in stands was calculated on the basis of the estimated quality of a total of 7,154 beech trees from 2,088 plots and a small sample of harvested trees. The established timber assortment structure and the prices of timber ex forest road were then used to determine the value of beech wood. The study showed that the assortment structure was extremely unfavorable, with the share of trees with sliced and peeled veneer quality lower than 1%, and the share of sawlogs totaling 1.6% of the net volume of analyzed trees. The quality is highest in beech trees measuring 50–55 cm in diameter at breast height. Quality values were higher in rejuvenation stands, and in stands with a comparatively high share of fir or a low share of conifers (spruce). The value drops with altitude and inclination and is higher on slope sites. Undamaged trees from higher social layers are also more valuable. Furthermore, beech timber value is positively affected by high harvesting intensity. In dense stands with a basal area exceeding 60–65 m2/ha, the value begins to decline. In order to improve the quality structure of beech stands, beech would have to be grown in beech-dominated clusters or stands, and thinned at the correct time. Particular attention needs to be paid to minimize the damage to dominant trees during harvesting. Quality assessment carried out within the forest inventory enables to assess the stand quality and value potential of forests at different spatial scale. The research has highlighted several possibilities to use quality data in connection with other parameters to enhance the efficiency of forest management.

Keywords: European beech, stem quality, assortment structure, timber value, influential factors, the Karavanke range

1. Introduction – Uvod European beech (Fagus sylvatica L.) is one of the most common tree species in Europe (Ellenberg 1986). In Slovenia, beech is present on almost 89% of total forest area. With the share of growing stock greater than 50%, it occupies merely 25% of total forest area (Ficko et al.

2008). In the Alpine territory, beech has been generally recognized as an economically less valuable species since the beginning of forest management. Beech wood was used as firewood and as raw material in charcoalmaking (Veber 1986). There were also several attempts to replace beech with more profitable conifers, in particular spruce (Johann 2006). In Slovenia, the distribuCroat. j. for. eng. 34(2013)1

tion range of beech has been declining for centuries; however, in the recent decades its distribution range and abundance increased (Poljanec et al. 2010), predominantly as a result of changed forest management concepts (Mlinšek 1968) and natural disturbances (e.g. windthrow and insect attacks) resulting in a decreasing proportion of conifers (mainly spruce).

With increasing abundance of beech and the changing role of beech timber, however, the beech management measures in the Alpine territory should be determined in detail. Such an investigation requires various types of information, including the data on the quality structure of forest stands, wood value characteristics of tree species, and the factors influencing wood value.

151

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Such information is required to define forest management goals and silvicultural objectives, determine the harvest maturity of trees and stands, and decide on the intensity and frequency of various forest management measures. Moreover, it is important to consider these data in assessing the value of forest stands, in optimizing yields and in determining the expected economic effects (Rebula and Kotar 2003).

The quality and value of beech trees have been the subject of several studies in Slovenia and abroad (e.g.

Petráš and Nociar 1991; Rieder 1997; Gfeller 1998; Krpan 2003; Kadunc 2006; Prka 2010). Most of these focused on the average quality (Petráš and Nociar 1991) and value (Omahen 1998; Prka 2003a) of trees in certain harvesting sites and on the factors influencing beech quality and value (e.g. Rebula and Kotar 2003). Numerous studies investigated the occurrence of heartwood discoloration in beech, the factors behind this defect (Knoke 2003; Prka 2003b; Kadunc 2006), and the impact of heartwood discoloration on the value of harvested trees (e.g. Šmajdek 2001). The research into the quality and value of standing trees and beech stands (e.g. Rebula 2005) has been less frequent. Normally, such studies are based on timber quality assessments of standing trees from various research plots (Pirc 1997) or on the assessments of the quality of trees measured during forest inventories (Rebula 2005; Čavlović 2010, Rebula 2011). The first approach, based on the stem analysis of a small sample of harvested trees, is destructive; however, it gives very accurate results, and enables an appropriate determination of internal stem defects. The second approach, which is based on the assessment of

standing trees during a forest inventory, is non-destructive and, normally, representative for larger areas. In this paper we combine the strengths of both presented methods in order to improve the quality assessments at PSPs. Therefore the quality assessments from PSPs were improved with a detailed analysis of the quality obtained from a small sample of harvested trees in comparable sites (Kadunc and Poljanec 2011).

Apart from the methodology developed within the framework of this research to estimate the timber assortments of standing trees from PSPs, the primary aim of the study was: Þ to analyze the quality and value-based structure of beech trees and stands, Þ to develop models to examine the mutual influence of tree, stand, site, and forest management variables on the quality and value of standing mature beech trees in the Karavanke region.

The Karavanke mountain range was chosen due to its geological diversity (Buser 1991) and because the Kara vanke mountains are high enough to cover the complete range of altitude and soil gradients in which beech constitutes an important or even a dominating share.

2. Materials and methods – Materijal i metode 2.1 Study area – Područje istraživanja The Karavanke is a mountain range which stretches west-east for 120 km from Tarvisio, Italy to Slovenj

Fig. 1 Location of the study area and distribution of PSPs Slika 1. Područje istraživanja i distribucija trajnih pokusnih ploha (TPP)

152

Croat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Gradec, Slovenia (Fig. 1). The study area is characterized by a considerable diversity of relief and geological composition. The altitude of the forest area ranges from 400 m to 1,900 m above sea level. The area belongs to the alpine continental climate zone, but local climatic conditions can change rapidly as a result of varied terrain. The temperature extremes vary according to the relief, with the lowest temperatures between –24°C and –21°C, and the highest between 25°C and 32°C (ARSO 2004a). Precipitation decreases from west to east. The western part of the Karavanke receives up to 3,600 mm in precipitation, whereas the lower-lying areas of the eastern part receive only 1,300 mm (ARSO 2004b). Karstic terrain and clastic sedimentary rock cover most of the territory, while the presence of felsic, igneous and various metamorphic rocks is sporadic.

Forests cover 82% of the total study area (Fig. 1).

They are characterized by a small-scale system management, using irregular shelterwood and group systems. The mean growing stock amounts to 322 m3/ha.

There are 38 different tree species recorded in the study area; Norway spruce (Picea abies Karst.) (66%) and European beech (16%) account for the highest proportion of growing stock, followed by European larch (Larix decidua Mill.) with 6% (SFS 2010b). The zonality of forest vegetation in the Karavanke is quite clearly defined due to distinctive orographic factors, different soil substrata and climatic conditions. According to the terminology used in forestry practice (Kutnar et al.

2012), the forests with beech in the study area could be classified into 12 forest types: Adenostylo-Fagetum, Anemone-Fagetum, Arunco-Fagetum, Castaneo-Fagetum, Luzulo-Fagetum, Ostryo-Fagetum, Hacquetio-Fagetum, Homogyno-Fagetum, Aceri-Fraxinetum, Aposeri-Piceetum, Bazzanio-Abietetum and Pinetum subillyricum.

2.2 Data collection and statistical analysis Prikupljanje podataka i statistička analiza Two datasets were used to analyze the quality and value of beech trees in the Karavanke region. The first dataset was composed of a timber quality assessment of standing trees on PSPs, which is maintained by the Slovenia Forest Service (SFS 2010a). PSP inventory covers the entire area of Slovenia and is carried out by overlaying a systematic grid of PSPs, which are mainly distributed on a 250×500 m grid. The re-measurement interval is 10 years and a scheduled number of plots are surveyed every year. The size of an inventory plot is 500 m2 and consists of two concentric circles with radii of 7.98 m and 12.61 m, to measure small (10 ≥ DBH < 30 cm) and large trees (DBH ≥ 30 cm), respectively. The basic records for each plot include latitude, longitude, elevation, slope, aspect and topoCroat. j. for. eng. 34(2013)1

A. Poljanec and A. Kadunc

graphic position. On each plot, every sample tree is georeferenced according to its polar coordinates (i.e. distance and azimuth from the plot centre) and the following data are recorded in every measurement: tree species, DBH, social status, major stem and crown damages, identification code (standing tree, standing dead tree, thinned tree, recruited tree) and timber quality. Timber quality is limited to larger trees (DBH ≥ 30 cm), where every tree is categorized in one of five quality classes (Table 1). For the purpose of this study 7,154 large beech trees on 2,088 PSPs in the Karavanke region were selected from the entire original dataset (Fig. 1).

The second dataset contained 26 plots (900 m2 each), established in mature, fully stocked stands dominated by beech on 8 locations in the wider Alpine region. All trees within the plot with DBH equal or greater than 10 cm were measured and cut (a total of 1,057 trees; 495 trees with DBH ≥ 30 cm were included in the research). For each tree species, social status, DBH, height, crown length and volume were recorded (Kotar 1989, 1991). All trees were cut into 7–10 sections for the purpose of conducting a stem analysis. The quality of trees was assessed during harvesting: log parts were classified into assortment classes according to JUS standards (1979), and internal trunk defects (e.g. heartwood discoloration) were also considered.

The size of this sample was small and therefore only used as an auxiliary sample to convert the quality assessed on PSPs into assortment classes (Table 1).

Firstly, the gross volumes of standing trees measured at PSPs were converted into net volumes on the basis of previous studies (Čokl 1981; Rebula 2002). The established conversion factor was between 0.85 and 0.98 and was DBH dependent. Furthermore, on the basis of DBH analyses and recommendations by Kotar (2003) and Kadunc (2006), the trees were assigned volume shares by individual segments. For Čokl’s (intermediary) tariffs up to the tariff class 4 and for Schaeffer’s tariffs up to the tariff class 4/5 trees were assumed to be conical, which meant that the butt log took up 43.54% of the stem volume, the second segment made up for 29.89%, the third for 18.09%, and the fourth segment took up 8.48% of the stem volume. For other tariff classes a more cylindrical shape was assumed, with the butt log taking up 37.50% of the stem volume, the second segment 29.17%, the third segment 20.83%, and the fourth segment 12.50% of the stem volume.

The quality classes of beech trees from PSPs (SFS 2010a) were further transformed into six assortment classes (sliced veneer, peeled veneer, sawlog I, sawlog II, sawlog III including sleeper, firewood). As some quality classes contained more than one assortment class for certain segments (e.g the first segment of trees

153

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Table 1 Conversion of quality classes evaluated on PSPs into assortment classes according to JUS standards (1979) Tablica 1. Konverzija razvrstavanja kakvoće ocijenjene na TPP u klase prema JUS-u (1979) Quality class Klasa kakvoće

Segment

DBH, cm

Prsni Segment promjer, cm 30–37 38–43 44–55 1.

56–60 61–65

Excellent (1 segment is veneer, peeled veneer, or sawlog I, 2nd segment is at least sawlog II)

> 65

Odlična

30–34

(prvi je segment furnirski trupac, trupac za ljuštenje ili pilanski trupac I. klase, drugi je segment najmanje pilanski trupac II. klase)

35–39

st

40–54 2.

55–59

Assortment class Klasa sortimenata Sawlog I, 100% – Pilanski trupci I. klase, 100 % Sawlog I 60%, peeled veneer 40% Pilanski trupci I. klase 60 %, trupci za ljuštenje 40 % Sawlog I 55%, peeled veneer 35%, sliced veneer 10% Pilanski trupci I. klase 55 %, trupci za ljuštenje 35 %, furnirski trupci 10 % Sawlog I 45%, peeled veneer 55% Pilanski trupci I. klase 45 %, trupci za ljuštenje 55 % Sawlog I 30%, peeled veneer 70% – Pilanski trupci I. klase 30 %, trupci za ljuštenje 70 % Sawlog I 35%, peeled veneer 65% Pilanski trupci I. klase 35 %, trupci za ljuštenje 65 % Sawlog II 100% – Pilanski trupci II. klase 100 % Sawlog II 96%, sawlog I 4% Pilanski trupci II. klase 96 %, pilanski trupci I. klase 4 % Sawlog II 70%, sawlog I 30% Pilanski trupci II. klase 70 %, pilanski trupci I. klase 30 % Sawlog II 55%, sawlog I 45% Pilanski trupci II. klase 55 %, pilanski trupci I. klase 45 % Sawlog II 40%, sawlog I 50%, peeled veneer 10%

> 60

Pilanski trupci II. klase 40 %, pilanski trupci I. klase 50 %, trupci za ljuštenje 10 %

3.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

4.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

Very good

1.

≥ 30

Sawlog II 100% – Pilanski trupci II. klase 100 %

(1st and 2nd segment are sawlog II, or 1st segment is of higher quality, and 2nd segment is of slightly lower quality)

2.

≥ 30

Sawlog II 100% – Pilanski trupci II. klase 100 %

3.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

(prvi i drugi segment su pilanski trupci II. klase ili je prvi segment bolje kakvoće, a drugi malo slabije kakvoće)

4.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

Good

1.

≥ 30

Sawlog II 100% – Pilanski trupci II. klase 100 %

(1st segment is sawlog II, 2nd segment is sawlog III or sleeper)

2.

≥ 30

3.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

4.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

Vrlo dobra

Dobra (prvi segment pilanski trupac II. klase, drugi segment je pilanski trupac III. klase ili trupac za željezničke pragove)

154

Sawlog III or sleeper 100% Pilanski trupci III. klase ili trupci za željezničke pragove 100 %

Croat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Satisfying (1st and 2nd segment are sawlogs III or sleepers, or 1st segment is of higher quality, and 2nd segment is of slightly lower quality)

Sawlog III or sleeper 100%

1.

≥ 30

2.

≥ 30

3.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

4.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

Zadovoljavajuća (prvi i drugi segment su pilanski trupci III. klase ili trupci za željezničke pragovove, ili je prvi segment bolje kakvoće, dok je drugi segment malo slabije kakvoće)

A. Poljanec and A. Kadunc

Pilanski trupci III. klase ili trupci za željezničke pragove 100 % Sawlog III or sleeper 100% Pilanski trupci III. klase ili trupci za željezničke pragove 100 %

Sawlog III or sleeper 40%, firewood 60% 1.

≥ 30

Pilanski trupci III. klase ili trupci za željezničke pragove 40 %, drvo za ogrjev 60 %

2.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

3.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

4.

≥ 30

Firewood 100% – Drvo za ogrjev 100 %

Bad st

(1 segment is sawlog III, sleeper, or lower, 2nd segment is industrial wood or firewood) Loša (prvi segment je pilanski trupac III. klase, trupac za željezničke pragove ili slabije, drugi segment je industrijsko drvo ili drvo za ogrjev)

classified as excellent, could either be veneer, peeled veneer or sawlog I quality; see also Table 1), the conversion of quality classes into assortment structure was based on the ratios assessed from the second (harvested) dataset. For the purpose of conversion, each tree from the PSP was assigned the average quality value, although in reality this phenomenon is frequency-based. For example, a tree of excellent quality and a 50 cm DBH was assigned 10% of sliced veneer, 35% of peeled veneer and 55% of sawlog I although veneer quality is only achieved in 10%, peeled veneer quality in 35%, and sawlog I in 55% of trees classified as excellent.

The calculations of wood value ex forest road were based on the price lists for 2011 published by 10 major beech traders in Slovenia. In order to obtain the value of the net volume of trees ex forest road (in €/m3), the volume of each quality class was multiplied by the average price and then the sum was divided by the net volume of the tree. The influence of the actual harvesting, bucking and sorting of the logs on final tree value was not considered in the calculations. Furthermore, the value was not reduced by the costs of silvicultural measures, harvesting operations, construction and maintenance of forest roads, overhead costs, public forest service, and various taxes and duties (Table 2).

Croat. j. for. eng. 34(2013)1

Table 2 Average prices of beech timber assortments ex forest road, €/m3 Tablica 2. Prosječne cijene drvnih sortimenata bukve fco. šumska cesta, €/m3 Yearly average price ex Assortment class

forest road, €/m3

Klasa sortimenata

Godišnja prosječna cijena fco. šumska cesta, €/m3

Veneer logs Furnirski trupci Peeled veneer Trupci za ljuštenje Sawlogs I Pilanski trupci I. klase Sawlogs II Pilanski trupci II. klase

192.83 100.04 71.85 56.19

Sawlogs III, sleepers Pilanski trupci III. klase, trupci za željezničke pragove Firewood Drvo za ogrjev

42.28

40.44

155

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Table 3 Independent variables tested in the binary logistic and multivariate linear regression model Tablica 3. Nezavisne varijable testirane u binarnom logističkom i multivarijantnom linearnom regresijskom modelu Level

Variable

Razina

Varijabla

Tree-level variables Varijable na razini stabala

Variable type and dependency form Tip varijable i oblik ovisnosti

DBH, cm

Continuous/parabolic

Prsni promjer, cm

Kontinuirana/parabolična

Basal area increment, cm2/year

Continuous/parabolic

Prirast temeljnice, cm2/god.

Kontinuirana/parabolična

Dummy variables coding Kodiranje »dummy« varijabli Name

Transformation

Naziv

Transformacija

–

–

–

– 1 = trees in upper stand layer;

Social status Socijalni status

0/1

Upper layer

0 = trees in other layers

Gornji sloj

1 = stabla u gornjem sloju sastojine; 0 = stabla u ostalim slojevima 1 = presence of major damages;

Damage presence Prisutnost oštećenja

0/1

Damage presence

0 = no damages

Prisutnost oštećenja

1 = prisutnost većih oštećenja; 0 = bez oštećenja 1 = rejuvenation stand;

Stand type Tip sastojine

Rejuvenation stand

0 = other

Pomlađene sastojine

1 = pomlađene sastojine; 0 = ostalo

0/1 Uneven-aged stand

Stand-level variables Varijable na razini sastojine

Stand basal area, m2/ha

Continuous/parabolic

Temeljnica, m2/ha

Kontinuirana/parabolična

Spruce share in stand basal area, %

Continuous/linear

Udio smreke po temeljnici, %

Kontinuirana/linearna

Fir share in stand basal area, %

Continuous/linear

Udio jele po temeljnici, %

Kontinuirana/linearna

Share of coniferous tree species in stand basal area, %

Continuous/linear

Udio četinjača po temeljnici, %

Kontinuirana/linearna

Altitude, m

Continuous/parabolic

Nadmorska visina, m

Kontinuirana/parabolična

Inclination, °

Continuous/linear

Site-level variables

Nagib, °

Kontinuirana/linearna

Varijable na razini staništa

Landscape position Položaj u krajoliku Aspect Reljef

156

0/1

0/1

1 = uneven-aged stand, two-layer stand; 0 = other

Preborne šume

1 = preborne šume, dvoslojne sastojine; 0 = ostalo

–

–

–

–

–

–

–

–

–

–

–

–

Ridge position

1 = ridge; 0 = other

Greben

1 = greben; 0 = ostalo

Slope position

1 = slope; 0 = other

Padine

1 = padina; 0 = ostalo

Sun-exposed site

1 = SE/JI, S/J, SW/JZ, 0 = other

Prisojna strana

1 = SE/JI, S/J, SW/JZ, 0 = ostalo

Croat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Forest management variable Varijabla gospodarenja šumama

Harvesting intensity, m2/ha, year

Continuous/linear

Intenzitet sječe, m2/ha, godišnje

Kontinuirana/linearna

Timber value as well as the quality of beech trees was assumed to depend on tree traits and stand characteristics, site conditions, and forest management variables. All independent variables were acquired from the PSP inventory (Table 3). The influence of independent variables on the quality and timber value was tested using bivariate and multivariate statistical analyses. The Kruskal-Wallis (KW) test was used to assess the existence of differences between the sliced veneer and peeled veneer shares by forest types for the trees classified into diameter class 45–49 cm or 50–54 cm, which normally achieves the highest value. Differences in the average value of beech wood between forest types were tested using univariate analysis of variance, while differences across diameter classes were tested using a median test. Since the relationship between the quality or timber value and the independent variable is often parabolic (e.g. Knoke 2003; Rebula 2005), a quadratic term was added in the models for certain continuous variables when so indicated by the bivariate analysis and data survey. All categorical variables with n categories were transformed into n –1 dummy variables (Table 3).

Given the binary response of the dependent variable (1 if quality was sliced veneer or peeled veneer; 0 for other assortment classes), the probability of the presence of high quality beech timber was modeled using a binary logistic regression. In order to avoid multicollinearity, only variables with a tolerance factor higher than 0.2 were included in the model (Knoke 2003). The model was estimated using a backward stepwise procedure where removal testing was based on the probability of the likelihood-ratio statistic based on the maximum partial likelihood estimates (Kleinbaum and Klein 2002). The Nagelkerke pseudo-R2 was used to evaluate goodness-of-fit for the model.

The relation between timber value and independent variables was examined with the standard multivariate linear regression. The model was estimated using a stepwise procedure. Independent variables used in the modeling procedure were the same as those used in the binary logistic model (Table 3). As in the logistic model, the variables were tested for multicollinearity. The final model was derived by using the adjusted R2 as the criterion for goodness-of-fit. All statistical analysis was carried out in the PASW SPSS Statistics 18.

Croat. j. for. eng. 34(2013)1

–

A. Poljanec and A. Kadunc

–

3. Results – Rezultati 3.1 Quality and assortment structure of beech trees – Kakvoća i struktura sortimenata bukovih stabala The share of trees classified as excellent or very good amounted to 22.9%. The number of excellent beech trees was highest in diameter class 50–54 cm, whereas the highest number of very good quality beech trees was found in diameter class 55–59 cm. The share of good quality timber was highest in the lowestdiameter trees and decreased with thickness. Conversely, the share of satisfactory and poor quality timber was highest in the thickest trees and rose with diameter (Table 4; upper part). In the assortment structure, firewood was followed in proportion by sawlog II and sawlog III. Sliced veneer, peeled veneer, and sawlog I amounted to a 2.3% (Table 4, lower part). The proportion of veneer quality wood was highest in trees measuring approximately 50 cm in DBH, whereas the share of peeled veneer was highest in very thick trees (> 65 cm).

The differences between the forest types were statistically significant for sliced veneer shares (KW = 24,527; P = 0.011) as well as for peeled veneer shares (KW = 23,854; P = 0.013). Forest types with the highest total shares of sliced and peeled veneer (above 2%) included Aceri-Fraxinetum, Aposeri-Piceetum and Hacquetio-Fagetum. In the forest types Castaneo-Fagetum, Bazzanio-Abietetum and Adenostylo-Fagetum sliced/ peeled veneer quality logs were not recorded (Fig. 2).

Similar relations between forest types as determined for sliced and peeled veneer shares were also ascertained in the timber class of sawlog I. The variability in the quality of beech trees can be partly explained by selected tree, stand, site, and forest management variables. In the binary logistic regression model, the main effects of eight independent variables significantly influenced the beech quality (Table 5, left part). The model explained 22% of the total variability (Nagelkerke R2 = 0.220).

In the group of variables denoting tree characteristics, the model included DBH, social status and presence of any major damage to the tree. There was a parabolic relationship between diameter at breast

157

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Table 4 The shares (in %) of beech trees by diameter classes with regard to the quality assessment and assortment structure (modes are marked bold) Tablica 4. Udio (%) bukovih stabala po debljinskim razredima s obzirom na procjenu kakvoće i strukturu sortimenata (modovi su označeni podebljano) Diameter class, cm – Debljinski stupanj, cm

Quality class (tree share), % Klasa kakvoće (udio u stablu), %

Average

30–34

35–39

40–44

45–49

50–54

55–59

60–64

65+

Prosjek

Excellent – Odlična

2.8

3.5

5.4

6.2

8.1

5.0

1.5

6.0

4.2

Very good – Vrlo dobra

16.9

20.0

19.3

20.1

18.8

22.3

18.2

14.9

18.7

Good – Dobra

50.1

47.5

45.8

43.3

39.4

28.7

33.3

32.8

46.2

Satisfying – Zadovoljavajuća

22.2

22.0

22.8

24.4

22.4

27.2

36.4

28.4

23.0

Bad – Loša

8.0

7.0

6.5

6.0

11.2

16.8

10.6

17.9

7.9

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Total – Ukupno

Diameter class, cm – Debljinski stupanj, cm

Assortment class (volume share), % Klasa sortimenata (udio u obujmu), %

Average

30–34

35–39

40–44

45–49

50–54

55–59

60–64

65+

Prosjek

Veneer logs – Furnirski trupci

0.0

0.0

0.1

0.3

0.3

0.1

0.0

0.0

0.1

Peeled veneer – Trupci za ljuštenje

0.0

0.2

0.8

0.9

1.2

1.0

0.4

1.8

0.6

Sawlogs I – Pilanski trupci I. klase

1.1

1.2

1.8

2.0

2.6

1.6

0.4

1.8

1.6

Sawlogs II – Pilanski trupci II. klase

32.0

33.0

32.1

31.8

29.8

27.3

25.2

23.6

30.9

Sawlogs III – Pilanski trupci III. klase

31.0

30.2

30.0

30.2

28.3

29.2

36.3

30.8

30.4

Firewood – Drvo za ogrjev

35.9

35.4

35.3

34.9

37.8

40.9

37.6

41.9

36.5

Total – Ukupno

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Fig. 2 The shares of highest-quality beech assortment classes by forest type Slika 2. Udio bukovih sortimenata odlične kakvoće po tipovima šuma

158

Croat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

A. Poljanec and A. Kadunc

3.2 The value of beech trees – Vrijednost debla bukovih stabala

height and the probability of highest-quality beech assortments (sliced veneer and peeled veneer), with the probability rising with DBH up to 55 cm, and then dropping again. The presence of considerable structural damage in beech trees decreased the probability of sliced veneer and peeled veneer while the trees reaching the upper stand layer were more likely to develop the highest-quality beech timber assortments.

The probability of this phenomenon increased with the basal area and the share of fir, but decreased in stands with a higher share of conifers. Furthermore, the chances for sliced veneer and peeled veneer timber quality increased as a result of forest management activities, which were evident through harvesting intensity.

In the group of tested site variables, only elevation was included in the model. The probability for sliced veneer and peeled veneer occurrence increased up to 1,100 m above sea level, and decreased at higher elevations.

The average value of beech wood in the study region was 46.3 €/m3 and varied significantly between forest types (univariate analysis of variance, DBH was used as a covariate, F = 7.357, P = 0.000), as well as across diameter classes (Median test statistic = 228.93, P = 0.000).

In most forest types, beech value was highest in diameter class 45–50 cm and diameter class 50–55 cm, whereas in forest types Bazzanio-Abietetum and AdenostyloFagetum it culminateed in the thickest trees. In terms of beech timber value, the best results were obtained on high-productivity beech sites (e.g. Hacquetio-Fagetum and Homogyno sylvestris-Fagetum), and the timber values were lowest in secondary spruce and pine forests (e.g. Aposeri-Piceetum Pinetum subillyricum).

The influence of selected variables on the value of beech timber (in €/m3) was tested using the multivariate regression analysis. Despite testing the same set of

Table 5 Parameters of binary logistic regression (the outcome variable is the presence of sliced veneer or peeled veneer logs) and multivariate regression (using value in €/m3 as dependent variable) Tablica 5. Parametri binarne logističke regresije (izlaznu varijablu označuje prisutnost rezanoga ili ljuštenoga furnira) i multivarijantne regresije (koristeći vrijednost u €/m3 kao zavisnu varijablu) Predictor – Independent variable Pokazatelj – Neovisna varijabla

Binary logistic regression

Multivariate regression

Binarna logistička regresija

Multivarijantna regresija

β

Exp(β)

P

β

SE(β)

P

Constant – Konstanta

–30.252

0.000

0.000

43.5778

0.9903

0.000

DBH – Prsni promjer

0.644

1.904

0.000

0.0901

0.0365

0.014

DBH – Prsni promjer

–0.006

0.994

0.000

–0.0013

0.0004

0.001

Altitude – Nadmorska visina

0.018

1.018

0.002

–

–

–

Altitude2 – Nadmorska visina2

–0.000

1.000

0.002

–0.0000

0.0000

0.001

–

–

–

–0.0319

0.0071

0.000

0.015

1.015

0.022

0.1147

0.0204

0.000

–

–

–

–0.0007

0.0002

0.001

–

–

–

–0.0001

0.0000

0.003

Harvesting intensity – Intenzitet sječe

0.423

1.527

0.000

0.8936

0.1238

0.000

Share of fir – Udio jele

1.386

4.000

0.045

2.1142

0.5799

0.000

Share of conifers – Udio četinjača

–1.064

0.345

0.009

–3.2919

0.2538

0.000

Landscape position – Položaj u krajoliku

–

–

–

0.7932

0.2389

0.001

Rejuvenating stand – Pomlađene sastojine

–

–

–

0.5365

0.2013

0.008

Upper layer – Gornji sloj

1.092

2.981

0.000

1.8303

0.1736

0.000

Lower layer – Donji sloj

–

–

–

–3.0394

0.3659

0.000

–0.918

0.399

0.057

–2.1289

0.2763

0.000

2

2

Inclination – Nagib Stand basal area – Temeljnica 2

2

Stand basal area – Temeljnica 2

2

Basal area increment – Prirast temeljnice

Damage presence – Prisutnost oštećenja Croat. j. for. eng. 34(2013)1

159

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Fig. 3 The probability of highest-quality beech timber assortments and beech timber value (€/m3) in relation to DBH, stand basal area and harvesting Slika 3. Vjerojatnost bukovih sortimenata odlične kakvoće i vrijednosti bukovih stabala (€/m3) u ovisnosti o prsnom promjeru stabla, temeljnici sastojine i sječe

160

Croat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

variables as in binary logistic regression, only 9.2% of variance (R2 = 0.092) was explained. The lower performance of the regression model could be explained by a high proportion of low and medium quality timber (e.g firewood, sawlogs III) with similar timber prices, which makes the dependant variable (tree value) less distinctive.

The differences in the value of beech trees were clearly explained through the following tree characteristics: DBH and basal area increment (both have a parabolic effect on timber value); social status (the upper layer has a positive effect on timber value); damage presence (displays a negative effect; Table 5, right part).

Regarding stand variables, the study showed that the values of beech trees were comparatively higher in rejuvenation stands than in other stand types. Moreover, the share of silver fir increased the value, while the share of other coniferous tree species (spruce is strongly dominating) decreased the value. On the other hand, the value of beech trees increased with increasing stand basal area up to 60 m2/ha, and then started lowering.

In the group of site factors, altitude and inclination have a negative effect on beech timber value. The timber value was also affected by the site location, with beech timber from sloped sites normally reaching higher values.

Harvesting intensity as an indicator of forest management activities undertaken in the area during the past ten years had a positive influence on the value of beech wood.

4. Discussion with conclusions – Rasprava i zaključci This study presents the quality and value potential of beech forest stands in the Karavanke region to illustrate the differences between forest types and to outline the factors influencing the quality and value of beech timber. The information illustrating the quality of complete stands derived from forest inventory data was used as the basis for the study. The forest inventory data offers clear advantages, but also raises certain concerns. The main advantages of quality assessment at PSPs are the following: it enables representative quality assessment for selected spatial units, it is simple to conduct, and, consequently, inexpensive. Furthermore, the quality assessment on PSPs is part of numerous measurements and estimates and as such enables researchers to determine the underlying principles and connections between variables. However, although multiple parameters have been measured and assessed on PSPs, it would be sensible to include Croat. j. for. eng. 34(2013)1

A. Poljanec and A. Kadunc

other parameters (e.g. rockiness), in particular if such a parameter is only recorded upon the first measurement and can be obtained through a simple process or assessment. It can therefore be concluded that the method is useful with the assumption that it is carried out correctly. Its main weakness, however, is that the quality assessment is only based on external signs, i.e. external tree defects. In order to calculate the percentage of various timber classes from the assessments of standing trees, an analysis should be made of a small sample of felled trees or certain assumptions should be set based on expert opinions or experiences (e.g. the ratio of sliced veneer : peeled veneer : sawlog I).

The assortment structure established for the Karavanke mountain range is considerably less favourable than the stands of comparable site quality determined for Slovenia by Kadunc and Kotar (2005), and by Kadunc (2006). For the Karavanke, low percentages of sliced veneer and peeled veneer logs were ascertained.

A higher percentage of veener quality trees and a better assortment structure than the one established for the Karavanke were also found in several other countries such as Croatia, Slovak Republic, and Switzerland (Petráš and Nociar 1991; Štefančič 1998; Gfeller 1998; Krpan 2003; Prka and Krpan 2007; Prka 2010).

The results of the Croatian National Forest Inventory showed a high percentage of excellent trees (14%) and an even higher share of very good trees (24%) (Čavlović 2010). On the other hand, the analysis of the quality grading of felled beech trees in the Austrian Federal Forests (in the period 1995/1996) showed an even more unfavorable assortment structure (Rieder 1997).

A certain degree of caution is required when comparing the quality classes obtained in various analyses, largely because the assessed quality grades are heavily dependent on relevant standards used (Prka and Poršinsky 2009), as well as on the assessors (Paladinić and Vuletić 2006; Prka and Krpan 2010). Although the differences in standards may lead to a difference in percentage for a particular timber class, the ratios between quality classes are more or less stable (Prka 2008). In all comparisons, a distinction must be made between forest stand assortment structure and harvest assortments (Prka 2006).

Nevertheless, since the assortment structure of a forest stand depends on the efficiency of silvicultural measures (Prka 2006), the less favorable assortment structure of beech trees in the study area is mainly the result of past forest management practice. In the Alpine territory, beech has always been considered a tree species used primarily for firewood, and forest owners relied on coniferous tree species, in particular spruce and larch, to generate revenue and produce technical

161

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

timber. Beech is frequently found as admixed species in stands (either as individual trees or in clusters), which is not promising for the growth of quality beech trees. Beech trees develop best among other trees of similar age. In addition, due to low marketing appeal, the beech stands in the study area were intensively tended, as was common practice in certain other parts of Slovenia (e.g. Kordiš 1993; Pirc 1997).

Multivariate analysis accounted for a low percentage of variability of timber quality and tree value, as can be expected for such research into beech trees (e.g.

Kadunc 2006), whereas in oak and fir a higher percentage of variability was explained (Kadunc 2009; Kadunc 2010). The formation of heartwood discoloration, commonly known as red heart, reduces the predictability of the quality and value of beech trees and timber. Nevertheless, we believe that there are a number of other methodological reasons for the low level of accountability. Evidently, several studies have shown that the efficiency of explaining the beech timber values could be substantially improved (Rebula and Kotar 2003; Rebula and Kotar 2004). Up to a certain extent, the models would be made more efficient by including other variables, e.g. rockiness, soil type (or other ground characteristics), and proximity to the nearest forest edge, which were unfortunately not available in the present case.

Diameter at breast height was found to have a significant influence on beech tree quality and timber value. The analyses showed that the diameter of the highest-quality and most-valuable trees is between 50 and 55 cm, which is in line with the findings of several other researchers (e.g. Štefančić 1998; Hapla et al.

2002; Prka 2003a; Kadunc 2006). In addition to DBH, the value of beech trees is also considerably affected by several other tree traits: the social status of a tree (higher value is assigned to trees in the upper stand layer, and vice versa), presence of damage (damaged trees carry a lower value), and basal area increment (fast growing trees develop larger crowns and, as a result, lower-quality trunks).

In the group of stand variables, stand basal area, percentage of fir, percentage of conifers, and stand type were also found to influence the value of trees. Analyses showed that stands of excessive density reduce the timber value, and, on the other hand, trees with very relaxed crowns seem to be deficient in quality. It is interesting to note that the presence of fir in stands has a positive influence on the value of beech wood. On the other hand, the presence of other coniferous tree species reduces this value. The high percentage of conifers might point to the fact that beech did not grow in a group of broadleaved trees, which would improve the

162

quality of their stems. The positive effect of fir, however, is indicative of better growing conditions. The fact that the quality was higher in rejuvenation stands than in mature forests and in uneven-aged stands of similar DBH showed that better-quality trees are left to grow longer in regeneration stands.

Altitude, inclination, and landscape position are included in the list of site variables that have a characteristic effect on the value of beech trees in the analyzed area. The effect of altitude is largely seen as the influence of the stand site quality, as one of the key entry data for the forest timber product table (Prka 2003a). The trees at higher elevations are less straight, their branches are stronger, and they grow to commercially attractive dimensions at a higher age, all of these factors contributing to less favorable quality structure. Adding to the extreme site conditions, inclination reduces the value of beech stems. The occurrence of damage caused to trees during harvesting operations is higher on steep sites; the crowns of trees on steep sites are asymmetrical (higher red heart probability, Kadunc 2006), and the shape of the stem deviates more frequently from the circular. It was also found that sites on slopes are more favorable than other sites, with ridge sites as the most prevalent. Moreover, the results of the analysis also showed that the value of trees rises with the felling in stands; which further stresses the importance of active management of stands where beech takes up a considerable percentage.

Prka and Krpan (2007) conclude that the quality of trees and/or stands is a result of various abiotic and biotic factors, including human action. Low management intensity of the studied stands is also evident in the low percentage of damage to trees during timber harvesting. Damage averaged at 6.2% of beech trees, which is lower than observed in intensively managed stands (Prka 2006).

The present beech timber value depends on the prices of forest timber products. When these prices, or the relations between them, change, other underlying principles may also be affected. Prka and Krpan (2010) conclude that the market for beech timber products is relatively unstable. The recent growth of beech firewood prices has been driven by the rise in the prices of fossil fuels. On the other hand, the decline in the production of beech hardwood furniture has led to a drop in the prices of beech sawlogs (EUWID 2011). According to the current market trends, the future of the technical use of beech timber looks rather uncertain.

Under these conditions, it is very difficult to make predictions into the assortment structure of beech trees, in particular because the occurrence of defects, their size and numbers are coincidental in nature and fail to corCroat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

relate with many tree parameters (Prka and Krpan 2010). Heartwood discoloration in particular presents a serious obstacle to beech yield planning (Prka et al.

2009). Highly volatile market poses another issue.

If the quality of beech trees and stands continues to be an important forest management goal in the future, it is sensible to cultivate beech trees in groups of broadleaved trees, favorably mixed up with fir, to strive to prevent damage to selected trees during harvesting, and to carry out thinning.

In our opinion, quality assessments carried out on PSPs enable the researchers to assess the stand value potential, the differences in the quality of certain forest areas and/or forest stand types (strata) as well as the impact of different silvicultural systems on stand quality and yield value. This will lead to enhanced forest management, both with regard to determining investment priorities as well as with regard to setting priority regeneration measures.

Acknowledgement – Zahvala

A. Poljanec and A. Kadunc

Ficko, A., Klopčič, M., Matijašić, D., Poljanec, A., Bončina, A., 2008: Razširjenost bukve in strukturne značilnosti bukovih sestojev v Sloveniji. Zbornik gozdarstva in lesarstva 87: 45–60.

Gfeller, B., 1998: Weisses und verkerntes Buchenholz in der Holzwirtschaft. Schweizerische Zeitschrift für Forstwesen 149(12): 943–954.

Hapla, F., Meggers, H., Militz, H., Mai, C., 2002.:Investigation on the yield and quality of sliced veneer produced from beech trees (Fagus sylvatica L.) containing red heartwood.

Holz als Roh- und Werkstoff 60: 440–442.

Johann, E., 2006: Historical development of nature-based forestry in Central Europe. In: Nature-based forestry in Central Europe (Ed. Jurij Diaci), Studia Forestalia Slovenica 126: 1–17.

JUS, 1979: D. B4. 020-029. Standardi za bukove hlode. Zvezni zavod za standardizacijo.

Kadunc, A., 2006: Kakovost in vrednost okroglega lesa bukve (Fagus sylvatica L.) s posebnim poudarkom ozirom na pojav rdečega srca. Gozdarski vestnik 64(9): 355–376.

Kadunc, A., 2009: Kakovostna zgradba in proizvodne dobe v hrastovih sestojih. UL, BF, Oddelek za gozdarstvo in obnovljive gozdne vire, Zaključno poročilo pri raziskovalnem projektu, 1–29.

The study was conducted in the frame of the karawanks@future.eu – Nature Based Economies in the European Future Region Karavanke project, which is carried out within the Operational Program SloveniaAustria 2007 – 2013, and was co-financed by the European Regional Development Fund and the Government Office of the Republic of Slovenia for Local Self-Government and Regional Policy. The study was additionally supported by Pahernikova ustanova (Contract Number 01/2011) and the Ministry of Agriculture and Environment (Project V4-1123).

Kadunc, A., 2010: Prirastoslovne značilnosti jelke (Abies alba Mill.) v Sloveniji. Gozdarski vestnik 9: 403–422.

5. References – Literatura

Kordiš, F., 1993: Dinarski jelovo bukovi gozdovi v Sloveniji.

Univerza v Ljubljani, Biotehniška fakulteta, Strokovna in znanstvena dela 112: 1–139.

ARSO, 2004a. Environmental Agency of Slovenia. Mean Annual Temperature Map M 1:250,000, Ljubljana.

ARSO, 2004b. Environmenta Agency of Slovenia. Mean Annual Precipitation Map M 1:250,000, Ljubljana.

Buser, S., 1991: Vodnik po Slovenski geološki poti. Ljubljana, Geološki zavod, 1–122. Čavlović, J., 2010: Prva nacionalna inventura šuma Republike Hrvatske. Ministarstvo regionalnog razvoja, šumarstva in vodnoga gospodarstva, Šumarski fakultet Sveučilišta u Zagrebu, Zagreb, 1–300. Čokl, M., 1981: Količina in struktura sečnih ostankov v gozdu. Gozdarski vestnik 39(2): 49–53.

Kadunc, A., Kotar, M., 2005: Volumenska in vrednostna zgradba ter priraščanje sestojev visokokakovostnih bukovih sestojev v Sloveniji. Zbornik gozdarstva in lesarstva 78: 69–96.

Kadunc, A., Poljanec, A., 2011: Quality and timber value of European larch (Larix decidua Mill.) trees in the Karavanke region. Zbornik gozdarstva in lesarstva 96: 23–34.

Kleinbaum, D. G., Klein, M., 2002: Logistic Regression. A SelfLearning Text. Second Edition, Springer Verlag, 1–513.

Knoke, T., 2003: Predicting red heartwood formation in beech trees (Fagus sylvatica L.). Ecological Modelling 169: 295–312.

Kotar, M., 1989: Prirastoslovni kazalci rasti in razvoja gozdov v Sloveniji. Zbornik gozdarstva in lesarstva 33: 59–80.

Kotar, M., 1991: Zgradba bukovih sestojev v njihovi optimalni razvojni fazi. Zbornik gozdarstva in lesarstva 38: 15–40.

Kotar, M., 2003: Gozdarski priročnik. Ljubljana, Oddelek za gozdarstvo in obnovljive vire, BF, UL: 49–50.

Krpan, A. P. B., 2003: Bukovi šumski proizvodi i tehnologije pridobivanja drva iz bukovih sastojina. In: Obična bukva u Hrvatskoj (Eds. Matić, S.), Hrvatske šume, Zagreb, 625–640.

Ellenberg, H., 1986: Vegetation Mitteleuropas mit den Alpen in ökologischer Sicht. 4. Auflage, Ulmer, Sttutgart, 1–989.

Kutnar, L., Veselič, Ž., Dakskobler, I., Robič, D., 2012.

Tipologija gozdnih rastišč Slovenije na podlagi ekoloških in vegetacijskih razmer za potrebe usmerjanja razvoja gozdov.

Gozdarski vestnik 70: 195–214.

EUWID, 2011: Germany: New uses sought for beech roundwood. Wood products and panels, Issue 48/2011 OF 30. 11.

2011.

Mlinšek, D., 1968: Sproščena tehnika gojenja gozdov na osnovi nege. Poslovno združenje gozdnogospodarskih organizacij, Ljubljana, 1–117.

Croat. j. for. eng. 34(2013)1

163

A. Poljanec and A. Kadunc

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

Omahen, R., 1998: Vrednostni prirastek sestoja in njegov pomen v gojenju gozdov. Diplomsko delo, Univerza v Ljubljani, Biotehniška fakulteta, 1–44.

Paladinić, E., Vuletić, D., 2006: Modeliranje sortimentne strukture dubećih stabala bukve. Radovi, Izvanredno izdanje 9: 279–296.

Petráš, R., Nociar, V., 1991: Sortimentačné tabuľky hlavných drevín. Slovenská akadémia vied, Bratislava, 1–304.

Pirc, S., 1997: Vpliv izbiralnih redčenj na rast, razvoj in kakovost sestojev v GGE Brezova reber. Graduation thesis, UL, BF, Oddelek za gozdarstvo, Ljubljana, 1–72.

Poljanec, A., Ficko, A., Bončina, A., 2010: Spatiotemporal dynamic of European beech (Fagus sylvatica L.) in Slovenia, 1970–2005. Forest Ecology and Management 259: 2183–2190.

Prka, M., 2003a: Vrijednosne značajke bukovih stabala prema vrsti sijeka u sječinama Bjelovarske Bilogore. Šumarski list 127(1-2): 35–44.

Prka, M., 2003b: Pojavnost neprave srži kod bukovih stabala i tehničke bukove oblovine iz prorednih i pripremnih sječa na području Bjelovarske Bilogore. Šumarski list 127(9–10): 467–474.

Prka, M., 2006: Značajke doznačenih bukovih stabala po vrsti sijeka u sječinama Bjelovarske Bilogore i utjecaj na strukturu sortimenata. Šumarski list 130(7–8): 319–329.

Prka, M., Poršinsky, T., 2009: Usporedba strukture tehničke oblovine jednodobnih bukovih sječina u sortimentnim tablicama izrađenih primjenom normi HRN (1995) i HRN EN 1316-1: 1999. Šumarski list 133(1–2): 15–25.

Prka, M., Zečić, Ž., Krpan, A. P. B., Vusić, D., 2009: Characteristics and share of European Beech False Heartwood in Felling Sites of Central Croatia. Croatian Journal of Forest Engineering 30(1): 37–49 Rebula, E., 2002: Izkoristek lesa pri sečnji bukovine. Zbornik gozdarstva in lesarstva 69: 197–213.

Rebula, E., 2005: Količinski in vrednostni prirastek drevja v revirju Mašun. Gozdarski vestnik 63(3): 115–130.

Rebula, E., 2011: Zakonitosti količinskega in vrednostnega priraščanja gospodarskih gozdov na primeru gozdnogospodarskih enot Mašun in Leskova dolina. Zveza gozdarskih društev Slovenije – Gozdarska Založba, Ljubljana, 1–136.

Rebula, E., Kotar, M., 2003: Vrednost bukovine in bukovega drevja. Gozdarski vestnik 3: 132–146.

Rebula, E., Kotar, M., 2004: Stroški sečnje in spravila bukovih dreves ter vrednost bukovine na panju. Gozdarski vestnik 4: 187–200.

Rieder, A., 1997: Einflußmöglichkeiten auf die Farbkernausbildung bei Rotbuche. Österreichische Forstzeitung 5: 34–36.

Prka, M., 2008: Određivanje sortimentne strukture jednodobnih bukovih sastojina primjenom norme HRN EN 13161: 1999. Šumarski list 132(5-6): 223–238.

Šmajdek, K., 2001: Vpliv rdečega srca pri bukvi v fitocenozah asociacij Lamio orvalae-Fagetum in Cardamini savensi-Fagetum na kvaliteto lesa. Diplomsko delo, Univerza v Ljubljani, Biotehniška fakulteta, 1–66.

Prka, M., 2010: Bukove šume i bukovina bjelovarskog područja: sortimentna struktura - čimbenici, planiranje, problemi i rješenja. Bjelovar, Hrvatsko šumarsko društvo, 1–252.

Štefančić, A., 1998: Udio drvnih sortimenata u volumenu krupnog drva do 7 cm promjera za običnu bukvu u jednodobnim sastojinama. Šumarski list 122(7-8): 329–337.

Prka, M., Krpan, A. P. B., 2007: Problem određivanja sortimentne strukture jednodobnih bukovih sastojina. Šumarski list 131(5-6): 219–235.

Veber, I., 1986: Gozdovi bohinjskih fužinarjev. Gozdnogospodarstvo Bled, Bled, 1–48.

Prka, M., Krpan, A. P. B., 2010: Impact of Tending Measures on Assortment Structure of Fellings in Central Croatian Beech stands. Acta Silv. Lign. Hung. 6: 171–182.

SFS 2010a. National forest inventory data. Data from permanent sampling plots, first and second survey. Slovenia Forest Service.

SFS 2010b. National forest inventory data, atabasis odseki. dbf, odsses.dbf, odssesdv.dbf. Slovenia Forest Service.

Sažetak

Kakvoća i vrijednost oblovine obične bukve (Fagus sylvatica L.) na Karavankama Bukva je u Alpama dugo vremena bila sinonim za ekonomsko manje vrijednu vrstu drveća. U posljednjim se desetljećima, posebno zbog promjene načina gospodarenja šumama, povećava udio bukve i njezina važnost u alpskom prostoru. Zato se nameće pitanje o načinu gospodarenja bukovim šumama u budućnosti. Za pravilnu procjenu stanja potrebne su mnoge informacije, kao što su poznavanje strukture kakvoće šumskih sastojina i vrijednosnih svojstava vrsta drveća te svih čimbenika koji na to utječu. Svrha je ovoga rada utvrditi kakvoću i vrijednost svojstava bukve u Karavankama i identificirati čimbenike koji utječu na vrijednost bukovih stabala.

Za ocjenu kakvoće stabala služili smo se podacima trajnih pokusnih ploha (7154 stabla na 2088 ploha) te podacima manjega uzorka posječenih stabala (495 stabala, 26 ploha). Vrijednost bukovih sortimenata izračunali smo

164

Croat. j. for. eng. 34(2013)1

Quality and Timber Value of European Beech (Fagus sylvatica L.) Trees... (151–165)

A. Poljanec and A. Kadunc

pomoću njihove utvrđene strukture te cijene sortimenata na kamionskoj cesti. Testirali smo čimbenike koji utječu na kakvoću ili vrijednost bukovih stabala. To su stablimični, sastojinski, stanišni i šumskogospodarski parametri koje prikupljamo za vrijeme inventure šuma na trajnim pokusnim plohama.

Istraživanje je pokazalo izuzetno nizak postotak stabala dobre kakvoće. Tako je udio furnirskih trupaca (klasa F) iznosio tek 0,1 posto, 0,6 posto je iznosio udio trupaca za ljuštenje (klasa L), dok je udio pilanskih trupaca prve klase (najviše kakvoće) bio 1,6 posto. Najveći zajednički udio furnirskih trupaca i trupaca za ljuštenje utvrdili smo na staništima asocijacija Aceri-Fraxinetum, Aposeri-Piceetum i Hacquetio-Fagetum, dok klasu F ili L nismo našli na staništima Castaneo-Fagetum, Bazzanio-Abietetum i Adenostylo-Fagetum.

Na kakvoću i vrijednost bukovih stabala ima, od svih istraženih parametara, najveći utjecaj prsni promjer. Analize su pokazale da stabla najviše kakvoće i vrijednosti imaju prsni promjeri od 50 do 55 cm. Od ostalih parametara na kakvoću bukovih stabala najviše utječu socijalni status, oštećenja na stablima te prirast temeljnice. Neoštećena su stabla u gornjem sloju općenito bolje kakvoće i njihova je vrijednost veća.

Od sastojinskih parametara na vrijednost stabala utječu temeljnica sastojine, udio jele, udio četinjača te tip sastojine. Analize su pokazale da sastojine veće gustoće smanjuju vrijednost, a, s druge strane, ni sastojne sa slobodnim krošnjama nisu najbolje kakvoće. Zanimljiv je rezultat istraživanja o pozitivnom utjecaju udjela jele te negativnom utjecaju udjela četinjača. Velik dio crnogoričnih stabala upućuje na to da bukva vjerojatno nije rasla u grupi listopadnoga drveća, što bi sigurno poboljšalo njezinu kakvoću. Pozitivan udio jele vjerojatno upućuje na povoljne stanišne uvjete. U pomlađenim je sastojinama kakvoća starih stabala veća nego u optimalnim razvojnim fazama, što je posljedica načina rada, kada se za vrijeme obnove šume najbolja stabla ostavljaju duže razdoblje. Analiza je također pokazala da vrijednost trupaca raste s povećanjem sječe u sastojini. To je svakako dokaz o potrebi aktivnoga gospodarenja sastojinama za većim udjelom bukve.

Nadmorska visina, nagib i položaj u krajoliku su varijable staništa koje su značajno utjecale na vrijednost bukovih stabala na području istraživanja. Utjecaj nadmorske visine uglavnom odražava utjecaj boniteta sastojine. Stabla na većim nadmorskim visinama manje su ispružena, grane su deblje, komercijalno zanimljive dimenzije stabla dostižu u starijoj dobi, što sve pridonosi nepovoljnijoj strukturi kakvoće. Povećanjem nagiba padina smanjuje se vrijednost trupaca: prilikom njihova privlačenja često dolazi do većih oštećenja, krošnje su stabala asimetrične (povećava se vjerojatnost nastanka crvenoga srca), oblik je trupaca obično također asimetričan. Pokazalo se da su padine primjerenije u usporedbi s položajima na grebenima.

Visoka kakvoća bukovih stabala i sastojina, kao važan cilj gospodarenja šumama u budućnosti, upućuje na potrebu uzgoja takvih sastojina u grupama listopadnih stabala kod kojih je poželjan manji udio jele. Posebnu pozornost valja posvetiti privlačenju drva radi smanjenja oštećenja dubećih stabala. Potrebno je provoditi redovite prorede. Ocjena kakvoće na trajnim pokusnim plohama može biti koristan pokazatelj za procjenu vrijednosnih mogućnosti sastojina te za utvrđivanje razlika u kakvoći pojedinih područja ili tipova (stratuma) šuma. To pridonosi boljemu upravljanju šumama, osobito pri odlučivanju o intenzitetu i učestalosti različitih uzgojnih mjera te o početku i načinu obnove šume.

Ključne riječi: obična bukva, kakvoća debla, sortimentna struktura, vrijednost drvne oblovine, utjecajni čimbenici, Karavanke

Authors’ address – Adresa autorâ:

Received (Primljeno): July 18, 2012 Accepted (Prihvaćeno): October 15, 2012 Croat. j. for. eng. 34(2013)1

Aleš Poljanec, PhD. e-mail: ales.poljanec@zgs.gov.si Zavod za gozdove Slovenije Večna pot 2 SI-1000 Ljubljana SLOVENIA Asst. Prof. Aleš Kadunc, PhD.* e-mail: ales.kadunc@bf.uni-lj.si University of Ljubljana, Biotechnical Faculty Department of Forestry and Renewable Forest Resources Večna pot 83 1000 Ljubljana SLOVENIA * Corresponding author – Glavni autor

165

Subject review – Pregledni rad

Recent Challenges of Forest Engineering Academic Education Marijan Šušnjar, Milan Oršanić, Tibor Pentek, Tomislav Poršinsky, Mario Šporčić Abstract – Nacrtak This paper presents the history and tradition of forestry academic education in Europe. In the last few decades a lot of changes have occurred and forestry academic education has lost a highlevel position in European society and economy.

In many European countries, higher education forestry programs are just a part of the educational system of universities of applied sciences, life sciences, agricultural or technical sciences.

Important changes of higher education systems have been made with the establishment of new educational programs according to the Bologna principles. Some problems and disadvantages will be discussed on the Croatian example, where the reform of teaching programs according to the Bologna process was made 7 years ago.

The most important problems are related to the employment of bachelors of forestry, decreasing number of students, and evaluation of forestry educational programs in forestry practice.

Also the field of forest engineering, which still has a quite strong position in forestry practice, is not adequately present in teaching programs in many countries.

In order to provide the place that belongs to higher forestry education in the Republic of Croatia and Europe, the existing study programs should be revised and updated with the development of new teaching methods, promotion of new and updated teaching materials and handbooks, encouragement of students for scientific and individual work.

With the aim to preserve and recognize the importance of education in the field of forest engineering, some existing principles of Bologna process could be accepted. First, the need should be stressed for the development of joint master studies in the framework of Erasmus Mundus program, which would strengthen the existing cooperation and serve as the basis for a more fruitful scientific-research and educational work.

It needs to be pointed out that the establishment of new graduate studies in the field of forest engineering would contribute to the development of higher education in the European environment, with an aim of providing student and academic staff mobility.

Keywords: educational programs, forestry, forest engineering, Bologna principles

1. Introduction – Uvod Forestry is a profession, science and art of managing and preserving forest ecosystems, whose purpose is to secure permanent benefits to man, society and nature. According to Helms (1998) forestry is the science and profession of creating, managing, using, conserving, and repairing forests and associated resources in a sustainable manner to meet desired goals, needs, and values for human benefit.

The first signs of forestry can be found in the Roman province of Gaul, where Pliny mentioned beautiCroat. j. for. eng. 34(2013)1

ful, managed forests called »sylvae caeduae« (Anon 1963). In the Middle Ages, the beginnings of forestry can be found in the statutes of cities, where certain provisions were mainly related to prevent excessive logging. With the time, they started giving instructions how to manage forests.

Forestry as a science appeared in the 18th century. It sprang from the need to provide a sustainable use of forest resources and preserve forests after deforestation (disappearance of water springs, onset of torrents, soil erosion, formation of bare rock, decrease in soil fertility, loss of forest resources) resulting from intensive cutting

167

M. Šušnjar et al.

Recent Challenges of Forest Engineering Academic Education (167–172)

operations in the 17th century and the first half of the 18th century. Forestry is based on the principle of sustainability, defined in 1712 by the German scientist H. C. W. Carlowitz, who published the famous paper »Sylvicultura oeconomica oder hauswirtliche Nachriht und naturmässige Anweisung zur wilden Baum-Zuht nebst gründlicher Darstellung wie zuförderst durch götliches Benedeyen den allenthalben und insgemein einreissenden grossen Holz-Mangel vermittels Säe-Pflanz- und Versetzung vielerhand Bäumen za rathen« (Anon 1998).

It is still the only proper example of sustainable development and sustainable management.

Roots of forestry science can be found in the works of the French naturalists R. Reaumur (1721); G. Buffon (1739); G. Duhamel du Monaceau (1764) and Varenne de Fenille (1791); they set some basic principles of forestry and defined the growing stock, increment, age classes in even-aged forest, absolute maturity of forest, etc. (Anić et al. 2012).

The Forest Order issued by the Austro-Hungarian Empress Maria Theresa in 1769 is an important document for forestry in Central and Eastern Europe as it provided felling maturity for the main tree species. In the Austro-Hungarian Empire professional forestry services, based on certain principles of forestry science, were legally prescribed and introduced in 1894 (Anić et al. 2012).

The development of forestry services and forestry science leads to the establishment of state institutions that systematically engaged in forest management and scientific research, as well as to the development of the first educational programs for forestry experts. The International Union of Forest Research Organizations, as the only international organization that coordinates forest science efforts world-wide, was founded in 1892 at Eberswalde University in Germany (www.iufro.com).

2. Forest Engineering in terms of academic education – Šumarske tehnike i tehnologije u sveučilišnom obrazovanju The first forestry schools in Europe were established at the beginning of the 19th century and namely: 1807 in Würtenberg (Germany); 1813 in Mariabrunn (Austria); 1816 in Tharandt (Germany); 1824 in Nancy (France); 1828 in Stockholm (Sweden); 1846 in Bánska Štiavnica (Check); 1872 in Florence (Italy) and in Vienna (Austria); 1885 in Zurich (Switzerland), etc.

(Anon. 1998). Very soon forestry schools became a part of universities and forestry education entered into the system of academic education.

The basic principles of forest sustainability, which were generally accepted, were introduced and evolved

168

through newly created courses of forestry academic curricula: forest management, forest silviculture, forest pedology, forest exploitation, etc.

The beginning of higher forestry education in Croatia dates back to 1860, when the High School of Agriculture and Forestry was established in Križevci and to 1898, when the Forestry Academy was founded in Zagreb. With this Academy, forestry education in Croatia gained university status and it was the fourth high school of the University of Zagreb.

A 4-year study was introduced at the beginning of the academic year 1907/08 and the study courses were divided into: common science courses, natural science courses, technical courses, legal and regulatory courses and forestry professional courses.

The above distribution of courses was in accordance with the notion that forestry, as a joint action of science and profession, should enable finding, adopting and implementing the most favorable solutions for the preservation and maintenance of forest ecosystem stability and realization of benefits that forests can offer. In order to do this in accordance with biological and ecological criteria (observing forest demands) and economic criteria (in accordance with economy laws), the solutions must be provided by the technical and technology component of forestry. Meanwhile, the term forest engineering has been introduced instead of technical and technology component of forestry.

In general, forest engineering could be defined as the application of engineering principles to the solution of forestry problems, such as those dealing with harvesting, forest transportation, materials handling, and mechanical silviculture, with regard to long-range environmental and economic effects.

The current development of study courses of forest engineering has enabled the introduction of mechanized labor, first replacing manual labor with manualmachine work, and later completely with machine work, all with the goal of increasing productivity and worker protection.

At the beginning of systematic forest management, timber harvesting was performed manually, using simple tools and animal power. In 1796, J. Watt invented the steam engine, and in 1878 the English engineer Ransome tried to use the steam engine for felling saws. The first chainsaw was designed by the Swedish engineer Westfalt in 1916 and he perfected it in 1924. The saw was known as »Sector«. In 1954 the manufacturer »Stihl« produced a chainsaw that could be handled by one worker. In 1959 the production of petrol chainsaw »Stihl Contra« was started. The use of motor powered chainsaws raised the level of productivity in forestry and humanized tree felling and processing.

Croat. j. for. eng. 34(2013)1

Recent Challenges of Forest Engineering Academic Education (167–172)

The mechanization of timber skidding, by using agricultural tractors (mostly crawler tractors), began in the period just before the World War II, and after the war it became quite intense.

As early as 1951, the first skidder was produced in USA and in 1954 the first skidder with articulated steering was produced. In the Scandinavian countries, skidders have been used in forestry since 1962, in Austria and Germany since 1964, and in Croatia since 1968.

Further development of special forest machines continued with the design, manufacturing and use of the first processors (1965, Canada), forwarders (1961, Sweden), feller-bunchers (1957, Canada), single-grip harvesters (1966) (Drushka and Konttinen 1997).

The period of intensive mechanization of forestry operations, particularly timber harvesting lasted until the ‘80s of the last century. Forestry academic education followed forestry practice and educational programs and courses formed the field of forest engineering. It developed quickly in order to cover the new areas: implementation of new technologies, research productivity and cost efficiency of mechanized work, organization of forestry works, optimization of harvesting systems, forest opening and construction of forest roads, ergonomics in forestry.

Soon, scientists and experts in the field of forest engineering faced a new task: assessing the impact of forest machines on forest environment. The International Society for Terrain-Vehicle Systems (ISTVS) was founded in 1962 at the request of the General Assembly of the First International Conference on the Mechanics of Soil Vehicle Systems held in Turin, Italy. The motivation behind this was to acquire and advance knowledge of the mechanics of terrain-vehicle systems and machinery interacting with soils in all environments (www.istvs.org).

The firsts Symposium »Forestry Mechanization« (today »FORMEC«) was held in 1966 at Zvolen in former Czechoslovakia, as a meeting of European professors from the field of forest harvesting. The original idea of the meeting was to give an opportunity to scientists from Eastern and Central European countries to gather and discuss about forest engineering matters, deepen the knowledge in that field of science and overcome gaps due to different technology levels (Šušnjar 1999).

At that time, forestry academic education was provided by a number of universities across Europe.

Modern forestry generally embraces a broad range of concerns, including assisting forests to provide timber as raw material for wood products, wildlife habitat, natural water quality management, recreation, landscape and community protection, employment, aesthetically appealing landscapes, biodiversity manCroat. j. for. eng. 34(2013)1

M. Šušnjar et al.

agement, watershed management, erosion control, and preserving forests as ‘sinks’ for atmospheric carbon dioxide.

Forest ecosystems have come to be seen as the most important component of the biosphere, and forestry has emerged as a vital field of science and technology.

Kennedy and Koch (2008) explained changes in forestry and forestry education caused by the rise in awareness of the importance of forests and forest ecosystems. They defined »human-ecosystem relationship« as the concept of managing natural resources for multiple and diverse social values, which were developed to make sense of all the different sociocultural, economic and political/legal systems and impacting traditional forest management. The new approach, the so called social value model of ecosystem management, increasingly required from forest professional educators and managers to include social considerations.

Today, the forestry profession includes a wide diversity of jobs, with different educational requirements ranging from bachelor’s degrees to PhDs for highly specialized work.

Over the past centuries, forestry was regarded as a separate science. With the rise of ecology and environmental science, there has been a reordering in the applied sciences.

As the result of reorganization of universities across Europe (excluding faculties as organizational units) and the increasing awareness of the importance of forest social functions, forestry in the traditional sense disappears from the names of many former forestry faculties, particularly in Western Europe.

In many European countries the higher education forestry programs are just a part of educational systems of universities of applied sciences, life sciences, agricultural or technical sciences. Forestry faculties under the same name have remained only in the southeastern and eastern Europe.

3. Challenges of Bologna process – Izazovi bolonjskoga sustava Important changes of high education systems have been made with the establishment of new educational programs according to the Bologna principles. The reform of the higher educational system was initiated at the Ministerial Conference in Bologna in 1999 with the issue of the famous Bologna Declaration and has been further developed at subsequent biannual conferences of the European Union Ministers. The Bologna Declaration is the core document for the reform of the higher educational system in the European Area of Higher Education (EHEA).

169

M. Šušnjar et al.

Recent Challenges of Forest Engineering Academic Education (167–172)

Guidelines for the development of higher educational and research processes in the European area are contained in a number of political and professional decisions and documents of the European Union. The European Council convened in Lisbon in March 2000, where the representatives of the governments of the European Union passed knowledge-based strategic goals intended to increase employment and strengthen economic reforms and social cohesion in the decade to come (»to become the most competitive and dynamic knowledge-based economy in the world capable of sustainable economic growth with more and better jobs and greater social cohesion«).

The reform of the higher educational system was completed at the Berlin Ministerial Conference in 2003, where doctoral studies were put in a three-tier cycle of the higher education – undergraduate, graduate and doctoral.

The higher educational system has been reformed with the goal of providing more purposeful education and enabling competent and skilled young experts to join global economic trends and create their own careers. The criteria of excellence, competence and mobility rank high in this reform.

According to the agendas from SILVA Network annual conferences in the last few years, almost the same problems about the adoption of the Bologna process in forestry academic education have occurred in all European countries. The SILVA Network is an association of institutions that offers educational programs leading to Master’s and Doctoral Degrees in forestry with the primary objective to stimulate and facilitate educational co-operation in the field of forestry in Europe.

Some problems and disadvantages will be discussed on the Croatian example, where the reform of teaching programs according to the Bologna process was made 7 years ago.

The most important problems are related to the employment of bachelors of forestry, decreasing number of students, and evaluation of forestry educational programs in forestry practice. The solution of these problems lies in the cooperation between forestry education-science and forestry practice.

In our opinion, the problem is also in the fact that the government has promised to secure employment of bachelor graduates. In case of forestry, it is important to take into account the professional needs for certain qualifications. With the development of qualification frameworks, it is necessary to align the learning outcomes of study programs in the field of forest engineering with working competencies and demands for performing forest operations.

170

Many employers in the field of forest engineering (private contractors, state enterprises and forestry institutions) often criticize the system of forestry education in accordance with the Bologna process, or the acquired competencies of bachelors and masters in forestry. On the other hand, they have mostly not adjusted the systemization of their jobs to the professional status of persons with the academic education. The problem occurs primarily with the employment of Bachelors in Forestry, whose education is considered incomplete for managing all forest operations, while Masters in Forestry are still compared with the previous Diploma Engineers. Therefore, most bachelor graduates pursue a master study. It is quite common for the bachelors to look for employment for some time and to continue the master studies after a year or two.

Due to this situation and poor relationship between forestry practice and higher forestry education, students have lost interest for forestry academic programs. Only in countries with a strong forestry tradition, the number of forestry students has not fallen.

Strained relations between forestry practice and forestry higher education irreversibly leads to the decline of the role and importance of forestry profession. The increasing emphasis on social values of forests and the protection of natural areas and forest ecosystems also leads to a decreasing importance of higher education in the field of forest engineering. These problems are not expressed only in countries where forestry as an economic sector takes a considerable part in the economy of the country.

Therefore, steps should be taken promptly and without delay to stop this negative trend and to change its direction in near future. It is necessary to develop a new brand related to these undergraduate studies that will be recognizable among potential, future students. The following steps should be made in order to solve the said problem: Þ The existing programs of undergraduate studies should be revised and updated, Þ The revision of the curricula need to be done taking into account stakeholder interests, Þ Employment profiles of bachelor graduates need to be more defined, Þ Promotion of undergraduate studies should be continuous and intensive, and love towards forests and forestry should be encouraged with future students at an early age.

On the other hand, the improvement and recognition of bachelor graduates and better opportunities for their employment can result in a decreased number of master study students.

Croat. j. for. eng. 34(2013)1

Recent Challenges of Forest Engineering Academic Education (167–172)

Another experience gained is the problem related to the regulation of admission to master studies and the flexibility within the master studies. This led to some problems in master courses as students’ knowledge was inadequate, and also at the labor market as the graduates failed to meet the expectations of the employers.

It is necessary to further analyze the requirements for admission to graduate studies in forestry, especially in the narrow field of forest engineering. In recent years, at the Faculty of Forestry, University of Zagreb, bachelors of science in agronomy and even in technical fields (construction, transport) have shown interest for enrolling in master study »Techniques, Technologies and Management in Forestry«. Differences in their learning outcomes in relation to the undergraduate forestry students are enormous, and each student should prove that he/she meets the admission requirements for the master study.

Another important topic is the impact of the Bologna system on internationalization through mobility of students and teaching staff.

In general, the Bologna system has actually failed to further enhance the exchange of students between universities and countries. The recognition of foreign studies/courses is not easy, as the curricula are not harmonized.

When applying for one semester mobility within Erasmus or CEEPUS program, students require the Learning Agreement and want to choose courses equivalent in semester to courses of their home university.

Due to the non-harmonized curricula, choosing equivalent courses in semester becomes impossible.

By choosing equivalent courses, undergraduate students in Erasmus exchange program often attend graduate study courses at the foreign higher education institution. By attending equivalent courses from different studies, students gain knowledge that is recognized as relevant, meaning that they do not have to prolong their studies by attending courses and taking exams for compensating differences between semester courses at the domestic and foreign higher education institution.

Student mobility imposes a need to conduct more courses in English or another foreign language, or to establish graduate studies in English.

It should be pointed out that the establishment of new graduate studies from the field of forest engineering would contribute to the development of higher education in the European environment, with the aim of providing student and academic staff mobility.

In order to preserve and recognize the importance of education from the field of forest engineering, some existing principles of Bologna process could be acceptCroat. j. for. eng. 34(2013)1

M. Šušnjar et al.

ed. First, the need should be stressed for the development of joint master studies, which would strengthen the existing cooperation and serve as the basis for a more fruitful scientific-research and educational work.

4. Conclusions – Zaključci This paper deals with the problems and challenges faced by the higher education in the field of forest engineering in Europe, caused by the reform of higher education system in Europe and changes in forestry caused by an ever evolving society. In doing so, emphasis is given to only some of the problems and challenges that are currently most evident. To consider all advantages and disadvantages, and provide guidance for further development of higher education in the field of forest engineering is a very complex job that needs to be thoroughly implemented with the involvement of all relevant factors - universities, scientificresearch, government and professional institutions, forestry employers. Currently, the most important task is to firmly connect the enumerated factors and to provide effective transfer of forestry engineering knowledge and technology from higher education and scientific institutions to forestry practice as well as to the industrial sector of forest machinery and equipment.

5. References – Literatura Anić, I., Meštrović, Š., Matić, S., 2012: Značajni događaji iz povijesti šumarstva u Hrvatskoj. Šumarski list 123 (3–4): 169–177.

Anon., 1963: Šumarska nastava u Hrvatskoj 1860–1960. Eds: N. Neidhardt, M. Androić. Šumarski fakultet Sveučilišta u Zagrebu 1998, 618 p.

Anon., 1998: Sveučilišna šumarska nastava u Hrvatskoj 1898–1998. Knjiga druga – Sto godina sveučilišne nastave u Hrvatskoj. Ed. S. Matić. Šumarski fakultet Sveučilišta u Zagrebu, 709 p.

Anon., 2005: ECTS usersž guide. Directorate general for Education and Culture. Brussels, 51 p.

Drushka, K., Konttinen, H., 1997: Tracks in the forest. Timberjack group Oy. Otava printing works, Keuruu, Finland, 254 p.

Gerzabek, M. H., Stampfer, K., 2010: Experiences of the University of Natural resources and Applied Life Sciences (BOKU) in the Bologna system with the special reference to studies in forestry. Paper presented at the SILVA Network Conference June 17th – June 19th, Zagreb, Croatia, 6 p.

Helms, J. A., 1998: Dictionary of Forestry. The Society of American Foresters. Bethseda, USA. Kennedy, J. J., Koch, N. E., 2008: Forest educations in a changing World. Proceedings of the SILVA Network Conference May 15th – May 17th. Eds: P. Schmidt, S. Lewark, N. Strange.

University of Copenhagen, Denmark, p. 9–17.

Krpan, A. P. B., 2007: Faculty of Forestry, University of Zagreb and Department of Forest Engineering in the Bologna Process.

Croatian Journal of Forest Engineering 28 (1): 123–136.

171

M. Šušnjar et al.

Recent Challenges of Forest Engineering Academic Education (167–172)

Pentek, T., Oršanić, M., Papa, I., 2008: Prošlost, sadašnjost i budućnost sveučilišne šumarske nastave u Republici Hrvatskoj. Nova mehanizacija šumarstva 29 (1): 61–72.

Pentek, T., Poršinsky, T., 2008: 110 Years of University Forestry Education in the Republic of Croatia. Croatian Journal of Forest Engineering 29 (2): 109–112.

Sever, S., 1980: Istraživanje nekih eksploatacijskih parametara traktora kod privlačenja drva. Disertacija, Šumarski fakultet Sveučilišta u Zagrebu, 301 p. Šušnjar, M., 1999: Ususret FORMEC 1999. Mehanizacija šumarstva 22(3): 181. www.istvs.org

Sažetak

Izazovi u sveučilišnom obrazovanju iz područja šumarskih tehnika i tehnologija Članak prikazuje povijest i tradiciju šumarskoga sveučilišnoga obrazovanja u Europi. Tijekom posljednjih nekoliko desetljeća dogodile su se mnoge promjene koje su donekle smanjile razinu nekadašnjega položaja šumarskoga sveučilišnoga obrazovanja u europskom društvu i gospodarstvu.

U mnogim su europskim zemljama visokoškolski šumarski programi samo dio obrazovnoga sveučilišnoga sustava najčešće primijenjenih znanosti, prirodnih znanosti, poljoprivrednih ili tehničkih znanosti.

Važne promjene u visokim obrazovnim sustavima napravljene su s uspostavom novih obrazovnih programa u skladu s bolonjskim načelima. Na hrvatskom primjeru, gdje je bolonjski sustav uveden prije sedam godina, prikazani su neki od problema i nedostataka novoga sustava.

Najvažniji se problemi javljaju u zapošljavanju prvostupnika šumarstva, u smanjenju broja novih (upisanih) studenata te procjeni šumarskih obrazovnih programa koju daje šumarska praksa. Također, smjer šumarskih tehnika i tehnologija, koje same po sebi još uvijek imaju vrlo važan položaj u šumarskoj praksi, nisu u dovoljnoj mjeri zastupljene u nastavnim programima mnogih europskih zemalja.

Kako bi se u budućnosti osigurao dostojan položaj sveučilišnoga šumarskoga obrazovanja u Republici Hrvatskoj, ali i u cijeloj Europi, postojeći studijski programi trebaju se preraditi i ažurirati u skladu s razvojem novih nastavnih metoda, trebaju se promicati novi nastavni materijali i priručnici te se trebaju poticati studenti na znanstveni i samostalni rad.

Radi očuvanja i prepoznavanja važnosti obrazovanja iz područja šumarskih tehnika i tehnologija neka se od postojećih načela bolonjskoga sustava mogu prihvatiti. Primarno valja istaknuti potrebu za uspostavom zajedničkih diplomskih studija u okviru programa Erasmus Mundus koji će ojačati postojeću suradnju i poslužiti kao osnova za kasnije uspješno zajedničko znanstvenoistraživačko djelovanje i obrazovanje.

Istaknuta je važnost osnivanja novih diplomskih studija iz područja šumarskih tehnika i tehnologija, što će pridonijeti razvoju visokoga obrazovanja u Europi radi povećanja pokretljivosti studenata i sveučilišnoga osoblja.

Ključne riječi: obrazovni programi, šumarstvo, šumarske tehnike i tehnologije, bolonjski sustav obrazovanja

Authors’ address – Adresa autorâ:

Received (Primljeno): November 17, 2011 Accepted (Prihvaćeno): September 11, 2012

172

Assoc. Prof. Marijan Šušnjar, PhD. e-mail: susnjar@sumfak.hr Prof. Milan Oršanić, PhD. e-mail: milan.orsanic3@zg.t-com.hr Prof. Tibor Pentek, PhD.* e-mail: tpentek@sumfak.hr Prof. Tomislav Poršinsky, PhD. e-mail: porsinsky@sumfak.hr Assoc. Prof. Mario Šporčić, PhD. e-mail: sporcic@sumfak.hr Zagreb University Faculty of Forestry Svetošimunska 25 10 000 Zagreb HRVATSKA * Corresponding author – Glavni autor Croat. j. for. eng. 34(2013)1

ISSN 1845-5719

9 771845 571000