Mar vasto final article

Hazard Evaluation in ValparaĂ­so: the MAR VASTO Project

Pure and Applied Geophysics pageoph ISSN 0033-4553 Volume 168 Combined 3-4 Pure Appl. Geophys. (2010) 168:543-582 DOI 10.1007/ s00024-010-0164-3

1 23

Your article is protected by copyright and all rights are held exclusively by Springer Basel AG. This e-offprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to self-archive your work, please use the accepted author’s version for posting to your own website or your institution’s repository. You may further deposit the accepted author’s version on a funder’s repository at a funder’s request, provided it is not made publicly available until 12 months after publication.

1 23

Author's personal copy Pure Appl. Geophys. 168 (2011), 543–582 2010 Springer Basel AG DOI 10.1007/s00024-010-0164-3

Pure and Applied Geophysics

Hazard Evaluation in Valparaı´so: the MAR VASTO Project MAURIZIO INDIRLI,1 HOBY RAZAFINDRAKOTO,2,3 FABIO ROMANELLI,4 CLAUDIO PUGLISI,1 LUCA LANZONI,5 ENRICO MILANI,6 MARCO MUNARI,7 and SOTERO APABLAZA8 Abstract—The Project ‘‘MAR VASTO’’ (Risk Management in Valparaı´so/Manejo de Riesgos en Valparaı´so), funded by BID/ IADB (Banco InterAmericano de Desarrollo/InterAmerican Development Bank), has been managed by ENEA, with an Italian/ Chilean joined partnership and the support of local institutions.

Valparaı´so tells the never-ending story of a tight interaction between society and environment and the city has been declared a Patrimony of Humanity by UNESCO since 2003. The main goals of the project have been to evaluate in the Valparaı´so urban area the impact of main hazards (earthquake, tsunami, fire, and landslide), defining scenarios and maps on a geo-referenced GIS database. In particular, for earthquake hazard assessment the realistic modelling of ground motion is a very important base of knowledge for the preparation of groundshaking scenarios which serve as a valid and economic tool to be fruitfully used by civil engineers, supplying a particularly powerful tool for the prevention aspects of Civil Defense. When numerical modelling is successfully compared with records (as in the case of the Valparaı´so, 1985 earthquake), the resulting synthetic seismograms permit the generation of groundshaking maps, based upon a set of possible scenario earthquakes. Where no recordings are available for the scenario event, synthetic signals can be used to estimate ground motion without having to wait for a strong earthquake to occur (pre-disaster microzonation). For the tsunami hazard, the available reports, [e.g., SHOA (1999) Carta de Inundacion por Tsunami para la bahia de Valparaı´so, Chile, http://www.shoa.cl/servicios/ citsu/citsu.php], have been used as the reference documents for the hazard assessment for the Valparaı´so site. The deep and detailed studies already carried out by SHOA have been

1

ENEA (Italian National Agency for New Technologies, Energy and Sustainable Economic Development), Bologna and Rome, Italy.

2 Laboratory of Seismology and Infrasound, Institute and Observatory of Geophysics, Antananarivo, Madagascar.

3 ICTP (The Abdus Salam International Centre for Theoretical Physics), Trieste, Italy.

4 Department of Geosciences, University of Trieste, Via Weiss 4, 34127 Trieste, Italy. E-mail: romanel@units.it 5 IUSS (University Institute for Higher Studies), University of Ferrara, Ferrara, Italy.

6 Department of Engineering, University of Ferrara, Ferrara, Italy.

7 Department of Structural and Transportation Engineering, University of Padua, Padua, Italy.

8 Board of Architects of Chile, Valparaı´so, Chile.

complemented with (a) sets of parametric studies of the tsunamigenic potential of the 1985 and 1906 scenario earthquakes; and (b) analytical modelling of tsunami waveforms for different scenarios, in order to provide a complementary dataset to be used for the tsunami hazard assessment at Valparaı´so. In addition, other targeted activities have been carried out, such as architectonic/urban planning studies/vulnerability evaluation for a pilot building stock in a historic area and a vulnerability analysis for three monumental churches. In this paper, a general description of the work is given, taking into account the in situ work that drove the suggestion of guidelines for mitigation actions.

Key words: Hazard, vulnerability, risk, GIS, earthquake scenario, tsunami.

1. Introduction 1.1. The Project ‘‘MAR VASTO’’ and its Partnership ‘‘MAR VASTO’’ (MAR VASTO, 2007; http:// www.marvasto.bologna.enea.it), funded by BID/ IADB (Banco InterAmericano de Desarrollo/InterAmerican Development Bank), has been coordinated

by ENEA (Italian Agency for New Technologies, Energy and the Environment), with the participation of several partners (Italy: Ferrara University, Departments of Architecture and Engineering; Padua University, Department of Structural and Transportation Engineering; Abdus Salam International Centre for Theoretical Physics/Trieste University; Chile: Technical University Federico Santa Maria of Valparaı´so, Civil Works Department; University of Chile in Santiago, Division Structures Constructions Geotechnics), and support of local stakeholders.

Valparaı´so being included since 2003 in the UNESCO Word Heritage List of protected sites, the project’s main goals have been the following: to collect and elaborate existing information and articulate a satisfactory evaluation of main hazards; to

Author's personal copy 544

M. Indirli et al.

develop a geographic information system (GIS) digital archive that is user-friendly and easily implemented, including hazard maps and scenarios; to provide a vulnerability analysis for three historical churches (La Matrı´z, San Francisco del Baro´n, Las Hermanas de la Divina Providencia, made of various materials—masonry, concrete, wood and adobe—and located in different city sites) and for a building stock in the Cerro Cordillera (partially inside the UNESCO area); and to suggest guidelines for future urban planning and strengthening interventions.

Pure Appl. Geophys.

Valparaı´so Italian Community. Also important was the contribution of Geocom Santiago, providing the survey laser-scanner equipment. Moreover, qualified professionals of the Valparaı´so Municipality have been provided short fellowships in Italy, funded by the Italian Latin American Institute (IILA, Istituto Italo Latino Americano).

2. The Project 2.1. A Brief Description of Valparaı´so

1.2. The Support of Local Institutions During the work, many Chilean Organizations cooperated with the Italian team: above all, the Valparaı´so Municipality, providing logistic and technical support; the Regional Authority (Intendencia V Region Valparaı´so); the Church (Archbishop of Valparaı´so and Franciscan Order of Friars Minors); the Civil Defense (OREMI); the Chilean Navy Hydrographic and Oceanographic Service (Servicio Hidrogra´fico y Oceanogra´fico de la Armada de Chile, SHOA); the PRDUV (Programa de Recuperacio´n y Desarrollo Urbano de Valparaı´so); the Firemen (Bomberos) and the Sea Rescue (Bote Salvavidas) Corps of Valparaı´so; the Valparaı´so Board of Architects, several professionals and other universities; the Police (Carabineros de Chile); and the

The Valparaı´so Bay was reached by the Spanish conquerors in 1536, who first settled in the ancient nucleus of the ‘‘Puerto’’, and then expanded into the Almendral area (Fig. 1). The city represents a distinctive case of growth, inside a remarkable landscape, of an important Pacific Ocean seaport (over the XIX-XX centuries). It achieved strategic importance in shipping trade, but declined after the Panama Canal opening (1914). Valparaı´so tells the never-ending story of a tight interaction between society and environment, stratifying different urban and architectonic layers, sometimes struck by disasters and always in hazardous conditions.

Valparaı´so’s morphology can be roughly divided into two main sectors: a flat harbour area (growing on lands reclaimed over centuries, see Fig. 2) and the

Figure 1 Valparaı´so: origin (left) and present situation (right)

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

545

Figure 2 Growth of Valparaı´so on reclaimed lands

Figure 3 Valparaı´so: the flat area and the harbor

hill quarters. Large neoclassic masonry buildings, some previous colonial style constructions (some of which are still standing, spared by earthquakes and following fires) and more recent architecture occupy the commercial district, with straight streets, highways and rail tracks parallel to the coast. A wide area is occupied by the port facilities up to the waterfront.

Otherwise, the steep forty-nine hills, cut by ravines (quebradas) and climbed by narrow and snaky lanes, are deeply populated by small and squat houses, typically constructed of wooden frames, adobe panels and covered by zinc tinplate (calamina). In addition to these pervading clustered homes, notable historical buildings are also present (Figs. 3, 4, 5, 6). In fact,

Valparaı´so shows an irregular urban tissue and its building inventory is very inhomogeneous. Therefore, it is possible to say that Valparaı´so is a city ‘‘with and without architects’’ (Fig. 7), in which the work of anonymous citizens has accumulated, together with the construction of remarkably designed buildings, sometimes copies of European ones but built with different materials.

Several old cable cars (ascensores) ascend the slope (Fig. 8). The historic district (Barrio Puerto, protected by UNESCO) lies in Southern Valparaı´so and embraces a sector which, starting from the flat, reaches the hills (red line, Fig. 9). Furthermore, the Valparaı´so Municipality declared all the city lying

Author's personal copy 546

M. Indirli et al.

Pure Appl. Geophys.

Figure 4 Valparaı´so: buildings in the flat area

Figure 5 Valparaı´so: the hills quarters

within the hills amphitheatre to be a protected area (green line, Fig. 9). Certainly, the city is subjected to various natural hazards (seismic events, but also tsunamis, landslides, etc.) and anthropic calamities (mainly wild and human-induced fires). These features make Valparaı´so a paradigmatic study of hazard mitigation, and risk factors must be very well evaluated during future restoration phases.

2.2. The Project Architecture ‘‘MAR VASTO’’ can be summarized as shown by Fig. 10: horizontal lines give the ‘‘general purpose’’ activities, while targeted investigations are reported in the two columns.

As it would be impossible to manage deep investigations for the entire Valparaı´so historic area

(due to limited resources in funds and time), a common decision with Chilean partners and stakeholders has been taken on structures/areas to be investigated, with the highest priorities identified as: – a building stock in the Cerro Cordillera (partially included in the UNESCO zone); – three important historical churches (‘‘La Matriz’’, ‘‘San Francisco del Baro´n’’, ‘‘Las Hermanas de la Divina Providencia’’), made of different materials and located in different city sites (Fig. 11).

3. The GIS Database The first ‘‘general purpose’’ activity was the organization of a Geographic Information System

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

547

Figure 6 Valparaı´so: buildings in the hills

Figure 7 Valparaı´so: a city ‘‘with and without architects’’

Figure 8 Examples of Valparaı´so cable cars

(GIS) geo-referenced database encompassing all of Valparaı´so, building (at ENEA) a detailed Digital Elevation Model (DEM) of the Valparaı´so area

by generating ortho-photos from the very helpful aerial photos provided by SHOA. Digital cartography (streets, buildings, quoted points, and other

Author's personal copy 548

M. Indirli et al.

Pure Appl. Geophys.

Figure 9 Valparaı´so: hazards and safeguarded areas

Figure 10 Brief description of the ‘‘MAR VASTO’’ Project

information) provided by the Valparaı´so Municipality, often not very accurate, did not match the aerial photos of the Valparaı´so area. Therefore, a field

survey using Differential Global Positioning System (DGPS) was been carried out in situ (a pattern of 33 points) in order to check aerial photos and

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

549

Figure 11 Location of the selected churches in Valparaı´so

cartography, verifying the GIS database from the topographic point of view, removing uncertainties, and clarifying unequivocally the real geographic positions. The GIS platform organized in clear and user-friendly maps a huge amount of data of general interest, including aerial and satellite photos; cartography and topo-bathymetry; GIS urban layers such as buildings, open spaces and viability; geo-referenced historic maps, etc. (see examples in Fig. 12).

The GIS also includes information targeted on specific hazards and the building inventory of the Cerro Cordillera pilot study sector (MAR VASTO, 2007; INDIRLI, 2009).

All the details are in DGPS survey (2008) and GIS database, (2008) (http://www.marvasto.bologna. enea.it).

4. Hazard Maps 4.1. Introduction Hazard maps have been developed for natural (earthquake, tsunami, landslide) and anthropic (fire) disasters and then stored in the GIS database (MAR VASTO, 2007; INDIRLI, 2009). In this article more

attention is given to earthquake and tsunami hazards, while the others are briefly described.

4.2. Seismic Hazard 4.2.1 General Information Chile is one of the most earthquake-prone countries in the world; it was struck by the most powerful seismic event ever recorded (1960 Valdivia earthquake and tsunami). Valparaı´so has been hit by other major earthquakes (Table 1). In particular, the 1906 event was the most destructive; the damages (classified using 1906 pictures) were concentrated mostly in the ‘‘El Almendral’’ neighborhood of the Valparaı´so harbor (Fig. 13a). Chilean partners have provided microzonation studies (performed in order to identify local soils effects), new evaluation of earthquake intensities, isoseismal maps in the damaged area. Three 1906 centennial undamaged surviving buildings also survived 1985 Chile earthquake (see Fig. 13b; ASTROZA, 2007; SARAGONI, 2007; STURM, 2008).

Specific studies on seismic hazard have been carried out. It is worth noting that the neo-deterministic approach (PANZA et al., 2001) has been followed

Author's personal copy 550

M. Indirli et al.

Pure Appl. Geophys.

Figure 12 Organization of the GIS database for Valparaı´so

in the ‘‘MAR VASTO’’ Project, in order to evaluate the seismic input in the Valparaı´so area for certain earthquake scenarios (in general), and in some sections underneath the churches locations (in particular). A complete description of the methodology, from the definition of the hazard to the seismic input calculation for the design of a building, is given in ZUCCOLO et al. (2008). Case studies indicate the limits of probabilistic seismic hazard analysis (PSHA), the currently used methodology, supplying indications that it can be useful but not sufficiently reliable (DECANINI et al., 2001; INDIRLI et al., 2006; KLU¨GEL et al., 2006). Though deeply rooted in engineering practice, PSHA has been called into question by recent examples (the earthquakes of: Michoacan 1985, Kobe 1995, Bhuj 2001, Boumerdes 2003, Bam 2003 and E-Sichuan 2008 events), that acted as catalysts for the use of zoning in seismic risk management. A drastic change is required in the orientation of zoning, which must be a pre-disaster

Table 1 Strong earthquakes striking Valparaı´so (e.g. LOMNITZ, 1971; 1983) Date Year

Month

Day

1647 1730 1822 1906 1965 1971 1985

05 07 11 08 03 07 03

13 08 19 17 28 09 03

Location

M

Valparaı´so, Chile Valparaı´so, Chile Valparaı´so, Chile Valparaı´so, Chile La Ligua, North Valparaı´so, Chile Valparaı´so Region, Chile Offshore Valparaı´so, Chile

8.50 8.75 8.50 8.20 7.10 7.50 7.80

activity performed to mitigate the effects of the next earthquake using all available technologies. Seismic zoning can use scientific data banks, integrated in an expert system, by means of which it is possible not only to identify the safest and most suitable areas for urban development, but also to define the seismic input that is going to affect a given building.

personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

551

Figure 13 The Valparaı´so 1906 earthquake

Seismic hazard assessment, necessary to design earthquake-resistant structures, can be performed in various ways, following a probabilistic or a deterministic approach. National seismic codes and zonations are often based on seismic hazard assessments computed with PSHA (CORNELL, 1968; SSHAC, 1997; GSHAP; TANNER and SHEDLOCK,

2004). An example for Chile using the PSHA approach is shown by Fig. 14.

Nevertheless, the PSHA cannot be sufficiently reliable to completely characterize the seismic hazard, because of the difficulty in defining the seismogenic zones and in correctly evaluating the occurrence of earthquakes (frequency–magnitude relations) and the

personal copy 552

M. Indirli et al.

Pure Appl. Geophys.

structures. Modelling can be done at different levels of detail, depending on the available knowledge of geological, geophysical, seismological and seismotectonical setting. The realistic modelling of the ground motion is a very important base of knowledge for the preparation of groundshaking scenarios that represent a valid and economic tool to be fruitfully used by civil engineers, supplying a particularly powerful tool for the prevention aspects of Civil Defense (pre-disaster microzonation).

Where the numerical modelling is successfully compared with records (as in the case of the Valparaı´so 1985 earthquake), the synthetic seismograms permit the generation of groundshaking maps, based upon a set of possible scenario earthquakes.

Figure 14 Seismic hazard map (PGA in m/s2 with 10% probability of exceedence in 50 years) of Chile using the probabilistic approach (see http://www.seismo.ethz.ch/gshap/)

propagation of their effects (attenuation laws). A more adequate description of the seismic ground motion can be achieved following a neo-deterministic approach, which allows for a realistic description of the seismic ground motion due to an earthquake of given distance and magnitude (PANZA et al., 2001). This approach, which can be feasibly applied at urban scales, is based on modelling techniques that have been developed from the knowledge of the seismic source generation and propagation processes. It is very useful because it permits engineers to define a set of earthquake scenarios and to compute the associated synthetic signals, without having to wait for a strong event to occur.

• Synthetic signals can be produced in a short time and at a very low cost/benefit ratio, and can be used as seismic input in subsequent engineering analysis aimed at the computation of the seismic response of

Concerning the seismic input, the major goal of the ‘‘MAR VASTO’’ Project has been to provide a dataset of synthetic time series representative of the potential ground motion at the bedrock of Valparaı´so, especially at selected sites (e.g. the three important churches located in the Valparaı´so urban area: La Matriz, San Francisco, Las Hermanas de la Divina Providencia), for four scenarios, taking into account two fault rupture typologies (unilateral and bilateral) in the urban Valparaı´so area, whose magnitudes are reported in Table 2. The occurrence periods and risk levels in Table 2 are intended solely for an engineering analysis and not in the sense of a return period (for this purpose see e.g. COMTE et al., 1986).

The characteristics of the calculated signals (e.g. amplitude, frequency content and duration of shaking) are determined by the earthquake source process and the wave propagation effects of the path between the source and the site. The generation and selection of realistic time series for design is essentially a problem of choosing appropriately from among a number of Table 2 Earthquakes scenarios for Valparaı´so M

Occurrence period

Tm & 120– 140 years Tm & 200– 250 years 8.3 1906 event Rare Tm & 500 years 8.5 Scenario Exceptional Tm & 1,000 years event

7.5 Scenario Occasional event 7.8 1985 event Sporadic

Strong Very strong Disastrous Catastrophic

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

future earthquake scenarios, whose most important characteristics are their magnitude and the distance to the site. From hazard maps, it is possible to define the scenario earthquakes to be used in planning exercises and earthquake engineering studies. Such an analysis is accomplished by hazard deaggregation, in which the contributions of individual earthquakes to the total seismic hazard, their probability of occurrence and the severity of the ground motions are ranked in the order.

Using the individual components (‘‘deaggregating’’ the events driving the hazard at the target region) of these hazard maps, the user can properly select the appropriate scenarios given their location, regional extent, and specific planning requirements. Thus, the definition of earthquake scenarios depends on many factors (e.g. historical seismicity, seismotectonic studies, engineering considerations) and is not intended as an earthquake prediction. That is, no one knows in advance when or how large a future earthquake will be, but making assumptions about the size and location of a hypothetical future earthquake, one can make a reasonable prediction of the effects (e.g. groundshaking) for planning and preparedness purposes.

Intermediate-term middle-range earthquake prediction is possible at different scales (PERESAN et al., 2005).

One of the most difficult tasks in earthquake scenario modeling is the treatment of uncertainties, since each of the key parameters has an uncertainty and natural variability, which often are not quantified explicitly. A possible way to handle this problem is to vary the modeling parameters systematically. Actually, a severe underestimation of the hazard could come by fixing a priori some source characteristics and thus the parametric study should take into account the effects of the various focal mechanism parameters (i.e. strike, dip, rake, depth, etc.). The analysis of the parametric studies will allow the researcher to generate advanced groundshaking scenarios for the proper evaluation of the site-specific seismic hazard, with a complementary check based on both probabilistic and empirical procedures. Once the gross features of the seismic hazard are defined and the parametric analyses have been performed, a more detailed modelling of the ground motion can be carried out for sites of specific interest. Such a detailed analysis should take into account the source characteristics, the path and the local geological and

553

geotechnical conditions. This neo-deterministic modelling goes well beyond the conventional deterministic approach taken in hazard analyses—in which only a simple wave attenuation relation is invoked— in that it includes full waveform modelling.

4.2.2 Seismic Hazard at Regional Scale Neo-deterministic seismic zoning (PANZA et al., 1996; PANZA et al., 2001) is one of the newest and most advanced approaches; it has been applied successfully to many areas worldwide (e.g. PARVEZ et al., 2003). It can be used as a starting point for the development of an integrated approach that combines the advantages of the probabilistic and deterministic methods, thus minimizing their respective drawbacks.

This approach addresses some issues largely neglected in PSHA, namely how crustal properties affect attenuation: ground motion parameters are not derived from overly simplified attenuation functions, but rather from synthetic time histories. Starting from the available information on the Earth’s structure, seismic sources, and the level of seismicity of the investigated area, it is possible to estimate peak ground acceleration, velocity, and displacement (PGA, PGV, and PGD) or any other parameter relevant to seismic engineering, which can be extracted from the computed theoretical signals. This procedure allows us to obtain a realistic estimate of the seismic hazard in those areas for which scarce (or no) historical or instrumental information is available and to perform the relevant parametric analyses.

Synthetic seismograms can be constructed to model ground motion at sites of interest, using knowledge of the physical process of earthquake generation and wave propagation in realistic media.

The signals are efficiently generated by the modal summation technique (PANZA and SUHADOLC, 1987; PANZA et al., 2001), so it becomes possible to perform detailed parametric analyses at reasonable costs. The flowchart of the procedure is shown in Fig. 15. The first problem to tackle in the definition of seismic sources is the handling of seismicity data. Basically, what is needed is an evenly spaced distribution of the maximum magnitude over the territory, but the data available from earthquake catalogues are widely scattered. Furthermore, earthquake catalogues are

Author's personal copy 554

M. Indirli et al.

both incomplete and affected by errors, so a smoothed distribution is preferable (PANZA et al., 2001).

Specific seismograms have been computed at the nodes of a grid, with step of 0.2 , that covers the region of the Central Chile territory. 107 events have been selected from the dataset collected by the Servicio Sismologico Universidad de Chile (http:// ssn.dgf.uchile.cl/home/terrem.html), that have been considered as the most representative between the destructive earthquakes occurred in Chile, from 1570 till present, with a magnitude greater than 7. Four different seismogenic zones have been defined using the information about the seismicity and tectonics of the area. The seismicity is discretized into 0.2 9 0.2 cells, assigning to each cell the maximum magnitude recorded within it; a smoothing procedure is then applied to account for spatial uncertainty and for source dimensions. Since the aim of this step is the average regional definition of seismic input, a double-couple point source was chosen, with a focal mechanism consistent with the regime of the pertinent seismogenic zone (PANZA et al., 2001), as shown in Fig. 16.

REGIONAL POLYGONS

FOCAL MECHANISMS

STRUCTURAL MODELS

SEISMOGENIC ZONES

EARTHQUAKE CATALOGUE

SEISMIC SOURCES

SITES ASSOCIATED WITH EACH SOURCE

TIME SERIES PARAMETERS

P-SV SYNTHETIC SEISMOGRAMS

VERTICAL COMPONENT

SH SYNTHETIC SEISMOGRAMS

HORIZONTAL COMPONENTS

EXTRACTION OF SIGNIFICANT GROUND MOTION PARAMETERS

Figure 15 Flow-chart of the neo-deterministic procedure for seismic hazard assessment at regional scale (see Panza et al., 2001). The vertical component is routinely not used

Pure Appl. Geophys.

To define the physical properties of the source-site paths, the territory is characterized by a structural model composed of flat, parallel anelastic layers that represent the average lithosphere properties at a regional scale. We have defined the regional bedrock structural model (the elastic and anelastic parameters of the uppermost layers are shown in Table 3) starting from the model proposed by MENDOZA et al.

(1994) and the related references.

The seismograms have been computed for an upper frequency content of 10 Hz and the point sources were scaled for their dimensions (Size Scaled Point Source) using the relatively simple spectral scaling laws by GUSEV (1983). Such a simple source model gives a reliable upper bound of the PGA and, at the same time, permits a realistic estimate of the PGD and PGV. In Figs. 17, 18, 19 and 20 only some examples of the wide set of results are shown. Next, we focused (deaggregating the hazard) on the two most important earthquake scenarios for Valparaı´so: the 1985 and 1906 events, belonging to seismogenic zone 4. These events were chosen not only because of their large magnitude, but also because of the damage that they generated in Valparaı´so (Saragoni, 2006).

We used the simple model of source (SSPS) to define the upper bound of PGA for the entire study region; it turns out to be about 1.2 or 0.5 g (after deaggregation), in good agreement with the Intensities values reported for the 1906 event (Saragoni, 2006; ASTROZA, et al., 2006). The use of microzonation, considering more complex and realistic sources, allowed us to better delimit the zones at an urban scale where the largest PGA can be expected.

Therefore, our procedure consisted of: (a) expeditious and low-cost determination of hazard using simple and easily available information about seismic sources and their surrounding medium (this section), (b) more realistic modeling (next sections) focused on items of special interest for which more expensive procedures are necessary and fully justified by the exposed value.

4.2.3 Extended Source Models A further step towards realism, would be to consider the rupture process at the source and the related directivity effect (i.e. the dependence of the radiation

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

Zone 1 Zone 2 Zone 3 Zone 4

date 16/10/1981 28/03/1965 06/07/1979 16/08/1906

mag 7.2 7.3 6.0 8.2

strike 345 350 11 3

555

dip 86 80 54 15

rake -93 -100 105 117

Figure 16 Seismogenic zones and their representative events Table 3 Elastic and anelastic parameters of the uppermost layers of the bedrock structural model Thickness (km)

Density (g/cm3)

P-wave velocities, Vp (km/s)

S-wave velocities, Vs (km/s)

P-waves quality factor, Qp

S-waves quality factor, Qs

1.0000 3.0000 4.0000 8.9000 5.5000 15.000 30.000

1.70 2.00 2.30 2.50 2.65 2.80 3.28

4.000000 4.750000 5.560000 6.070000 6.530000 7.000000 8.000000

2.310000 2.740000 3.210000 3.500000 3.770000 4.040000 4.500000

100.00 200.00 400.00 500.00 600.00 600.00 600.00

50.00 100.00 200.00 250.00 300.00 300.00 300.00

at a site on its azimuth with respect to the rupture propagation direction). To this end, extended source models have been considered, using the algorithm for the simulation of the source radiation from a fault of finite dimensions, named pulse-based wideband synthesis (PULSYN), developed by GUSEV and PAVLOV (2006). The seismic waves due to an extended source are obtained by approximating it with a rectangular plane surface, corresponding to the fault plane on which the main rupture process is assumed to occur.

Effects of directivity and of the energy release on the fault can be easily modeled, simulating the wide-band

radiation process from a finite earthquake source/ fault. To represent an extended source, PULSYN uses the main features of the HASKELL (1964) model and discretizes the rectangular fault plane with a grid of point sub-sources. The arrival of the rupture front at a subsource switches its slip. However, unlike Haskell model, the spatial distribution of slip and the rupture velocity are treated as a random process and characterized in a stochastic manner. In fact, the small-scale details of the rupture process, connected to heterogeneities in the stress distribution, though, in general, too complicated to be

Author's personal copy 556

M. Indirli et al.

Pure Appl. Geophys.

Figure 17 Horizontal PGD distribution and Period in seconds of its maximum

exactly characterized, generate high frequency waves when reached by the rupture front. In this way the code PULSYN generates a source (phase and amplitude) spectrum, which is close in amplitude to GUSEV’S (1983) empirical curves and reproduces the directivity effects (see Fig. 21) as in the theoretical Haskell model.

With this approach we can simulate the time histories using (1) Extended Source (ES) and (2) Space and Time Scaled Point Source (STSPS) methods. In the ES case, the source is represented as a grid of point subsources, and their seismic moment rate functions are generated, considering each of them as realizations (sample functions) of a non-stationary random process. Specifying in a realistic way the source length and width, as well as the rupture velocity, one can obtain realistic source time functions, valid in the far-field approximation.

Finally, to calculate the ground motion at a site, Green functions are computed with the highly efficient and accurate modal summation technique, for each subsource–site pair, and then convolved with

the subsource time functions and, at last, summed over all subsources. Furthermore, assuming a realistic kinematic description of the rupture process, the stochastic structure of the accelerograms can be reproduced, including the general envelope shape and peak factors. The extended seismic source model allows us to generate a spectrum (amplitude and phase) of the source time function that takes into accounts both the rupture process and directivity effects, also in the Near Source region. In the second case (STSPS), we use a mixture of extended and point sources. We sum up the source time functions generated by the distributed (point) subsources in order to obtain the equivalent single source, representative of the entire space and time structure of the extended source, and the related Green Function. In this way it is possible to perform expeditious parametric studies useful for engineering analysis (ZUCCOLO et al., 2008) to investigate the dependence of the ground motion (in the time and frequency domain) on source parameters (geometry, energy release, etc.).

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

557

Figure 18 Horizontal PGV distribution and Period in seconds of its maximum

4.2.4 Parametric Tests The synthetic signals database has been greatly expanded by performing a parametric study of ground motion, taking into account variations due to the choice of focal mechanism parameters. Varying the geometry of the seismic source, different ground motions at the Valparaı´so site have been studied in order to consider maximum excitation in both longitudinal (P-SV motion) and transverse (SH motion) direction, and in order to consider (starting from the Maximum Historical Earthquake) both the Maximum Credible Earthquake and the Maximum Design Earthquake.

Computations of synthetic seismograms (displacements, velocities and accelerations for the radial, transverse and vertical components) were carried out with a cut-off frequency of 10 Hz. All the focal mechanism parameters of the original source models obtained from the seismic catalogues were varied in order to find the source mechanism which produces the maximum amplitudes of the

various ground motion components. The preliminary parametric test was performed to estimate the dependence of the radiation pattern on the orientation of the fault plane. This analysis allows for limited changes in the assumed strike-receiver angle (which is uniquely determined by the strike of the fault and the coordinates of the epicentre and the receiver) to avoid the case when one of the three components of motion corresponds to a minimum of radiation. The conventions adopted are explained in Fig. 22.

Firstly, the hypothetical receiver was taken to be at an epicentral distance of 100 km; two focal mechanisms have been adopted as starting models: the one proposed by CHOY and DEWEY (1988), with strike of 360 , dip 35 , rake 105 and the result given by the CMT Catalogue of Harvard (http://www. globalcmt.org/).

Then, the hypothetical receiver was taken to be at an epicentral distance of 60 km; one focal mechanism has been adopted as starting model, given by the CMT Catalogue of Harvard (http://www.globalcmt. org/), but with depth 25 km.

personal copy 558

M. Indirli et al.

Pure Appl. Geophys.

Figure 19 Horizontal PGA distribution and Period in seconds of its maximum

To simulate the 1906 event at Valparaı´so, the receiver was taken to be in the Valparaı´so urban area, at an epicentral distance of about 48 km. The results of these parametric studies allowed us to fix some of the focal mechanism parameters for the computation of synthetic signals along detailed profiles for this scenario earthquake: the same values of dip and rake (respectively 15 and 105 ) were adopted; but, for the strike receiver, we used values of 230 and 300 , respectively, for the transverse and radial components of motion, corresponding to the maximum ground motion amplitude.

To simulate the 1985 event at Valparaı´so, the receiver was taken to be in the Valparaı´so urban area, at an epicentral distance of about 30 km. The results of these parametric studies allowed us to fix some of the focal mechanism parameters for the computation of synthetic signals along detailed profiles for this scenario earthquake.

Details of this procedure are given in Earthquake hazard (2008) (http://www.marvasto.bologna.enea.it).

Quantitative validation of the neo-deterministic results was made using the only available observed

signals in Valparaı´so urban zone, i.e., the ones recorded during the 1985 earthquake at the El Almendral (ALMEN) and Universidad Santa Maria (UTFSM) recording stations (SARAGONI, 2006). Figure 23 shows an example of extended source model, with the geographical reference. Many investigators have studied the source parameters of the Chilean earthquake of March 3, 1985 (BERESNEV and ATKINSON, 1997; SOMERVILLE et al., 1991); in this study, we have chosen the one proposed by CHOY and DEWEY (1988), with a strike of 360 , a dip 35 , a rake of 105 and the result given by the CMT Catalogue of Harvard (http://www.globalcmt.org/).

4.2.5 Comparison between Computed Results and Recorded Signals Figure 24a shows the first ‘‘blind test’’ simulated seismograms (respectively displacement, velocity and acceleration) along the NS and EW components using the extended source models embedded in the bedrock structural model and calculated at the El Almendral station. The signals recorded at such a station, located

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

559

Figure 20 Horizontal PGA distribution and Period in seconds of its maximum after deaggregation

on a sedimentary cover, are shown in Fig. 24b and confirm that the extended source and bedrock models allow us to reproduce the observed amplitude and duration of the ground motion. The comparison between the simulated and the recorded signals and the related response spectra at the UTFSM station (located on bedrock) are shown in Fig. 25, and demonstrate excellent agreement. Table 4 summarizes the results of this ‘‘blind test’’, i.e., without any tuning process or validation procedure, confirming that the extended source and bedrock models are successfully validated for the computation of the seismic input. Of course, the visible differences reported in Table 4 between the observed and simulated maximum values of the ground motion (i.e. acceleration, velocity and displacement) would be reduced with a more targeted inversion procedure.

4.2.6 Seismic Input at Urban Scale The methodology explained above allowed us to generate a set of groundshaking scenarios at bedrock

in the urban area of Valparaı´so associated to different ‘‘scenario’’ earthquakes. The scenario earthquakes can be classified, according to their different: (a) magnitude, (b) occurrence period, Tm, and (c) risk level (see Table 2). These are intended solely for an engineering analysis and not in the sense of a return period (for this purpose see e.g. COMTE et al., 1986).

For every scenario two rupture styles (unilateral North to South and bilateral) have been considered and the synthetic signals (displacements, velocities and accelerations) for the two horizontal components of motion (N–S and E–W) have been computed at a dense grid (step of approximately 0.02 km) of sites in the Valparaı´so urban area. The peaks of the ground motion and their period of occurrence, have been extracted in a matrix form; these results have been GIS processed and graphically rendered in a set of 96 maps. The results (see an example in Fig. 26) are extremely important, since the information they carry can be mapped in terms of intensity scenarios, allowing comparison with the available measured intensities for the 1985 and 1906 scenarios. PANZA et al. (1997)

personal copy 560

M. Indirli et al.

Pure Appl. Geophys.

Figure 21 a Unilateral and bilateral rupture processes, b directivity effect and c an example of source spectra for a unilateral rupture at three different sites, in the case of forward (red), neutral (green) and reverse (blue)

have produced new relations between Intensity I and the peak values of acceleration, velocity and displacement, valid for the Italian territory (see Table 5). They used two different versions of the GNDT earthquake catalogue (NT3.1 and NT4.1.1) and two sets of observed intensity maps for the Italian territory (ING and ISG data) and exploited advanced modelling methods for seismic wave propagation (PANZA et al.,

2001). The results obtained for accelerations do not differ significantly from the earlier results of CANCANI (1904). From Tables 5, 6 and 7, it is evident that the generated groundshaking scenarios at bedrock match very well with the average intensities measured in Valparaı´so for the 1985 (VIII MSK) and the 1906 (IX MSK) events as reported by SARAGONI (2006) and Astroza et al. (2006).

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

561

Figure 22 Seismic source parameters and conventions adopted

4.2.7 Seismic Input along Selected Profiles: Site Response Estimation Coordinates of three sites collected during the in situ DGPS campaign, see DGPS survey, 2008 were selected as strategic targets for the whole project.

These were the locations of the three churches: La Matriz, San Francisco and Las Hermanitas de la Providencia (see Fig. 11) The full seismic input (acceleration time histories and related response spectra) at the bedrock was specifically computed for these three locations. As an example (Fig. 27) the time histories computed for the 1906 scenario (unilateral rupture) at the La Matriz church are shown; it is the only one located (see Fig. 28) on a bedrock site, while the other two churches are located near the El Almendral area characterized by the presence of a sedimentary basin. Thus, the computation of the seismic input at these sites can be affected by local soil amplifications (as discussed later).

To deal both with realistic source and structural models, including topographical features, a hybrid method has been developed that combines modal summation and the finite difference technique (FA¨H and PANZA, 1994), and optimizes the use of the advantages of both methods. Wave propagation is treated by means of the modal summation technique from the source to the vicinity of the local, heterogeneous structure that we want to model in detail. A laterally homogeneous anelastic structural model is adopted which represents the average crustal properties of the region. The generated wavefield is then introduced into the grid that defines the heterogeneous area and is propagated according to the finite differences scheme. With this approach, source, path and site effects are all taken into account, and it is therefore possible to conduct a detailed study of the wavefield that propagates even at large distances from the epicentre. This methodology has been successfully applied to many urban areas worlwide

Author's personal copy 562

M. Indirli et al.

Figure 23 Source and station geometry for the 1985 Valparaı´so, Chile, earthquake: map view showing horizontal projection of fault; numbers in fault elements represent slip in meters (modified from Somerville et al., 1991)

(PANZA et al., 1999; ZUCCOLO et al., 2008), and to strategic buildings and lifelines (ROMANELLI et al., 2003, 2004; VACCARI et al., 2005).

In the hybrid scheme (see Fig. 29), two local heterogeneous models have been coupled with the average regional model used in the initial analysis for the detailed modelling of earthquake ground motion by computing synthetic seismograms in laterally heterogeneous media. The two profiles, eachabout 2 km long and 0.2–0.4 km deep, have been selected according to the available information and to their representativeness of the most important areas of the municipality (see Fig. 28). To define the sub-surface topography of the sedimentary layers, we used (VERDUGO, 1995, 2006a, b) the information given by ESPINOZA (2000), elaborated for the bedrock model.

We then performed some parametric tests, i.e. (a) including (or not) some very superficial layers, (b) changing the values of elastic and anelastic parameters, (c) choosing different bedrock models and (d) adopting the different scenarios previously defined.

In the following we show the results related to the two profile models shown in Fig. 30, together with

Pure Appl. Geophys.

the values of the adopted elastic and anelastic parameters. The minimum S-wave velocity present in the models shown is 660 m/s, and the mesh used for the finite differences is defined with a grid spacing of 7 m. This allows us to carry out the computations at frequencies as high as about 10 Hz The synthetic time signals (displacements, velocities and accelerations) have been calculated for the three components of motion, adopting a STSPS seismic source model for the 1985 earthquake scenario.

Site effects are then evaluated as spectral amplifications, described by the ratios (2D/1D) of the acceleration response spectra, with 5% damping, computed along the bedrock model (1D) profile and along the one containing the local model (2D).

The amplification factors computed for the horizontal components (up to 4) explain very well the pattern of the measured intensities in the Valparaı´so urban area associated to the 1985 and the 1906 events, as reported for example by SARAGONI (2006) and ASTROZA et al. (2006). In fact, a general result of our investigation is that the local effects due to the thickening of the sedimentary basin (up to 300 m) in the El Almendral zone can cause an increment greater than 1 unit in the seismic intensity experienced with respect to the average intensity affecting the urban area as a whole (see Fig. 13).

More details are given in Earthquake hazard (2008) (http://www.marvasto.bologna.enea.it).

4.3. Tsunami Hazard 4.3.1 General Information A tsunami occurs after a huge mass of water is displaced by some force from its equilibrium configuration. Gravity acts as a restoring force, tending to bring the displaced mass of water back to its original equilibrium state. Most tsunamis are generated by submarine earthquakes, but possible sources are also inland/coastal earthquakes, landslides and meteoric impacts. Due to their generation mechanism, periods and wavelengths associated with tsunamis are longer than those associated with ordinary wind-driven sea waves; for large submarine earthquakes their amplitudes can be very impressive, especially when the waves approach the shorelines.

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valpara覺織so

563

Figure 24 El Almendral station: horizontal (N50E) component of acceleration, velocity and displacement for the 1985 event: a computed; b recorded (thanks to R. Saragoni and S. Ruiz)

Author's personal copy 564

M. Indirli et al.

Pure Appl. Geophys.

Figure 25 1985 event at UTFSM station: a Horizontal recorded accelerations; b simulated accelerations; c comparison of response spectra: this study, recorded, and the one simulated by SOMERVILLE et al., 1991 Table 4 Peak ground motion values for (blind) simulated and observed signals Receivers Components CHOY

ALMEN UTFSM

N50 E S40 E N70 E S20 E

CMT

OBS

Amax (g) Vmax (cm/s) Dmax (cm)

Amax (g) Vmax (cm/s) Dmax (cm) Amax (g) Vmax (cm/s) Dmax (cm)

0.1516 0.1640 0.1467 0.1399

0.128 0.122 0.112 0.139

23.05 14.07 20.49 14.07

9.16 6.10 7.55 4.54

12.07 15.23 11.38 18.94

6.22 5.32 5.79 5.68

0.2907 0.1621 0.1767 0.1625

28.59 16.89 14.70 6.40

5.37 2.81 3.26 1.33

CHOY CHOY and DEWEY (1988), CMT centroid moment tensor, http://www.globalcmt.org/, OBS Observed

The Chilean coast is currently exposed to the effects of tsunamis generated in the Pacific Ocean (GUTIERREZ, 2005), and Valparaı´so has been inundated several times in the past. For instance, catastrophic events of the nineteenth century, 1868 and 1877, overwhelmed the coast of the northern region of the country. During the twentieth century,

the most important disaster was the 1960 earthquake and tsunami in Valdivia, in the south of the country.

This event had a great impact on the coasts of most of the neighbouring countries in the Pacific Ocean, primarily in the Hawaiian Islands and Japan. The last important event recorded along the Chilean coast was the ‘‘good tsunami’’ which occurred in Antofagasta,

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

565

Figure 26 Groundshaking scenario at the bedrock level in the Valparaı´so urban area for the 1985 event. NS component of velocities for bilateral rupture Table 5 ING–NT4.1.1 (MCS) (horizontal components) Intensity

Displacement (cm)

Velocity (cm/s)

DGA (g)

V VI VII VIII IX X XI

0.1–0.5 0.5–1.0 1.0–2.0 2.0–3.5 3.5–7.0 7.0–15.0 15.0–30.0

0.5–1.0 1.0–2.0 2.0–4.0 4.0–8.0 8.0–15.0 15.0–30.0 30.0–60.0

0.005–0.01 0.01–0.02 0.02–0.04 0.04–0.08 0.08–0.15 0.15–0.30 0.30–0.60

DGA design ground acceleration Table 6 ISG–NT4.1.1 (MCS) (horizontal components) Intensity

Displacement (cm)

Velocity (cm/s)

DGA (g)

VI VII VIII IX X

1.0–1.5 1.5–3.0 3.0–6.0 6.0–13.0 13.0–26.0

1.0–2.0 2.0–5.0 5.0–11.0 11.0–25.0 25.0–56.0

0.01–0.025 0.025–0.05 0.05–0.1 0.1–0.2 0.2–0.4

DGA design ground acceleration

1995. This historic situation has contributed to an awareness of the risk involved, and therefore to the development of research on the subject in Chile. The organization in charge of detecting and issuing the warning is the Hydrographic and Oceanographic Service of the Chilean Navy (SHOA; http://www. shoa.cl). The tsunami warning head office is located in the Department of Oceanography of SHOA. In the last few years, new developments in technology have made it possible to improve the quality of information used in assessing the potential risk of a tsunami event off the Chilean coast. Since 1995, a TREMORS System has been operating in Chile. This is system comprises seismic monitoring equipment that improves the existing seismic network and tsunami warning system in Chile, giving information in real time of seismic parameters and their relationship with some of the parameters of tsunami generation in order to estimate risk. A good example of the application and utility of the technology was the tsunami warning issued by SHOA for the 1996 Chimbote earthquake in Peru.

Author's personal copy 566

M. Indirli et al.

Pure Appl. Geophys.

Table 7 Comparison of seismic intensity scales

MM I II III IV

RF

JMA

MCS

MSK

II

I

III

II

IV

III

II

V

IV

III

VI

V

IV

VII

VI

I II

I

III IV V

V VI VI VII

VII

VIII VIII

V

VIII IX

VIII X

IX VI

X XI

VII

IX

XI XII

X

IX X XI

VII XII

XII

Adapted from Dolce et al. (2005) MM Modified Mercalli, RF Rossi-Forel, JMA Japanese Meteorological Agency, MCS Mercalli–Cancani–Sieberg, MSK Medvedev–Sponheuer–Karnik

As a very important complement to the operative work, SHOA has been working actively in the processing of inundation maps by tsunamis for the Chilean coast using the TIME project technology. Since 1996, after the TIME training course in Chile, the National Tsunami Warning System has been producing inundation charts of the main ports to help the Civil and Maritime Authorities to plan for and mitigate the effects of a tsunami. During the period 1997–2004, twenty-eight charts were produced under the project ‘‘Processing of Inundation Maps by Tsunamis for the Chilean Coast’’.

The cities included in these charts (http://www. shoa.cl/servicios/citsu/citsu.php) are: Arica, Iquique, Tocopilla, Mejillones, Antofagasta, Taltal, Caldera, Chan˜aral, Huasco, Coquimbo, La Serena, Los Vilos, Papudo, Quintero, Valparaı´so, Vin˜a del Mar, Algarrobo, San Antonio, Constitucio´n, Talcahuano, Penco, Lirque´n, Tome´, San Vicente, Coronel, Lebu, Corral y Ancud. Inundation maps have being used for tsunami

hazard planning by the national civil protection agency (ONEMI) and other government institutions.

The SHOA report (SHOA, 1999), discussed by the Italian team at the SHOA headquarters during the in situ visit, should be used as the reference document for the tsunami hazard assessment for the Valparaı´so site. The objective of this work is to complement the deep and detailed studies already carried out by SHOA, with (a) sets of parametric studies about the tsunamigenic potential of the 1985 and 1906 scenario earthquakes; (b) analytical modelling of tsunami waveforms for different scenarios, in order to provide a complementary dataset to be used for the tsunami hazard assessment at Valparaı´so.

4.3.2 Tsunami Simulation: Theory and Modelling The traditional approach to model tsunami generation is based on solving hydrodynamic equations

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

567

Figure 27 Example of seismic input computed at the La Matriz church: 1906 scenario, unilateral rupture. Displacements, velocities and accelerations for the two horizontal (North–South, NS, and East–West, EW) components of motion

with boundary conditions at the ocean floor corresponding to a static displacement caused by the earthquake source (HAMMACK, 1973; LEE and CHANG, 1980; OKAL, 1982; COMER, 1984a, b). Another welldeveloped approach is based on the modal theory (POD’YAPOLSKY, 1968; WARD, 1980; COMER, 1984a, b; PANZA et al., 2000). The former approach assumes the ocean and solid Earth to be partially coupled, whereas according to the latter they are fully coupled. Though the modal theory gives a solution corresponding to the exact boundary conditions, and it may be easily extended to models with slightly varying thickness of the water layer, it can be applied only when a source is located under the ocean. However, there are indications that sources near a coastline, and even inland, may cause intense tsunami waves. For the analysis of such a case a suitable approach may be that based on the Green’s function technique, as proposed firstly by KAJIURA (1963) for the analysis of tsunamis excited by an impulsive source.

4.3.3 Modal Summation Technique: Tsunamis Generated by Offshore Earthquakes The approach we use here for modelling tsunamis generated by offshore earthquakes is an extension (PANZA et al., 2000) to the case of tsunami propagation, of the well-known modal theory (POD’YAPOLSKY, 1968; WARD, 1980; COMER, 1984a, b) and therefore we simply refer to it as ‘‘modal method’’. In this approach it is assumed that the ocean and the solid Earth are fully coupled. From the mathematical point of view, in the modal approach the equations of motion are solved for a multi-layered model structure (according to HASKELL, 1964), so the set of equations is converted into a matrix problem in which to look for eigenvalues and eigenfunctions. In general, the modal theory gives a solution corresponding to the exact boundary conditions, and so it is easily extended to models with slightly varying thickness of the water layer. Therefore, the modal method allows us to calculate synthetic signals for both

Author's personal copy 568

M. Indirli et al.

Pure Appl. Geophys.

Figure 28 Bedrock model (depth) at El Almendral (VERDUGO, 1995; SARAGONI, 2006) and the position of the two proďŹ les

Distance from the source

Free surface

A Depth

Source

Reference layered model

Artificial boundaries, limiting the FD grid.

A

Zone of high attenuation, where Q is decreasing toward the artificial boundary.

Local heterogeneous model

Adjacent grid lines, where the wave field is introduced into the FD grid. The incoming wave field is computed with the mode summation technique. The two grid lines are transparent for backscattered waves (Alterman and Karal, 1968).

Site

Figure 29 Scheme of the hybrid (modal summation plus ďŹ nite differences scheme) method (e.g. PANZA et al., 2001)

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

569

Figure 30 Local profiles (top: red line of Fig. 28; bottom: blue line of Fig. 28) with their elastic and anelastic parameters

laterally homogeneous (1D) and laterally heterogeneous (2D) structures. For the 2D case, the structural model is parameterized by a number of 1D structures arranged in series along the profile from the source to the receiving site. The liquid layer is considered to be homogeneous and incompressible, no vertical stratification of the water is considered. The parameterization of the bathymetry is important for the longer source-site paths, since it can strongly influence travel times. In our calculations the number of model structures varies from 2 to 14, depending mainly on the number of slope-trending variations along each path. It is a useful rule to keep the parameterization as simple as possible. The modal method has a major limitation: due to its intrinsic mathematical formulation, it can be applied only when a source is located under the ocean (i.e. is applicable only to the offshore source case).

4.3.4 Green’s Function Approach: Tsunamis Generated by Inland/Coastal Earthquakes There are several indications that sources near, or even inside, a coastline may cause intense tsunami waves. For the analysis of such cases, a suitable approach to compute synthetic mareograms has been

developed by Yanovskaya et al. (2003) with the Green’s function technique, which solves the problem of modelling tsunamis generated by inland/ coastal sources. This method uses the representation theorem together with the Green’s function as first proposed by KAJIURA (1963) for the calculation of tsunamis generated by an extended source under an infinite water layer of constant thickness. This case is then extended with the addition of a coastline, considering a semi-infinite water layer of constant thickness. The exact solution for the Green’s function in the liquid layer is represented in an integral form, and therefore, to solve the problem, it is necessary to adopt an approximation. The approximation adopted is the well-known asymptotic representation of the integral solution by Hankel’s functions, which allows calculation only for the far-field case.

A rough evaluation, in the case of tsunamis in a shallow water domain, fixes the lower limit for source-site distances that can be considered in this approximation at about ten kilometers.

4.3.5 Wave Propagation Since we use two-dimensional and one-dimensional models, we can compute mareograms only along

Author's personal copy 570

M. Indirli et al.

straight segments from the source to the receiver sites, neglecting all three-dimensional effects, such as refraction and diffraction; this is a limitation of our method. When analyzing the results one has to take into consideration that variations of the sea depth can cause refraction and thus focusing or de-focusing of the wave in some regions. Diffraction of the wave front may also play a significant role in the presence of obstacles such as an island or a peninsula.

Moreover, a number of local effects can generally occur in proximity to the coast due to the thinning of the liquid layer, strongly influencing both travel time and maximum amplitude. The ensamble of this phenomena is often called shoaling and is responsible for the final tsunami run-up. The major contribution is the amplification of the wave approaching the coast due to the progressive thinning of the water layer.

The principle of conservation of energy requires that the wave energy, when the tsunami reaches shallow waters, is redistributed into a smaller volume, which results in a growth of the maximum amplitude.For the shoaling amplification factor linear theory gives a simple expression, known as Green’s law. Typically the shoaling factor ranges from 1 (no growth) up to several units (amplification) depending on the considered domain (WARD, 2002). Shoaling amplification acts until the wave amplitude is approximately less than half the sea depth (WARD and DAY, 2008), then nonlinear phenomena cause the waves to break and eventually turn them backward. WARD and DAY (2008) suggest that due to complications of wave refraction and interference, runup is best considered as a random process that can be characterized by its statistical properties. Models and observations hint that runup statistics follow a single skewed distribution spreading between 1/2 and 2 times its mean value. Another phenomenon contributing to the wave amplification is the overlapping of the signal, due to the fact that waves travel more slowly in shallow than in deep waters, so the front of the wave packet that first reaches shallow waters, is overtaken by the tail of the signal. This often results in a growth of the maximum amplitude.

When dealing with very long source-site distances (hundreds of kilometers), an additional effect on tsunami maximum amplitude becomes relevant due to the phenomenon of dispersion, i.e., the fact that the

Pure Appl. Geophys.

components at low frequency of the signal travel faster than the higher ones. After a certain distance the slower high-frequency components tend to migrate to the tail of the wavetrain where they no longer contribute to the main peak amplitude.

4.3.6 Hazard Scenarios The main purpose in modelling a hazard scenario is to assess the maximum threat expected from a studied phenomenon in a certain area and to give specific directives to the local authorities in order to prevent and mitigate serious consequences on the population, the infrastructures and the environment. By means of modelling, we have calculated the maximum amplitude of the vertical displacement of the water particles on the sea surface and the travel time of the maximum amplitude peak, since they are the most relevant aspects of the tsunami wave and also are the only characteristics always recorded in the chronicles and therefore in catalogues. The horizontal displacement field has been calculated, too, and, in average, it exceeds the vertical one by approximately an order of magnitude (this accounts for the great inundating power of tsunami waves with respect to wind-driven waves). To calculate tsunami hazard scenarios (see Table 8) we have first adopted the scenario events (1985 and 1906) and the source model described by SHOA (SHOA, 1999) and then we proceeded to model the tsunami for other possible scenarios.

It is important to mention that the extremely efficient analytical modelling techniques (computation times are of the order of seconds and are bound to decrease with the natural rate of improvement of computers) for real time simulations can be utilized also for a Tsunami Warning System, since they can be compared with real time incoming open-sea level data, in order to validate, or close, an impending alarm.

4.3.7 Parametric Studies The quick, accurate and efficient analytical modelling techniques are used to generate a preliminary dataset of synthetic mareograms, performing parametric studies to define the influence of the focal

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

571

Table 8 Tsunami scenarios for Valparaı´so M 7.0 7.5 7.8 8.3 8.5

Occurrence period Scenario event Scenario event 1985 event 1906 event Scenario event

Frequent Occasional Sporadic* Rare* Exceptional

Tm Tm Tm Tm Tm

& & & & &

70–80 years 120–140 years 200–250 years 500 years 1,000 years

Moderate/strong Strong Very strong Disastrous Catastrophic

* From SHOA source models and simulations

mechanism (strike, dip, rake and focal depth) on the tsunamigenic potential of the seismic sources associated to the scenario events. In particular, the modal technique is used for the most significant tsunami events with source located offshore (whose hypocentre is located under the sea bottom).

Varying the geometry of the seismic source, different tsunamis at the Valparaı´so site have been studied, in order to consider the maximum tsunamigenic excitation and in order to consider (starting from the Maximum Historical Earthquake) both the Maximum Credible Earthquake and the Maximum Design Earthquake. All the focal mechanism parameters of the original source models obtained from the seismic catalogues were varied in order to find the source mechanism producing the maximum amplitude of the tsunami. A preliminary parametric test was performed to estimate the dependence of the radiation pattern on the orientation of the fault plane (see Fig. 22 for the convention adopted in the focal mechanism parameters). The starting source model is the one proposed by SHOA for the 1985 and 1906 Valparaı´so earthquakes (SHOA, 1999), as shown in Table 9.

The calculations were performed using the 1985 scenario as the reference event and a laterally homogeneous oceanic model with a water layer of 1.5 km. The value of 1.5 km for the thickness of the oceanic layer represents the average bathimetric depth from the source area to the Valparaı´so site, taken to be at a distance of about 50 km. This simple model gives a reliable upper bound of the height of the tsunami (about 3 meters) and the signal computed with this configuration represents the ‘‘reference’’ signal for the other simulations.

Details are given in Tsunami hazard (2008) (http://www.marvasto.bologna.enea.it).

4.3.8 Laterally Heterogeneous Oceanic Models With the modal approach it is very easy to perform expeditious computations for laterally heterogeneous oceanic models (PANZA et al., 2000; PAULATTO et al., 2007) and we computed the tsunami signals at the Valparaı´so site for different cases; the results are shown in Fig. 31. In such a 2D case, the structural model is parameterized by a number of 1D structures arranged in series along the profile from the source to the receiving site.

Details are given in Tsunami hazard (2008) (http://www.marvasto.bologna.enea.it).

4.3.9 Extended Sources For source-site distances comparable with the dimension of the source (near-source), the space extension of the fault may be relevant. In that case the point source approximation may be too crude for the Table 9 Fault parameters for the simulation of the 1906 and 1905 Tsunamis (SHOA, 1999) Parameters

Tsunami 1906

Tsunami 1985

South extreme Slip Length Width Strike Dip Depth Rake

35.1 Lat.S-72 Lon.W 4.6 m 330 km 130 km N10 E 18 15 km 90

34.38 Lat.S-72 Lon.W 2.8 m 200 km 90 km N10 E 18 17 km 105

Author's personal copy 572

M. Indirli et al.

Pure Appl. Geophys.

Figure 31 Tsunami signals computed for the reference case (1D) and different laterally heterogeneous models (2D)

Figure 32 An example of 2D final slip function and rupture history on the fault plane, obtained with Pulsyn (GUSEV and PAVLOV, 2006)

estimation of arrival times, so we adopted an extended source model (see also Earthquake Hazard 2008). To obtain mareograms for the extended source we have developed FORTRAN code that uses the data of the slip distribution along the fault obtained by stochastic procedures using another program (PULSYN) developed by A. Gusev (GUSEV and PAVLOV, 2006). This last program discretizes the fault and assigns a value of the slip and of rupture time to each subsource (see Fig. 32). The characteristic of each subsource is then used as an independent source to model the tsunami and the sum of all the

signals obtained gives us a final mareogram for the extended source.

Using this approach, the tsunami time series have been computed, with a laterally homogenous model, at the Valparaı´so site for different magnitudes which can be associated with different earthquake scenarios.

The results are shown in Fig. 33.

Summarizing the work, using as a base of knowledge the inundation map provided by SHOA (SHOA, 1999; see Fig. 34) associated to the 1906 event, an upper bound of the multiplication factor for the tsunami hazard to be used for the different

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

573

Figure 33 Tsunami signals computed at the Valparaı´so site (about 50 km) for different magnitudes (from 7.5 to 8.7) considering extended source models

Figure 34 SHOA tsunami inundation map

scenarios can be read in Fig. 35. The figure shows the tsunami heights, computed with a scaled and an extended source, are plotted versus magnitude and the associated amplifications (using as reference the 1906 level, whose effective measured

height is debatable since the reports are contradictory).

From the results it emerges that the coastal line in the Valparaı´so harbour zone could be considered exposed to a relatively high risk of flooding.

Author's personal copy 574

M. Indirli et al.

Pure Appl. Geophys.

Figure 35 a Maximum height and b amplification compared to the reference event (1906 earthquake) for the scenario earthquakes considered

Figure 36 Landslide susceptibility maps

4.4. Landslide Hazard Thanks to the indispensable support of SHOA, Valparaı´so Municipality and local universities, slope, landslide inventory and susceptibility maps (Fig. 36) have been provided through in-field campaign (in particular in the pilot sector of the Cerro Cordillera), reconstruction of past landslide events from historic archives, pluviometric analysis and digital/analogical aerial photos elaboration. Landslide hazard is very high in the entire Valparaı´so amphitheatre. The upstream hillside is characterized mainly by muddebris flow events, triggering a couple of times in the year, concentrated in the summer season. The intensity of those phenomena can vary widely, but the presence of densely populated urban settlements in ravine beds, escarpment sides and valley heads (often artificially terraced) makes the associated risk very high. The

coastal flat is reached by moved materials only when the event is intense or when several activated areas merge and flow together in the same bed. Fall events are punctual and characterized by local effects, but often destructive, at the basis of the sub-vertical sides. Certainly, seismic ground shaking as starting point of landslide phenomena should be carefully investigated.

The complete study is available in Landslide hazard (2008) (http://www.marvasto.bologna.enea.it).

4.5. Fire Hazard Fires certainly are the most frequent and dangerous Valparaı´so disasters. The ‘‘state-of-the-art’’ information has been provided by Firemen Corp and Valparaı´so Municipality, with particular regard

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

575

Figure 37 Fire hazard map for Valparaı´so

to the Calle Serrano tragedy. In fact, on February 3rd, 2007, a violent explosion due to a gas leak killed four people, destroyed some heritage buildings and damaged others in Calle Serrano, in the core of the UNESCO zone. Despite the expertise of local firemen, fires occur in the urban area (due to bad maintenance of electric systems and gas pipelines, building materials, lack of education and vandalism), but also in the surroundings forests and bushes (mainly human-made events). The risk is worsened by usual windy weather, narrow and tortuous hill roads, presence of wooden houses and sometimes insufficient water pressure in the hydrants. The presence of the close harbor facilities represents a further risk factor. Moreover, important monuments were burned during the 1906 earthquake, but also damaged by recent fires (as the Church of ‘‘San Francisco del Baro´n’’ in 1983). Figure 37 shows the hazard map, marking the most fire-prone Valparaı´so locations. The work has been verified by a couple of recent fire events: they occurred exactly in one of the areas identified as most fire-prone in the GIS database.

The complete study is available in Fire hazard (2008) (http://www.marvasto.bologna.enea.it).

5. The Cerro Cordillera Investigation Geo-referred hazard maps must interact with a detailed land and building inventory, in which urban planning and single construction features (architecture, structural characteristics, vulnerability, present status, etc.) are linked to the surrounding environmental and social context. The pilot zone of the Cerro Cordillera is an historically ‘‘virgin’’, socially complicated and poor sector, partially inside the UNESCO area, delimited by Calle Serrano (plane side), the San Agustin cable car upper station (hill side), and by the two opposite ravines of San Francisco and San Agustin (Fig. 38).

The architectonic/urban planning investigation encompassed 230 buildings, 4 public areas and about 50 road network stretches. The information (function, architectonic style, general condition, etc., see Fig. 38a1–a3) has been picked up through in situ

personal copy 576

M. Indirli et al.

Pure Appl. Geophys.

Figure 38 Investigation in the Cerro Cordillera: architectonic/urban planning analysis

Figure 39 Investigation in the Cerro Cordillera: vulnerability analysis

surveys (by using an investigation form specifically designed for Valparaı´so), and then stored in the GIS. Different indexes properly overlapped (for example, high architectonic quality and bad conditions, Fig. 38a4), enabled us to identify rehabilitation priorities.

On the basis of the above work, earthquake vulnerability investigation incorporated 70 structures (Fig. 39), when exhaustive cadastral data were

available (plans, prospects, sections, construction details, geotechnical features, etc.), excluding informal and illegal houses.

A special form was elaborated for Valparaı´so, modeled upon established Italian procedures (GNDT, 1999). Almost one half of the analyzed units shows a high vulnerability index IV (22% 0 \ IV \ 30 low vulnerability; 20% 30 \ IV \ 45 average vulnerability; 16% 45 \ IV \ 60 high

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

577

Figure 40 Laser scanner, geometric, photographic and damage survey on churches

vulnerability; 42% 60 \ IV \ 100 very high vulnerability).

The complete study is available in Cerro Cordillera (2008) (http://www.marvasto.bologna.enea.it) and the procedure to set up a building vulnerability inventory in INDIRLI (2009).

6. The Investigation on Churches Thanks to Church Authorities and Valparaı´so Firemen, three important churches, located in different sites and built of different materials, were investigated (La Matrı´z, San Francisco del Baro´n, Las Hermanas de la Divina Providencia, Fig. 40). The following steps were carried out (MAR VASTO, 2007; INDIRLI, 2009): (i) historic data collection; (ii) laser scanner/photographic survey, visual investigation and evaluation of maintenance and damage; (iii) vulnerability evaluation; (iv) execution of preliminary numerical calculations, if necessary; (v) indication of rehabilitation actions.

Vulnerability has been evaluated by using a well known Italian procedure, completing specific survey forms conceived for churches (MOLISE, 2003). The complete study is available in Evaluation of the vulnerability of three churches (2008) (http://www. marvasto.bologna.enea.it).

6.1. Iglesia del Salvador, Matrı´z de Valparaı´so Periodically destroyed by earthquakes, tsunamis and fires, the present fourth version of La Matrı´z was constructed from 1837 to 1842 (and modifications after 1897), in the location of the original first chapel, built after the discovery of the Valparaı´so Bay in 1559, in the ancient nucleus of the ‘‘Barrio Puerto’’.

The church, in simple neoclassic style, is built of adobe perimetral walls (height 12 m and thickness 1.30 m), a masonry fac¸ade, with a roof of clay tiles. The bell-tower (height 40 m), modified at the end of the XIX century, is wooden with an iron spiral staircase inside. The internal colonnades,

Author's personal copy 578

M. Indirli et al.

forming the naves, are also made of wood. In the XX century a certain amount of damage occurred, due to seismic activity, scarce maintenance and termite attacks. Partial renovations were done between 1971 and 1988. The most relevant damage mechanism is in-plane shear actions in the fac¸ade, but the global vulnerability index is about 8%, which is a very low value. In conclusion, ‘‘La Matrı´z’’ can be considered in sufficiently good static conditions, but a general restoration is suggested anyway for fire, materials degradation and termite attacks prevention.

6.2. San Francisco del Baro´n The neo-baroque tower and fac¸ade (brick masonry connected by lime) were erected in 1890–1892, thanks to the project of the architect Eduardo Provasoli. The church faced several earthquakes (mainly 1906 and 1985) without collapse, but severe damage was found mainly in the bell-tower and the arcades during our investigation. The construction seems to be (in the fac¸ade and in the bell-tower) a very regular masonry brickwork, but diagnostics testing is strongly recommended. The building shows heavy widespread structural damage and lack of effective antiseismic protections. The most relevant damage mechanisms are out-of-plane fac¸ade overturning and collapse of the bell-tower. The global damage index is about 33%, but the local damage index in the fac¸ade (66%) is very high. The present damage situation must be considered very worrying, because partial or total collapse (especially in the bell-tower and in the fac¸ade) can occur in case of an earthquake (i.e. medium to high magnitude seismic excitations, as expected in the Valparaı´so area); in fact, the church is unsafe and urgently must be closed partially or totally, implementing both prompt safety measures and overall strengthening as soon as possible.

After several technical meetings with Regional and Church Authorities, a Chilean–Italian team prepared a proposal for a prompt intervention to be done quickly in the beginning of 2009, as an activity developed thanks to the ‘‘MAR VASTO’’ Project cooperation.

Pure Appl. Geophys.

6.3. Las Hermanas de la Divina Providencia The congregation of ‘‘Las Hermanas de la Divina Providencia’’ was constructed in the ‘‘Puerto’’ after 1867. The first chapel underwent various modifications until the fire of 1880. Then, a second version was erected on the Merced Hill (1880–1883), but collapsed almost completely due to the 1906 earthquake and was later demolished. The present building (designed by the architect Victor Auclair in a neorenaissance style but made by a rare primitive reinforced concrete) began in 1907. Las Hermanas Chapel is located in the Almendral at the Merced foothill, exactly where the 1906 earthquake Intensity reached the highest X value. The church was severely damaged by the 1985 earthquake, declared unsafe and almost completely closed without any rehabilitation. The monument is characterized by many critical parameters (fac¸ade tympanum overturning, in-plane shear mechanism in the fac¸ade, transversal response of nave and transept, collapse of the dome, apse overturning, apse and presbytery vaults, and wall shear rupture). The global vulnerability index is about 58%. The present damage situation must considered very worrying, because partial or total collapse (in several structural parts, due to widespread weakness) can occur in case of an earthquake (i.e. medium to high magnitude seismic excitations, as expected in the Valparaı´so area). Due to the particular typology of the construction materials (a primitive reinforced concrete very rare in the world), a strengthening intervention with conventional techniques can be ineffective or very invasive; a solution should be planned only after detailed design work. As a suggestion, an innovative solution can be imagined, in order to reduce drastically the seismic input, involving the introduction of a base isolation system (with all the due precautions, avoiding elevation and foundation wall cutting, by means of the insertion of a new subfoundation system);, this seems possible due to the apparent absence of a crypt.

7. Conclusions The ‘‘MAR VASTO’’ Project showed importance and effectiveness of GIS databases in studying

Author's personal copy Hazard Evaluation in Valparaı´so

Vol. 168, (2011)

historic centers, important for their patrimonial value, prone to natural/anthropic disasters. At the present research stage, the methodology has been sufficiently defined in case of earthquake (hazard mapping; building inventory; architectonic/urban planning, structural vulnerability analyses; intervention proposals; etc.). On the other hand, further standardization in data storing and application of different vulnerability functions for a larger set of building typologies (including specific algorithms already developed by the scientific community) will be necessary. ‘‘MAR VASTO’’ hazard and vulnerability studies covered most of the project resources (limited in time and funds). Starting from the above described results, the authors will try to perform a risk analysis in the future. In fact, the identification of a global risk factor for a given area (or a building) needs deeper investigation. Hopefully, further projects can take advantage of new ongoing studies such as the running EU C26 Action (COST, 2006), in which the Vesuvius eruption is the study case.

Finally, ‘‘MAR VASTO’’ originated important initiatives and further cooperation between Chile and Italy, now in progress, regarding heritage protection.

Acknowledgments We acknowledge the contribution of the anonymous reviewers that helped us to more precisely define some critical points. In the framework of the ‘‘MAR VASTO’’ Project, many people need to be thanked; Italian team: Lorenza Bovio, Fabio Geremei, Francesco Immordino, Lorenzo Moretti, Augusto Screpanti and Edi Valpreda (ENEA); Claudio Alessandri, Marcello Balzani, Daniel Blersch, Paolo Ceccarelli, Daniel Chudak, Gianfranco Franz, Marco Miglioli, Enrico Milani, Gian Paolo Simonini and Antonio Tralli (University of Ferrara); Nieves Lopez Izquierdo (ENEA and University of Ferrara); Claudio Modena (University of Padua); Cristina La Mura, Giuliano Panza, Franco Vaccari and Elisa Zuccolo (ICTP/University of Trieste). Chilean partners: Rodolfo Saragoni H., Maximiliano Astroza I. and Thomas Sturm (Chile University of Santiago); Carlos Aguirre A., Luis Alvarez, Raul Galindo U., Marcela

579

Hurtado S., Gilberto Leiva H. (Federico Santa Maria University of Valparaı´so); Geocom Santiago (Osvaldo Neira F. and Marco Quevedo T.), which provided Laser-Scanner equipment and personnel.

Furthermore, the support of Andres Enriquez, Dante Gutierrez and other SHOA (‘‘Servicio Hidrogra´fico y Oceanogra´fico de la Armada de Chile’’) researchers was wonderful. During the work in Valparaı´so many local Organizations cooperated with the Italian team: Valparaı´so Municipality: above all, Mauricio Gonzalez L., Cristian Palma V., Carolina Avalos A., Claudia Zun˜iga J. (Valparaı´so Municipality professionals at the time of ‘‘MAR VASTO’’, which joined the Italian team also in the framework of some bursaries provided in Italy by the Istituto Italo-Latino Americano); Mayors of Valparaı´so Aldo Cornejo and Jorge Castro, Vice-Mayor Omar Jara A.; other Valparaı´so Municipality professionals, starting from Paulina Kaplan D., director of the ‘‘Oficina de Gestion Patrimonial’’, with many others; Intendencia V Region Valparaı´so: Intendente Ivan de la Maza, Karina Englander K., Juan Carlos Garcia P. de Arce and others; Church Authorities: Father Fernando Candia (San Francisco Church), Mons. Gonzalo Duarte Garcı´a de Corta´zar (Bishop of Valparaı´so) and others; Other Chilean Institutions: the Ministry of Culture (‘‘Consejo Nacional de la Cultura y Las Artes’’); Ana Maria Icaza and Francisco Saavedra (Programa de Recuperacion y Desarrollo Urbano de Valparaı´so-PRDUV); Guillermo De La Maza (OREMI, Civil Defense); Enzo Gagliardo L. (Head), Vicente Maggiolo O. and colleagues (‘‘Bomba Italia’’) of the Valparaı´so (‘‘Bomberos’’) Firemen; the Bote Salvavidas personnel (Valparaı´so Sea Rescue Corp); the Police (‘‘Carabineros de Chile’’); Nelson Morgado L. and many others of the Valparaı´so Board of Architects; other Universities (‘‘Pontificia U. Catolica de Valparaı´so’’, ‘‘U. de Valparaı´so’’, U. de Playa Ancha Valparaı´so); Luis Enriquez, Javier Troncoso (‘‘Gerencia Barrio Puerto’’, the historic district of the City); ‘‘Junta de Vecinos’’ of the Cerro Cordillera; Chilean professionals: above all, Milagros Aguirre D., always very helpful and kind; Luis Bork V., Fabio Mezzano P., Octavio Pe´rez A., Alfonso Salinas, Francisco Silva I., Gunther Su¨hrcke and many others; grateful thought for the great support to: the Italian Embassy in

Author's personal copy 580

M. Indirli et al.

Santiago; Roberto Santilli, Maruzzella Giannini and other office workers of the Italian Trade Commission in Chile; Pablo Peragallo of the Valparaı´so Italian Community. Last but not least, special thanks to Arcindo Santos and other professionals of BID/IADB (Banco Interamericano de Desarrollo/InterAmerican Development Bank). We used GMT software (Wessel and Smith, 1991) in the preparation of some figures.

REFERENCES ASTROZA, M. I., NORAMBUENA, A., and ASTROZA, R. (2006), Reinterpretacion de las intensidades del terremoto de 1906, Proc.

International Conference Montessus de Ballore 1906 Valparaı´so Earthquake Centennial ASTROZA, M. I. (2007), A re-interpretation of the Valparaı´so 1906 earthquake intensities/Reinterpretacio´n de las intensidades del terremoto de 1906. Proc. VI Congreso Chileno de Geote´cnia, Valparaı´so, Chile, 28–30 November 2007 BERESNEV, I. A. and ATKINSON, G. M. (1997), Modeling finite-fault radiation from the omega n spectrum. In Bulletin of the Seismological Society of America, 87(1):67–84 CANCANI, A. (1904), Sur l’emploi d’une double echelle seismique des intesites, empirique et absolue, G Beitr 2, 281–283 Cerro Cordillera (2008), Cerro Cordillera pilot project (Valparaı´so), MAR VASTO Project Technical Report. http://www. marvasto.bologna.enea.it CHOY, G. L. and DEWEY, J. W. (1988), Rupture process of an extended earthquake sequence: teleseismic analysis of the Chilean earthquake of March 3, 1985, J Geophys Res 93, 1103–1118 COMER, R. P. (1984a), The tsunami mode of a flat earth and its excitation by earthquake sources, Geophys J Astr Soc 77, 1–28 COMER, R. P. (1984b), Tsunami generation: a comparison of traditional and normal mode approach, Geophys J Astr Soc 77, 29– 41 COMTE, D., EISENBERG, A., LORCA, E., PARDO, M., PONCE, L., SARAGONI, G.R., SIGH, S.K., SUAREZ, G. (1986), The 1985 central Chile earthquake: a repeat of previous great earthquakes in the region?, Science 233, 449–452 CORNELL, C. A. (1968), Engineering seismic risk analysis, Bull Seism Soc Am 58, 1583–1606 COST (2006), COST, European Cooperation in the field of Scientific and Technical research, Transport and Urban Development, COST Action C26: ‘‘Urban Habitat Constructions Under Catastrophic Events’’. In Proc. Symposium on ‘‘Urban habitat construction under catastrophic events’’, Working Group 4 ‘‘Risk assessment for catastrophic scenarios in urban areas’’, Session 6: Volcanic hazard and the Vesuvius study case; Malta, 23–25 October 2008 DECANINI, L., MOLLAIOLI, F., PANZA, G.F., ROMANELLI, F., and VACCARI, F. (2001), Probabilistic vs deterministic evaluation of seismic hazard and damage earthquake scenarios: a general problem, particularly relevant for seismic isolation. Proc. 7th International Post-Smirt Seminar on Seismic Isolation, Passive Energy Dissipation and Active Control of Vibration of Structures, Assisi, Italy, 2–5 October, 2001

Pure Appl. Geophys.

DGPS survey (2008), DGPS survey in the city of Valparaı´so, MAR VASTO Project Technical Report. http://www.marvasto. bologna.enea.it DOLCE, M., MARTELLI, A., and PANZA, G.F. (2005), Proteggersi dal terremoto: le moderne tecnologie e metodologie e la nuova normativa sismica. Seconda edizione. 21mo Secolo, 336 pagine, ISBN 88-87731-28-4 EARTHQUAKE HAZARD (2008), Earthquake hazard in the city of Valparaı´so, MAR VASTO Project Technical Report. http:// www.marvasto.bologna.enea.it ESPINOZA, P. (2000), Amplificacion sismica en suelos y microzonificacion de los sectores planas de Vina del Mar y Valparaı´so, Civil Engineering Thesis. Departamento de Obras Civiles, Universidad Tecnica Federico Santa Maria, Valparaı´so, Chile Evaluation of the vulnerability of three churches (2008), Evaluation of the vulnerability of three churches in Valparaı´so and numerical calculation, MAR VASTO Project Technical Report. http://www.marvasto.bologna.enea.it FA¨H, D. and PANZA, G.F. (1994), Realistic modelling of observed seismic motion in complex sedimentary basins. Annali di Geofisica 37, 1771–1797 FIRE HAZARD (2008), Fire hazard in the city of Valparaı´so, MAR VASTO Project Technical Report. http://www.marvasto. bologna.enea.it GIS database (2008), A GIS database for the city of Valparaı´so, MAR VASTO Project Technical Report. http://www.marvasto. bologna.enea.it GNDT (1999), Second Level vulnerability form for masonry buildings/Scheda di vulnerabilita` di 2 livello per edifici in muratura, 1999 GSHAP. The Global Seismic Hazard Assessment Program, http://seismo.ethz.ch/gshap/index.html GUSEV, A. A. (1983), Descriptive statistical model of earthquake source radiation and its application to an estimation of short period strong motion, Geophys J R Astron Soc 74, 787–800 GUSEV, A. A. and PAVLOV, V. (2006), Wideband simulation of earthquake ground motion by a spectrum-matching, multiplepulse technique. Proc. First European Conference on Earthquake Engineering and Seismology (a joint event of the 13th ECEE & 30th General Assembly of the ESC), Geneva, Switzerland, 3–8 September 2006. Paper Number: 408 GUTIERREZ, D. (2005), Advances in the Chilean tsunami warning system and application for the TIME project on the Chilean Coast. Presentation at GFZ Potsdam HAMMACK, J. L. (1973), A note on tsunamis: their generation and propagation in an ocean of uniform depth, J. Fluid Mechanics 60, 769–799 HASKELL, N. (1964), Total energy and energy spectral density of elastic wave radiation from propagating faults, Bull Seismol Soc Am 56, 1811–1842 INDIRLI, M. (2009), The organization of a GIS database on natural hazards and structural vulnerability for the historic center of San Giuliano di Puglia (Italy) and the City of Valparaı´so (Chile), Int J Archit Herit, accepted paper, publication in progress INDIRLI, M., CARPANI, B., PANZA, G., ROMANELLI, F., and SPADONI, B. (2006), Damage evaluation and rehabilitation of the Montorio medieval tower after the September 14th, 2003 earthquake. Proc. of the First European Conference on Earthquake Engineering and Seismology (a joint event of the 13th ECEE & 30th General Assembly of the ESC), Geneva, Switzerland, 3–8 September 2006, Paper Number: 666

Author's personal copy Vol. 168, (2011)

Hazard Evaluation in Valparaı´so

KAJIURA, K. (1963), The leading wave of tsunami, Bull Earthq Res Inst 41, 535–571 KLU¨GEL, J. U., MUALCHIN, L., and PANZA, G. F. (2006), A scenariobased procedure for seismic risk analysis, Eng Geol 88, 1–22 LANDSLIDE HAZARD (2008), Geomorphological hazard in the city of Valparaı´so, MAR VASTO Project Technical Report. http:// www.marvasto.bologna.enea.it LEE, J. J. and CHANG, J. J. (1980), Water waves generated by an impulsive bed upthrust of a rectangular block, Appl Ocean Res 2, 165–170 LOMNITZ, C. (1971), Grandes Terremotos y Tsunamis en Chile Durante el Periodo 1535–1955, Revista Geofı´sica Panamericana, 1(1), 151–178 LOMNITZ, C. (1983), On the epicenter of the great Santiago earthquake of 1647, BSSA 73, 885–886 MAR VASTO (2007), Risk Management in Valparaı´so/Manejo de Riesgos en Valparaı´so, Servicios Te´cnicos (acronym MAR VASTO), funded by BID/IDB (Banco Inter-Americano de Desarrollo/Inter-American Development Bank). Project ATN/II9816-CH, BID/IDB-ENEA Contract PRM.7.035.00-C, March 2007–October 2008. http://www.marvasto.bologna.enea.it MENDOZA, C., HARTZELL, S., and MONFRET, T. (1994), Wide-band analysis of the 3 March 1985 Central Chile earthquake; overall source process and rupture history. Bull Seismol Soc Am 84(2), 269–283 MOLISE (2003), Regione Molise, CNR. Second Level form for the evaluation of damage and vulnerability in the churches/Scheda chiese di secondo livello per la valutazione del danno e della vulnerabilita`, 2003 OKAL, E. A. (1982), Mode-wave equivalence and other asymptotic problems in tsunami theory, Phys Earth Planet Inter 30, 1–11 PANZA, G. F. and SUHADOLC, P. (1987), Complete strong motion synthetics, In Seismic strong motion synthetics, computational techniques 4, (ed. B. A. Bolt), Academic Press, Orlando, 153– 204 PANZA, G. F., VACCARI, F., COSTA, G., SUHADOLC, P., and FA¨H, D. (1996), Seismic input modelling for zoning and microzoning, Earthq Spectr, 12 529–566 PANZA, G. F., VACCARI, F., and CAZZARO, R. (1997), Correlation between macroseismic intensities and seismic ground motion parameters, Ann Geof 15, 1371–1382 PANZA, G. F., VACCARI, F., and ROMANELLI, F. (1999), The IUGSUNESCO IGCP Project 414: Realistic modeling of seismic input for megacities and large urban areas. Episodes 22(1), 26– 32 PANZA, G.F., ROMANELLI, F., and YANOVSKAYA, T. (2000), Synthetic Tsunami mareograms for realistic oceanic models, Geophys J Int 141, 498–508 PANZA, G. F., ROMANELLI, F., and VACCARI, F. (2001), Seismic wave propagation in laterally heterogeneous anelastic media: theory and applications to seismic zonation. Adv Geophys 43, 1–95 PARVEZ, I. A., VACCARI, F., PANZA, G. F. (2003), A deterministic seismic hazard map of India and adjacent areas, Geophy J Int 155(2), 489–508 PAULATTO, M., PINAT, T., and ROMANELLI, F. (2007), Tsunami hazard scenarios in the Adriatic Sea domain, Nat Hazards Earth Syst Sci 7, 309–325 PERESAN, A., KOSSOBOKOV, V., ROMASHKOVA, L., and PANZA, G. F. (2005), Intermediate-term middle-range earthquake predictions in Italy: a review. Earth Sci Rev 69, 97–132.

581

POD’YAPOLSKY, G. S. (1968), Generation of longperiodic gravitational wave in the ocean by seismic source in the crust, Izvestia AN SSSR, Fizika Zemli 1, 7–24 ROMANELLI, F., VACCARI, F., and PANZA, G. F. (2003), Realistic modelling of the seismic input: site effects and parametric studies. J Seismol Earthq Eng 5(3), 27–39 ROMANELLI, F., PANZA, G.F., and VACCARI, F. (2004), Realistic modelling of the effects of asynchronous motion at the base of bridge piers. J Seismol Earthq Eng 6(2), 19–28 SARAGONI H. R. G. (2006), Estudio comparativo de los efectos de los terremotos de Valparaı´so de 1906 y 1985 en el barrio El Almendral. Proc. International Conference Montessus de Ballore 1906 Valparaı´so Earthquake Centennial SARAGONI, H. R. G. (2007), The 1906 Valparaı´so earthquake/El terremoto de Valparaı´so de 1906. Proc. VI Congreso Chileno de Geote´cnia, Valparaı´so, Chile, 28–30 November 2007 SHOA (1999), Carta de Inundacion por Tsunami para la bahia de Valparaı´so, Chile. http://www.shoa.cl/servicios/citsu/citsu.php SOMERVILLE, P. G., SEN, M. K., and COHEE, B. P. (1991), Simulation of strong ground motions recorded during the 1985 Michoacan, Mexico and Valparı´so, Chile earthquakes, Bull Seism Soc Am 81(1), 1–28 SSHAC (1997), Senior Seismic Hazard Analysis Committee. Recommendations for probabilistic seismic hazard analysis: guidance on uncertainty and use of experts. NUREG/CR-6372 STURM, T. W. (2008), Valparaı´so: its historic heritage and earthquakes/Valparaı´so: su patrimonio histo´rico y los sismos’’, Civil Engineer Thesis, University of Chile, Faculty of Physical and Mathematical Sciences, Department of Civil Engineering, Santiago de Chile, April 2008 TANNER, J. G. and SHEDLOCK, K. M. (2004), Seismic hazard maps of Mexico, the Caribbean, and Central and South America, Tectonophysics, 390, 159–175 TSUNAMI HAZARD (2008), Tsunami hazard in the city of Valparaı´so, MAR VASTO Project Technical Report. http://www. marvasto.bologna.enea.it VACCARI, F., ROMANELLI, F., and PANZA, G. F. (2005), Detailed modelling of strong ground motion in Trieste. GT&A 2, 7– 40 VERDUGO, A. I. (1995), Estudio geofisico de los suelos de fundacion para una zonificacion sismica de Valparaı´so y Vina del Mar, Civil Engineer Thesis. Dept. Civil Engr. Fac. Ciencias Fisicas y Matematicas, Universidad de Chile, Santiago, Chile VERDUGO, A. (2006a), Amplificacio´n sı´smica en los sectores planos de Vin˜a del Mar y Valparaı´so, International Conference Montessus de Ballore 1906 Valparaı´so Earthquake Centennial, Valparaı´so, Chile VERDUGO, R. (2006b), Caracterizacio´n geote´cnica de Valparaı´so y su efecto en el Terremoto de 1906, International Conference Montessus de Ballore 1906 Valparaı´so Earthquake Centennial, Valparaı´so, Chile, 2006 WARD, S. N. (1980), Relationship of tsunami generation and an earthquake source, J Phys Earth 28, 441–474.

WARD, S. N. (2002), Tsunamis, In Encyclopedia of Physical Science and Technology (Academic Press, California) WARD, S. N., and DAY, S. (2008) Tsunami balls: a granular approach to tsunami runup and inundation, Commun Comput Phys 3(1), 222–249 WESSEL, P., and SMITH, W. H. F. (1991), Free software helps map and display data, EOS Trans AGU, 72, p. 441.

Author's personal copy 582

M. Indirli et al.

YANOVSKAYA, T. B., ROMANELLI, F., PANZA, G. F. (2003), Tsunami excitation by inland/coastal earthquakes: the Green function approach, Nat Haz Earth Syst Sci 3, 353–365

Pure Appl. Geophys.

ZUCCOLO, E., VACCARI, F., PERESAN, A., DUSI, A., MARTELLI, A., and PANZA, G.F. (2008), Neo-deterministic definition of seismic input for residential seismically isolated buildings. Eng Geol 101, 89–95

(Received May 15, 2009, revised March 25, 2010, accepted April 6, 2010, Published online July 6, 2010)