URBAN TREE CANOPY ASSESSMENT IN THE DISTRICT OF COLUMBIA
2006 to 2011
Prepared for: Metropolitan Washington Council of Governments & the District Department of Transportation’s Urban Forestry Administration Prepared by: AMEC Environment & Infrastructure and Plan-It Geo LLC
August 2013
page left intentionally blank
Acknowledgements The District Department of Transportation (DDOT)’s Urban Forestry Administration (UFA) would like to thank the U.S. Forest Service, Metropolitan Washington Council of Governments (MWCOG), and Office of the Chief Technology Officer (D.C.OCTO) for their cooperation and assistance on this project.
Funding for this project was provided by the USDA Forest Service Urban & Community Forestry Program.
The USDA is an equal opportunity provider and employer.
Prepared by: AMEC Environment & Infrastructure (2012) with updates by Plan-It Geo LLC (2013) Ian Hanou, Senior Analyst and Project Manager Richard Thurau, Ph.D., Geospatial and Environmental Scientist
page left intentionally blank
Table of Contents Section
Page
Table of Contents .............................................................................................................. i Executive Summary .......................................................................................................... ii Introduction ...................................................................................................................... 3 Major Findings .................................................................................................................. 3 Methodology .................................................................................................................... 6 Key Terms and Definitions ..................................................................................... 6 Imagery .................................................................................................................. 6 GIS Data.................................................................................................................. 7 Land Cover Classification ....................................................................................... 8 Urban Tree Canopy (UTC) and Possible Planting Area (PPA) Assessment .......... 10 Results ............................................................................................................................ 13 Land Cover & UTC in the District ................................................................................... 13 Tree Canopy Metrics by Wards ...................................................................................... 15 Tree Canopy by Advisory Neighborhood Commissions ................................................. 19 Tree Canopy by Sub-Watersheds................................................................................... 25 Tree Canopy Metrics by Land Use and Zoning Categories ............................................ 28 Tree Canopy Metrics by Census Block ........................................................................... 31 Tree Canopy by Parcel ................................................................................................... 32 Reasons for Gains and Losses of Urban Tree Canopy in the District ............................. 35 Recommendations.......................................................................................................... 39 Summary......................................................................................................................... 41 Appendix......................................................................................................................... 43
i
Urban Tree Canopy in the District of Columbia (2006 -2011)
page left intentionally blank
ii
Urban Tree Canopy in the District of Columbia (2006 -2011)
Executive Summary The District of Columbia (D.C.) Department of Transportation’s (DDOT) Urban Forestry Administration (UFA) and other Federal and local stakeholders care for and manage the urban forest of the District of Columbia. The Urban Forestry Administration is responsible for maintaining the health of the District’s street tree canopy including 148,000 trees growing in the street rights-of-way adjacent to homes, businesses, not-for-profits, and government. An Urban Tree Canopy (UTC) assessment provides a periodic mapping of the District’s canopy and is an essential tool for urban forest management, policy, and planning. D.C. UFA and other partners will use these results to identify strategies to reach the District’s UTC Goal of 40% by 2035. This report provides UTC results, tools, and management recommendations for implementing, attaining, and maintaining goals related to canopy cover, education, and outreach. Urban Tree Canopy in the District This assessment covered approximately 68 square miles (43,583 acres). Based on land area of 39,336 acres (excludes water), D.C. currently has approximately 37% tree canopy (14,601 acres), with 26% available planting space (10,343 acres), and 37% of land that is unsuitable for tree planting (14,392 acres). Overall, data indicates that canopy cover may have increased by nearly 1% from 2006 to 2011 (percentages are rounded to the nearest whole number). District of Columbia
2006
2011
Tree Canopy (acres)
14,246
14,601
Tree Cover (percent)
36%
37%
Across other planning and management boundaries, canopy cover ranged from 21% – 53% for Wards, 14% – 65% in Advisory Neighborhood Commissions, and 6% – 66% for land use categories. The street rights-of-way contain 3,005 acres of tree canopy, for an average 21% cover. Setting Canopy Cover Goals Tree cover is a critical measure of environmental health and sustainability in cities. Setting realistic canopy goals for different land use types and planning scales creates a shared vision to ensure that resources are allocated effectively and adaptive management can occur over time. Based on assessing the District’s land cover using a top-down approach, to reach the District’s goal of 40% cover by 2035, roughly 1,150 acres of new tree canopy is needed (even more when considering canopy loss from development, storms, and natural mortality). Results, tools, and techniques are presented to foster implementation at a variety of community planning scales.
1
Urban Tree Canopy in the District of Columbia (2006 -2011)
Using the Results of this Assessment Assessment data may sit on the shelf underutilized if information isn’t provided in formats appropriate for different users. To increase application of the data, several technical and non-technical tools were provided. These tools include:
Interactive neighborhood PDF maps of UTC and planting potential
A Canopy Calculator spreadsheet tool
Scenario planning tools to prioritize and tracking trees and their benefits
These sophisticated but easy-to-use products provide the District with a way to objectively target certain Wards, ANCs, and streets where trees are needed most. Recommendations and Summary Results indicate the District has added approximately 350 acres of canopy (70 acres/year). If that trend were to continue, D.C. could add 1,700 acres of canopy by 2035, reaching 41.5% tree canopy. However, 195 parcels were identified with ¼ acre of forest loss or more, totaling 227 acres, mostly due to new development. If that trend continues coupled with increasing threats from storm and insect damage, it will likely outpace the planting and growth of individual trees, working against the District’s UTC goal. Precisely identifying and realizing tree canopy growth in an ultra-urban environment such as the District is a challenge. This report discusses methods and accuracy of canopy cover change assessments. Even within a reasonable margin of error (e.g. .5%), comparing GIS tree canopy layers can report false UTC gains or losses. A statistical evaluation showed that misclassification accounted for 13% of false gains and 20% of false losses at the parcel-level. Therefore the data indicates gains and losses at a high degree of accuracy as defined within the remote sensing industry. Overall, this project advances the scientific body of knowledge and technologies associated with UTC assessments while providing creative and innovative maps, tools, and summaries to help reach and maintain the District’s commendable tree canopy goals. This Urban Tree Canopy Assessment provides a progress benchmark after years of committed tree planting, maintenance, non-profit, and education/outreach programs. In order to reach the Urban Tree Canopy Goal of 40% by 2035, numerous general recommendations are provided in this report. A few of the action items suggested for the District are: A. Target planting on large properties and in open spaces such as District and Federal parks. B. Develop individual UTC sub-goals based on land use within ANCs and/or Wards. C. Map and plan tree plantings that enhance forest cover in areas addressing forest fragmentation and in grass strips adjacent to large areas of impervious surfaces. D. Engage multiple community partners in planting and maintaining trees.
2
Urban Tree Canopy in the District of Columbia (2006 -2011)
Major Findings Based on the analysis of aerial imagery, land cover, land use and the resulting GIS mapping data, the following represent the major findings from this study:
The District has approximately 37% tree canopy (14,601 acres) based on 2011 imagery. Canopy loss from new development, though partially offset by the growth of existing trees and new tree planting efforts, most notably hinders attainment of the District’s 40% UTC goal by 2035. Possible planting area (PPA) including grass and parking lots cover 10,343 acres, or more than a third (36.6%) of the District. 2011 UTC in Wards ranged from 21% (Ward 6) to 53% (Ward 3). All Wards appear to have maintained or gained canopy cover between 2006 and 2011 with the exception of Ward 3 which lost roughly 3%. Forest cover in the District’s Advisory Neighborhood Commissions (ANCs) ranged from 14% (1A and 1B) to 65% (3G) 2011 UTC. Of all UTC in the District, 5,231 acres (36%) is found on open urban land, 28% comes from residential land uses, and 22% from transportation/rights-of-way (see Land Use on pages 11, and 28-30). Reaching 40% average canopy cover in residential areas (Low, Medium and High density) is achievable. It would require establishing 974 acres of new canopy, only ¼ of the available planting space (3,812 acres). Public rights of way adjacent to sidewalks cover 5.5% (2,182 acres) of the District’s total land area and make-up 6% (918 acres) of the total UTC; these are DDOT/UFA primary planting areas.
3
Urban Tree Canopy in the District of Columbia (2006 -2011)
Introduction Trees provide numerous economic, social, and environmental benefits as well as services that form the basis of livability in urban areas. It is important for urban development to work closely with urban forest health protection and management goals to maintain community livability. The recent influx of new urban residents and neighborhood revitalization in the District has created challenges for maintaining existing canopy and for providing planting areas for new trees. Geographic distribution of land ownership also plays a critical role in maintaining a uniform urban forest. Each category of ownership has unique management objectives and regulatory constraints. A primary example is D.C.’s land ownership distribution of which is 43% federal, 50% private and 7% D.C. government. This presents a unique challenge for the District of Columbia in providing comprehensive urban forest health management services. With funding from the U.S. Forest Service, MWCOG contracted with AMEC Environment & Infrastructure (AMEC) and later Plan-It Geo to update the District’s urban tree canopy (UTC) mapping. This assessment examines existing tree canopy, available planting in grassy areas, streets, and parking lots, and documents canopy cover change between 2006 and 2011. Specific objectives of this assessment are to: Map eight (8) land cover classes across the District. Map and quantify existing canopy, planting areas, and tree canopy change for eight (8) assessment boundaries. These included the District, Wards, Advisory Neighborhood Commissions, land use, subwatersheds, 2010 census blocks, parcel boundaries, and street rights-of-way. See Table 1 on page 11. Prioritize tree planting opportunities and provide technical as well as non-technical tools for implementation and outreach. Track and report on the District’s progress toward canopy cover goals using the results in a report and summary factsheet. Present the results to MWCOG and District of Columbia DDOT/UFA staff. In this report, tree canopy goals are discussed and compared to the District’s goal of 40% canopy cover by 2035. Innovative methods and tools for setting and implementing canopy cover goals are offered to facilitate this process. Gains and losses in canopy are also assessed to describe and illustrate the District’s changing urban forest landscape, followed by management and technical recommendations.
4
Urban Tree Canopy in the District of Columbia (2006 -2011)
It should be noted that this is not the first urban forest assessment in the District. The following is a list of forest assessments conducted in the District in the last decade: In 2002, American Forests conducted an “Urban Ecosystem Analysis for the Washington D.C. Metropolitan Area”, identifying tree loss hot spots and an initial estimate of the environmental benefits of trees in the region. In 2004 and 2009, Casey Trees Foundation conducted an i-Tree Eco study (previously UFORE) that revealed a 34% increase in the number of trees planted in the District. In 2009, the University of Vermont (UVM) and US Forest Service Northern Research Station conducted an Urban Tree Canopy study in the District based on 2006 QuickBird satellite imagery. In 2012, AMEC and Plan-It Geo re-evaluated the 2006 UVM data and utilized 2011 aerial imagery (National Agricultural Imagery Program – NAIP) to assess areas of UTC gains and losses and provide tools for highlighting tree planting opportunities. Additionally, UFA staff continually updates an inventory database of 150,000 trees in public spaces and street rights-of-way.
This assessment also provides data and tools to develop local and regional urban forestry goals, policies, outreach, and management plans to sustain and enhance the existing urban forest.
5
Urban Tree Canopy in the District of Columbia (2006 -2011)
Methodology The District’s UTC assessment was completed through a semi-automated land cover classification process, strict Quality Assurance / Quality Control (QA/QC), and spatial modeling and analysis. When combined, these steps are used to assess the distribution of UTC and land cover among selected assessment boundaries (Table 1). Data inputs and processes associated with the image classification and UTC Assessment are described below. Further technical details are provided in the Appendix. Key Terms and Definitions GIS – Geographic Information Systems are computer systems that were used to map tree canopy across the District. NAIP – National Agriculture Imagery Program is administered by the U.S. Department of Agriculture and provides 1-meter pixel resolution aerial photographs across the United States captured during the growing season, typically flown at 20,000’ average mean terrain. LiDAR – Light Detection and Ranging is an active aerial remote sensing modality involving pulses of near-infrared light collected through a series of mirrors recording elevation values of features including trees, buildings, bare ground, and more. UTC – Urban Tree Canopy is the layer of leaves, branches, and stems of trees that cover the ground when viewed from above. PPA – Possible Planting Areas are non-forest vegetation and paved surfaces where it is biophysically possible and feasible to plant trees. PPA sub-types are described on page 10. Unsuitable UTC – Areas where it is not feasible to plant trees. For this assessment it includes buildings, roads, water, barren land in construction sites, recreational sports fields, wetlands, and other specially designated areas such as select Federal properties (e.g. National Park service lands). Imagery Three (3) imagery data sets provided the foundation for land cover mapping in this study. Descriptions of each imagery set and how they were used are described in Figure 1. 2011 aerial photography at one-meter pixel resolution from the National Agricultural Imagery Program (NAIP) provided summer “leaf-on” imagery. This was used to map the District’s current tree canopy. Spring 2010 “leaf-off” aerial imagery at six-inch pixel resolution provided by D.C. GIS was used to map impervious surfaces, bare soil, and grass/open space vegetation. September 2006 QuickBird satellite imagery at two-foot pixel resolution was used to map 2006 tree canopy to compare with 2011 results and assess UTC change.
6
Urban Tree Canopy in the District of Columbia (2006 -2011)
2011 “Leaf-On” NAIP Ortho Aerial Photography
2010 “Leaf-Off” Ortho Aerial Photography
2006 “Leaf-On” QuickBird Satellite Imagery
(Four spectral bands, Onemeter pixel resolution)
(Three spectral bands, Six-inch pixel resolution)
(Four spectral bands, Two foot pixel resolution)
Figure 1: Sources of imagery used in this assessment
Note that LiDAR data flown in December 2008 was available but not used; given this project’s intent was to assess UTC change over five years (2006 to 2011), using LiDAR data flown at the exact midway point between the targeted assessment timeframe was not ideal. The use of LiDAR data is encouraged for the District’s next UTC Assessment assuming the timeframe and quality meet the project requirements. GIS Data GIS data layers seen in the map in Figure 2 (page 8) were provided by MWCOG, D.C. GIS and other sources to (1) improve the accuracy of the remotely-sensed land cover classification, and (2) assess UTC and land cover results by assessment boundaries. Land cover classification was augmented by using buildings, roads, sidewalks, and parking lots to refine impervious surface mapping. To map wetlands and water classifications, the National Wetlands Inventory (NWI), D.C. wetland restoration areas, and D.C. GIS water layers were used. Assessment boundaries were selected to understand land cover and UTC across different planning and management units of the District. Boundaries can be categorized by themes including political (District, Wards), management and planning (Land Use, Zoning, Ownership, Tax Parcels, Rights of Way), community, population and social demographics (Census Blocks, ANCs), and by environmental features (subwatersheds). See Table 1 on page 11 for more information on these UTC assessment boundaries.
7
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figure 2: GIS Data Inputs used in the Land Cover Mapping: water (blue), wetlands (magenta), roads (black), parking lots (yellow), buildings (red) and sidewalks (orange). Land Cover Classification Eight (8) land cover types were mapped through an object-based remote sensing classification of the 2011 NAIP imagery, the 2010 D.C. GIS aerials, and GIS inputs described above. A visual, manual QA/QC editing process was conducted at approximately 1�:2,000’ scale with particular focus on the tree canopy and impervious surface results. Erroneous polygons from the automation were digitized and classified correctly to achieve 95% overall accuracy. The final land cover classes shown in Figure 3 below are: 1) Tree canopy (forests and individual trees classified from the 2011 NAIP imagery) 2) Buildings (provided by D.C. GIS) 3) Roads (provided by D.C. GIS) 4) Other paved impervious surfaces (combination of D.C. GIS data and classification) 5) Other Green Vegetation (grass/open areas) 6) Water (provided by D.C. GIS with manual edits by Plan-It Geo) 7) Bare soil (dirt/sand/gravel/rock classified from the 2011 NAIP imagery) 8) Wetlands (provided by D.C. GIS and NWI)
8
Urban Tree Canopy in the District of Columbia (2006 -2011)
Bing Maps Base Layer
Figure 3: Classified Land Cover Data in the District with Inset Maps.
9
Urban Tree Canopy in the District of Columbia (2006 -2011)
Urban Tree Canopy (UTC) and Possible Planting Area (PPA) Assessment This UTC assessment maps four (4) main UTC types across the District, defined and described below. The area and percent of each UTC type was reported for all assessment boundaries (see Table 1). Detailed methods regarding the UTC metrics are provided in the Appendix. Note that UTC percentages (%) are calculated based on a total land area excluding water. Existing forest canopy is comprised of all forests and individual trees mapped from the 2011 summer NAIP imagery. Possible planting area (PPA) is defined as the total land area where (a) no tree canopy cover currently exists and (b) it is possible to plant trees. In this analysis, non-tree vegetation such as grass and open space (PPA – Vegetation) and suitable impervious surfaces such as parking lots (PPA – Impervious) are combined to report Total PPA. This does not equal potential canopy but rather the on-the-ground area available for consideration of planting opportunities. Areas considered unrealistic, unavailable, or undesirable for planting were removed for more reliable and accurate results. Examples of areas removed include tennis courts, sports fields, and lawns along the Downtown Mall owned and managed by the National Park Service (see Appendix for more details). Unsuitable UTC is the combination of bare soil, roads, buildings, and water. Soil and dry vegetation are considered unsuitable since the present land cover does not support biophysical health (vegetation is either dead or absent). However, some areas mapped as Unsuitable UTC could become PPA through natural and human processes over time.
From left to right, local area photos showing examples of Existing UTC, Possible Planting Area (Vegetation), and Possible Planting Area (Impervious). UTC Change was calculated by comparing 2006 imagery results with the 2011 canopy cover results. The 2006 canopy data layer was provided by D.C. GIS and compared against 2006 imagery, where it was determined that additional QA/QC was needed for a reliable change analysis (see Appendix for more details). UTC change calculated this way should be considered approximate. Overestimation or underestimation in the canopy layer from either time period may produce false gains or losses, in particular at finer-scale assessment boundaries such as neighborhoods or parcels. Generally speaking, losses are much easier to map accurately than gains, which are more subtle. 10
Urban Tree Canopy in the District of Columbia (2006 -2011)
Table 1: List and map of eight UTC Assessment Boundaries in the District
Assessment Boundaries
No. of Features
Description
District of Columbia
1
Official District boundary
8
D.C.‘s political boundaries (based on population; shown here labeled by Ward number)
39
Wards are divided into 39 sub-areas called Advisory Neighborhood Commissions.
10
Generalized land use classification scheme: low, medium, high density residential, commercial, mixed use, industrial, institutional, parks and open space, transportation and water (See Appendix).
Wards
Advisory Neighborhood Commissions (ANCs)
Land Use
11
Urban Tree Canopy in the District of Columbia (2006 -2011)
Map
Assessment Boundaries
No. of Features
Subwatersheds
4
Major drainages of D.C.’s municipal area
Census Blocks
6,507
2010 U.S. Census data provides demographic data available at the tract, block group, and block level. The most detailed ‘block’ level was used for this project.
Parcels
134,440
Parcels from tax assessors property database
N/A
This layer represents the public rights-of-way (PROW). It includes subtypes for roadway, sidewalks, parking, and service road.
Streets/Road Rights-of-Way
12
Description
Urban Tree Canopy in the District of Columbia (2006 -2011)
Map
Results Detailed land cover and UTC assessment and analysis results are presented below. UTC results focusing on three types (Existing UTC, Possible Planting Area, and UTC change) are presented starting with the District and other assessment boundaries.
Land Cover & UTC in the District Land cover in the District is dominated by forests and other vegetation, accounting for a combined 54% of the total land area. The ability of tree canopy to overlap other land cover types contributes to its landscape dominance. UTC metrics averaged across the District are reported in Table 2 below. Out of 39,335 land acres, 14,601 acres (37%) Figure 4. Distribution of were covered by tree canopy in 2011. land cover in acres Possible Planting Area (PPA) is broken out into vegetation (PPA Veg.) and impervious (PPA – Imp.) areas where it is suitable to plant trees. Across the District, a combined 10,343 acres (26%) are available for planting, demonstrating a high potential for UTC growth. Most impervious areas across the District (14,392 acres or 37% of the land area) are comprised of buildings and roads, where it is not suitable or feasible to plant trees. Table 2: Urban Tree Canopy Metrics for the District of Columbia. 1
The overall tree canopy goal (40% by 2035) is highlighted. Percentages are based on total land area, excluding area in the District that is water such as the Potomac and Anacostia Rivers and small bodies of water such as lakes and ponds.
1
Reflects the land area as determined in the 2004 Land Use Spatial Data
13
Urban Tree Canopy in the District of Columbia (2006 -2011)
UTC in the District, at a glance … 21,289 out of 134,323 parcels assessed have 0% tree cover 69,188 parcels have more than 20% tree cover 11,552 parcels have more than a 60% available planting space totaling 120 acres 25,482 parcels have less than 5% tree cover and more than a 25% total available planting space.
Figure 5: Existing, Possible Planting Area, and Unsuitable UTC for the District of Columbia.
District Tree Canopy Change Since 2006, tree canopy has expanded on average across the city from 14,246 acres (36%) in 2006 to 14,601 acres (37%) in 2011. Canopy cover increase resulted from tree growth and new tree plantings, countered only partially by natural mortality and some areas of clearing for land development.
Figure 6: Result of UTC Change Analysis in Acres
14
Urban Tree Canopy in the District of Columbia (2006 -2011)
Wards, at a glance…
Tree Canopy Metrics by Wards
Existing Canopy Percent Low – Ward 6 (22%) High – Ward 3 (53%)
Wards represent the most politically important assessment boundary in the District. Quite different from watersheds, Wards are drawn through political process and represent geographic areas where elected District councilmembers are responsible. This section quantifies Existing, Possible, Unsuitable UTC, and UTC change for D.C.’s eight Wards.
Possible Planting Area Percent Low – Wards 3 & 4 (22%) High – Ward 8 (34%)
Urban Tree Canopy and Possible Planting Area by Wards Results in Figure 8a (opposite) show that Wards 3, 4 and 7 currently contain canopy percent at or above the 2035 goal of 40%. More challenging are the Wards closer to downtown (Wards 1, 2 and 6) where roughly half of all land is unsuitable PPA and there is little opportunity to expand urban tree canopy.
Unsuitable Percent Low –Ward 3 (25%) High – Ward 6 (55%) UTC Change Percent Low – Ward 3 (-3.4%) High – Ward 2 (3.3%)
Table 3: Complete UTC results by Ward. Wards
2 2006 Acres UTC (excludes (Acres) water)
2006 UTC (%)
2011 UTC (Acres)
2011 UTC (%)
PPA Veg. (Acres)
PPA Veg. (%)
PPA Imp. (Acres)
PPA Imp. (%)
Total Total UTC UTC Possible Possible Change Change Planting Planting (Acres) (%) (Acres) (%)
Ward 1
1,581
320
20%
345
22%
257
16%
135
9%
392
25%
25
Ward 2
4,101
999
24%
1,135
28%
626
15%
329
8%
954
23%
137
3.3%
Ward 3
6,694
3,790
57%
3,560
53%
1,211
18%
235
4%
1,446
22%
-231
-3.4%
Ward 4
5,754
2,706
47%
2,882
50%
1,048
18%
214
4%
1,262
22%
177
3.1%
Ward 5
6,568
1,970
30%
2,007
31%
1,631
25%
316
5%
1,947
30%
37
0.6%
Ward 6
3,665
656
18%
764
21%
582
16%
321
9%
903
25%
108
2.9%
Ward 7
5,413
2,112
39%
2,150
40%
1,260
23%
313
6%
1,573
29%
38
0.7%
Ward 8
5,558
1,693
30%
1,757
32%
1,537
28%
327
6%
1,865
34%
64
1.2%
Total/Average
39,336
14,246
36.2%
14,601
37.1%
8,153
20.7%
2,190
5.6%
10,343
26.3%
355
0.9%
Figure 7: Existing canopy, planting space, and unsuitable urban tree canopy percent by Ward.
2
Reflects the area as determined in the 2002 DC Ward Spatial Data
15
Urban Tree Canopy in the District of Columbia (2006 -2011)
1.6%
Figure 8a: Percent of Existing Urban Tree Canopy (UTC) by Ward.
The District is comprised of 8 Wards with varying tree canopy cover as viewed from above, ranging from 22 to 53% of the total land area.
For each Ward, the combined space available for tree planting, that is the total grass, open space, and parking lot area, or Total PPA, ranged from 22% to 34%.
Figure 8b: Percent of Total Possible Planting Area (PPA) by Ward. 16
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figure 8c: Percent of Vegetated Possible Planting Area (PPA) by Ward.
For each Ward, the amount of grass and open areas where trees could be planted ranges from 15 to 28%. This is after removing ball fields and similar undesirable planting areas.
For each Ward, the total area of impervious surfaces ranges from 4 to 9% after removing buildings and roads where tree canopy could be established (e.g. parking lots).
Figure 8d: Percent of Impervious Possible Planting Area (PPA) by Ward. 17
Urban Tree Canopy in the District of Columbia (2006 -2011)
Tree Canopy Change by Wards Figures 9 and 10 illustrate total canopy evaluated by Ward for 2006 and 2011. Ward 3 contained the greatest canopy percent in both time periods, but was the only Ward to lose canopy. Ward 2 demonstrated the greatest increase (4%) while Wards 5 and 6 had only modest increases.
Figure 9: Percent in the change of tree canopy by Ward
60%
57%
53% 50% 50%
47%
39% 40%
40%
30% 32%
30% 31% 30%
28%
24% 20%
22%
2011 UTC (%)
21%
18%
20%
10%
0%
1
2
3
4
5
6
7
Figure 10: Comparison of percent UTC for 2006 and 2011 by Ward.
18
2006 UTC (%)
Urban Tree Canopy in the District of Columbia (2006 -2011)
8
Tree Canopy by Advisory Neighborhood Commissions The Advisory Neighborhood Commissions (ANCs) are the body of government with the closest official ties to the people among D.C.’s neighborhoods. Wards are divided into a total of 37 ANCs across the District. ANC’s are comprised of local neighborhoods clustered under a common legislature. Assessing UTC by ANC is important because many grass-roots policies and actions originating at the neighborhood level focus on urban greening and tree planting initiatives. Urban Tree Canopy and Possible Planting Area by ANC
ANCs, at a glance… Existing Canopy Percent Low – 1A (14%) High – 3G (65%) Possible Planting Area Percent Low – 2B (16%) High – 8C (37%) Unsuitable Percent Low – 3G (18%) High – 2B (66%)
Tree canopy and planting area percent are reported for ANCs and rights-of-way (ROW) within ANCs (see Figures 11 – 13). UTC Change Percent The highest canopy cover is in ANCs 4A, 3F, and 3G which all Low – 3D (- 8%) contain greater than 60% UTC, with 14 ANCs currently having High – 4C & 4D (8%) greater than 40% UTC. UTC and PPA percent specifically in the ROW by ANC is provided to highlight potential UTC in streets for each ANC.
Figures 11a & b: Existing canopy cover and possible planting area percent by D.C.’s ANCs. 19
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figure 12a: Percent of UTC by Rights-of-Way by ANC.
For ANCs, the percent of Existing Urban Tree Canopy (UTC), specifically within street rightsof-way, ranges from 20% to 56%.
For ANC’s, the percent of Total Possible Planting Area (PPA), specifically within street rights-of-way, ranges from 20% to 56%.
Figure 12b: Percent of PPA by Rights-of-Way by ANC. 20
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figure 13a: Percent of Possible Planting Area (PPA) Vegetation by ANC.
For ANCs, the amount of grass and open areas where trees could be planted ranges from 8% to 31%. This is after removing ball fields and similar undesirable planting areas.
For ANCs, the total area of impervious surfaces after removing buildings and roads where tree canopy could be established (e.g. parking lots) ranges from 2% to 12%.
Figure 13b: PPA Vegetation and PPA Impervious (right) by ANC. 21
Urban Tree Canopy in the District of Columbia (2006 -2011)
The highest canopy percentages are within ANCs around the periphery of the District, while values near the center tend to be lowest. Plantable space is often inversely related to canopy percentage, and this trend is reflected in ANC parcels. The greatest potential for increasing canopy will be targeting ANCs such as 8C, where low existing canopy (27%) is matched with high PPA (37%) (See Table 4 below). For many ANCs, canopy cover percent for the entire ANC is roughly equal to canopy cover within the rights-of-way (see Table 5 opposite page). However, ANCs in the center of the District that contain some of the lowest overall canopy percent also contain higher canopy percent within the rights-of-way. This is because DDOT/UFA plants aggressively in tree boxes as they are of but a few of the available planting spaces. Among ANCs with higher overall canopy, the opposite tends to be true where rights of way are below the area average. Table 4: Complete UTC assessment metrics by ANC. 3
3
Reflects the area as determined in the DC ANC Spatial Data
22
Urban Tree Canopy in the District of Columbia (2006 -2011)
Table 5: Complete UTC metrics by Rights of Way within ANCs.
Public rights of way provide valuable canopy cover and planting space for enhancing quality of life for the District’s businesses and residents.
23
Urban Tree Canopy in the District of Columbia (2006 -2011)
Tree Canopy Change by ANC Out of 37 ANCs, seven lost canopy by at least 1%, ten remained fairly constant (+/- 1%), and 20 gained at least 1% canopy cover. Several ANCs that lost canopy such as 3G, 3F, 3C, and 3D remain among the highest overall UTC percentages (ranging from 47-65%), however ANC 8A, which lost 6% canopy cover between 2006 and 2011, now contains only 30% canopy coverage.
Figure 14: Estimated percent gain or loss in tree canopy by ANC between 2006 and 2011.
24
Urban Tree Canopy in the District of Columbia (2006 -2011)
Tree Canopy by Subwatersheds The D.C. Metro Area is intersected by four (4) major watersheds named by the water body where incoming precipitation drains (Figures 15-17). For example, rain falling within the Rock Creek-Potomac River watershed will eventually flow into Rock Creek and the Potomac River. Given that tree canopy intercepts stormwater and filters pollutants before it reaches streams, measuring canopy cover by watershed helps direct tree management for water resources across the region. These are especially important considerations given the latest report by American Rivers designated the Potomac and Anacostia as America’s Most Endangered Rivers. UTC, PPA, and Canopy Change by Subwatershed Canopy cover by watershed illustrates that canopy percentages decrease from northwest to southeast in the District. Of the three Potomac River drainages, two average greater than 40% canopy cover. The Anacostia River drainage has the lowest overall canopy cover at 31%. Interestingly, the Oxon Run-Potomac River watershed has slightly higher canopy cover (32%) than the Anacostia River watershed, but also a larger percentage of plantable space (34%). (See Table 6 below and maps of existing UTC and PPA on the next page (Figures 16-17)). When averaged across the entire watershed, each basin gained canopy cover between 2006 and 2011 (Figure 15). The largest canopy gain was in the south (143 acres), while the two watersheds in the center of the District had no significant change (estimated increases of less than 1%).
Figure 15. Urban tree canopy change by subwatershed.
Table 6: Complete UTC results for the District’s subwatersheds. 4
4
Reflects the area as determined in the DC Boundary Spatial Data
25
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figure 16 a & b: Existing UTC ad Total Possible Planting Area percent by subwatershed. Total PPA summarizes vegetated areas such as grass and open space (PPA Vegetation) and impervious areas such as parking lots (PPA Impervious) where there is available space for tree planting.
26
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figures 17 a & b: Possible Planting Area (PPA) Vegetation (above left) and Possible Planting (PPA) Impervious (below right) by subwatershed. These figures represent the relative amount of available space for tree planting summarized at the watershed scale.
27
Urban Tree Canopy in the District of Columbia (2006 -2011)
Tree Canopy Metrics by Land Use and Zoning Categories Existing UTC & Possible Planting Area (PPA) was summarized for ten broad land use categories based on the Maryland Department of Planning’s protocol (see Appendix, page 47). Figure 18 shows the distribution of DC tree canopy and PPA. Below are selected UTC stats by land use:
Open urban land (parks and open spaces) accounts for the greatest proportion of tree canopy (5,231 acres, or 36% of all UTC); Transportation, including Rights-of-Way (ROW), ranks second with 3,284 acres of tree canopy, or 22.5% of all UTC.
More than 50% of High-Density Residential, Commercial, Mixed Use and Industrial land use areas are unsuitable for tree planting.
For Medium-Density Residential and Institutional land uses (combined 12,085 parcels), 3,383 have less than 10% existing UTC and greater than 20% PPA. While these land uses contain modest portions of total UTC (each comprises ~10% of the total UTC), they make up the greatest proportion of possible planting area (16% and 17% respectively).
Table 7: UTC Assessment Results by Land Use in the District based on 2011 imagery 5
Urban Tree Canopy Acres
Possible Planting Area Acres
Figure 18: Distribution of Existing UTC (left) and PPA (right) across ten land use categories. 5
Reflects the area as determined in the DC 2002 Land Use Spatial Data
28
Urban Tree Canopy in the District of Columbia (2006 -2011)
Table 8: UTC Assessment Results by Zoning Type in the District based on 2011 imagery. 6
Tree Canopy Change by Land Use Tree canopy change was statistically significant for medium density residential, open urban land and high-density residential which expanded by 2.6%, 1.7%, and 1.5% respectively (Table 7). Low-density residential (see density definitions in the Appendix, page 45) indicated a loss of 29 acres (-0.6%) in the five-year span. Given these metrics rely on the accuracy of both 2006 and 2011 canopy data sets which have an inherent margin of error, canopy trends should be considered as approximations rather than values of exact or precise change. At the parcel-level, features can be queried by land use type and/or UTC and PPA metrics (area or percent) and then symbolized by color ramps of UTC and PPA percentages. An example is on the following page, where GIS queries were used to only show residential properties and then symbolize the parcel boundary by existing tree canopy percent (Figure 19). 6
Reflects the area as determined in the DC Zoning Spatial Data
29
Urban Tree Canopy in the District of Columbia (2006 -2011)
Parcel boundaries
Parcel boundaries
Figure 19: Existing UTC for residential areas by Ward. Percent labels show the average UTC per Ward for residential land use only. Underlying parcel metrics are also shown. 30
Urban Tree Canopy in the District of Columbia (2006 -2011)
Tree Canopy Metrics by Census Block Every ten years, the U.S. Census Bureau provides a population estimate for the United States. Although data are collected house to house, to protect our anonymity, census data are summarized by several different geographic scales. Census blocks are the smallest (most specific) division of this demographic data released by the Census Bureau, and are generally similar to city blocks - drawn by intersecting roads or natural barriers. There are 6,507 census blocks reported across the District. These provide a link to socio-economic and demographic conditions in the District as well as a blocklevel scale for prioritizing tree planting and monitoring canopy cover change. UTC and PPA by Census Block
Census Blocks, at a glance… Existing Canopy Percent 1,784 Blocks above 40% 1,060 Blocks below 10% Possible Planting Area Percent 645 Blocks above 40% 786 Blocks below 10% Unsuitable Percent 3,575 Blocks above 40% 320 Blocks below 10% UTC Change Percent 747 Blocks gained > 10% 366 Blocks lost > -10%
Figure 20 illustrates current canopy cover percent for all census blocks. The large number of census blocks in the study area hampers visual interpretation. Subset examples of PPA impervious and PPA vegetation are provided to better illustrate the results at a larger scale. Canopy Change by Census Block More than half (3,448) of all census blocks have gained at least 1% canopy, 1,077 have remained fairly constant (changing by less than +/- one percent), and 1,982 blocks have lost at least 1% canopy between 2006 and 2011. Figure 20: Existing UTC Percent per Census Block. Note the fine resolution of Census Blocks as an assessment boundary for targeting neighborhood tree preservation and planting.
31
Urban Tree Canopy in the District of Columbia (2006 -2011)
Tree Canopy by Parcel The District contains 134,323 individual property parcels. Parcels provide the finest detail of UTC assessment in this analysis and report. Once UTC and PPA is evaluated for Wards and ANCs for broader planning and comparison purposes, the small size of parcels as an assessment boundary make them ideal for targeting individuals / land owners as private land is a target for implementation. Figures 21 a-d show the level of detail through a parcel-based UTC analysis.
Figure 21 a-d: UTC and PPA percentages for selected parcels (eastern side of D.C.)
32
Urban Tree Canopy in the District of Columbia (2006 -2011)
Putting Parcel-Level UTC Data to Work As described above, simple GIS queries can be used to filter and highlight specific tree planting opportunities using the data from this assessment. Two examples are provided here. Figure 22 a: At right, parcels colored orange have less than 25% tree canopy and have 25% or more possible planting area, resulting in nearly 54,000 parcels.
Figure 22 b: At left, a more constrained query identifies the parcels with the highest potential for expanding canopy. Parcels have less than 10% existing canopy and at least a 50% planting area, resulting in just over 15,000 properties.
33
Urban Tree Canopy in the District of Columbia (2006 -2011)
Putting Parcel-Level Data to Work (Continued) Trees in and around parking lots provide maximum environmental benefits: 
Interception of stormwater runoff and filtration of pollutants

Shade prolongs pavement life and lowers ambient air temperatures, reducing the formation of ozone (O3)

Trees beautify often harsh landscapes, making vibrant environments for business and retail
The map below in Figure 23 shows 859 parcels in the District (red polygons) with the best tree planting potential near parking lots.
Figure 23: Parcels (red polygons) shown at right have less than 20% tree canopy cover, more than a 50% available planting space, and have one or more parking lots.
34
Urban Tree Canopy in the District of Columbia (2006 -2011)
Reasons for Gains and Losses of Urban Tree Canopy in the District Understanding the reasons for canopy gains and losses is essential to meet the District’s canopy goal of 40% by 2035 and to improve urban forest management and policy planning. A quantitative, statistical assessment of tree canopy gains and losses was performed for 100 selected parcels to validate the accuracy of canopy change and determine the reason for change. Each parcel was visually inspected for both classification quality and observed reason for a reported canopy over change. Table 9 summarizes the four categories of qualitative gains and losses that were studied. Table 9: Types of canopy gains and losses in the District. Type of Change
Description
Significance/Comments
Loss: Development
Trees or forests removed for another land use.
Several examples were located in the District. Figure 24 below illustrates an example in the northwestern part of the District
Loss: Removals/Mortality
Natural mortality, storm events, power line vegetation clearance, disease and invasive species.
Many examples of large, single tree losses can be visually identified across the District.
Gain or Loss: Misclassification
Differences in aerial or satellite imagery pixel size, classification methods and quality assurance / quality control, or time of image capture can lead to a difference of results between time periods.
A detailed visual re-evaluation indicated an overall underclassification of 2006 UTC prior to re-analyzing. Manual and automated efforts to correct results between the two time periods improved comparisons but differences persisted. An example is provided in Figure 25.
Gain: new tree plantings, growth of existing trees and policy/regulatory controls
Landscaping codes, tree protection ordinances, public campaigns, wetland restoration, UFA’s aggressive rights-of-way planting policy, and D.C.’s Urban Forest Project contribute to expanding canopy cover.
Over time, tree plantings must keep pace with sources of canopy loss to meet the City’s canopy goals. Figure 25 below illustrates increasing canopy cover.
35
Urban Tree Canopy in the District of Columbia (2006 -2011)
The 100 parcels for the statistical evaluation were selected using the following five criteria: 1. The top 10 parcels ranked by total canopy area gain. 2. The top 10 parcels ranked by canopy percent gain. 3. The top 10 parcels ranked by total canopy area lost. 4. The top 10 parcels ranked by canopy percent lost. 5. The remaining 60 randomly selected parcels: a. Were larger than 100 square meters (to avoid erroneous sliver polygons). b. Had losses or gains greater than 10% of the total parcel area. Results are shown in Figure 24 below. Of the 100 parcels examined, 55 represented gains, and 45 represented losses. Misclassification produced 13% of false gains and 20% of false losses. Roughly two-thirds (67%) of canopy gains were due to growth (expansion) of existing tree canopy and new tree planting. Spatial shifts between imagery data mainly has an impact at the parcel-level where shifts can cause false gains/losses but is averaged across larger geographic assessment units (e.g. neighborhoods or city-scale boundaries). Reasons for Canopy Gain
Reasons for Canopy Loss
1.
18%
Tree Canopy Expansion
60%
31% Development
Tree Planting
Individual Tree Removal
Misclassification 20% Spatial Shift
Misclassification Spatial Shift
31%
Figure 24: Pie charts illustrate the distribution of reasons for Urban Tree Canopy gain (left) and losses (right) as percentages of examined parcels. Detailed methods can be found in the Appendix along with a discussion regarding canopy change results from differing imagery sources, classification techniques, and Quality Assurance / Quality Control (QA/QC) methods.
36
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figures 25 and 26 here provide aerial mapping examples of gains and loss from around the District. The areas shown were easy to locate by querying UTC metrics in the parcels layer.
Tree Canopy in 2006 Tree Canopy in 2011
This example shows a loss of urban tree canopy in the District during construction of a development. With the 2006 GIS tree canopy layer underneath the 2011 canopy data, it is clear to map and quantify the loss. Development and redevelopment can result in the concentrated losses of forests, large groups of trees, or individual trees. Also contributing to canopy loss, naturally occurring tree mortality from extreme weather, old age, insect damage and disease impact the District’s canopy cover goals.
Figures 25 a-c: Canopy loss due to development. Top left image shows areas of canopy in 2006 in yellow that were cleared during development and therefore not mapped in the 2011 imagery data. 37
Urban Tree Canopy in the District of Columbia (2006 -2011)
This example shows UTC gains from tree growth and new tree planting. While losses in canopy cover are easy to see and map, gains are difficult and more subtle. The images at bottom show dozens of trees with steady tree crown growth. When aggregated across the District, this can explain how UTC may increase even with losses from development. Education, outreach and stewardship programs reaching private property owners about tree benefits and the District’s 40% canopy goal are likely the most cost-effective way to meet goals.
2006
2011 Figures 26 a-b: An example of tree canopy growth (UTC gains) in the District.
38
Urban Tree Canopy in the District of Columbia (2006 -2011)
Recommendations Based on this assessment, the following general recommendations are offered for achieving tree canopy goals in the District of Columbia. 1. Locally and regionally, the District of Columbia government, MWCOG and others should use this report to guide: o Community forestry projects (new development, public street and park trees, tree protection); o Develop best management practices (BMPs) such as maintaining higher canopy cover in sensitive areas (wetlands), parking lots, schools, and more. 2. Target specific land use types at the parcel and/or ANC-level that are most in need of additional tree planting or have experienced the greatest losses. Specifically, Mixed Use and Industrial land uses have the lowest average canopy cover. 3. Conduct an audit of local policies to simplify or streamline tree-related policies and identify where codes are working against local goals. Additionally: o Collaborative planning across Wards can reduce costs and provide consistency for public works officials, planners, and stormwater and resource managers; 4. Create a regional UTC assessment report and plan with many partners (i.e., City of Alexandria, VA, Arlington County, Prince George’s and Montgomery County, MD, etc.) that make up the Metro D.C. Area. 5. Analyze the District’s UTC every 7-10 years to monitor trends and assess all efforts. o Ensure consistency by using comparable image resolution and classification and QA/QC techniques. LiDAR and sub-meter imagery acquired at similar times provides ideal data. Given large bodies of water in the District, it is critical to map water consistently calculating UTC percent is based on land area. 6. Utilize the land cover data from this project while it is current for other applications involving water supply planning, stormwater modeling and air quality monitoring. 7. Recommend that all universities, schools and institutional campuses in the District establish a forest management plan as their contribution towards the District’s 2035 UTC goal of 40%. 8. Explore all potential partnerships to achieve urban forest goals: public/private including corporate and academic sponsors, council representatives, air quality and stormwater associations, volunteers, non-profit organizations, and neighborhood associations. 9. Overlay PPA and redevelopment areas to identify green space creation with tree planting opportunity areas. 10. Promote hardy, climate-adapted, and long-lived tree species that are appropriate for the District’s environment to insure investments in trees achieve maximum benefits. 39
Urban Tree Canopy in the District of Columbia (2006 -2011)
11. To reach the District’s 40% UTC goal, use the Canopy Calculator tool provided in this study to discuss and set goals by land use types, starting with the public rights-of-way. 12. Use results to justify targeted public tree plantings based on existing canopy cover within the public rights-of-way broken out by ANC boundary. Next, within each ANC, prioritize planting at the block level including sites that provide optimal shade (west side of buildings), have low existing UTC and high impervious surface area. 13. Ground-truth Possible Planting Areas (PPA) to translate PPA area (acres or square feet) into the number of realistic planting locations. Then develop a national guide for other communities to use this information on their canopy assessment projects. 14. Identify private properties with planting space adjacent to streets where DDOT planting spaces are nonexistent. Offer tree subsidies in these areas similar to (or via) the District Department of Environment RiverSmart program (ddoe.D.C..gov/riversmarthomes). 15. Use the suite of planning/planting tools to meet D.C. MS4 permit requirements for planting 4,000 trees by 2017/2018. 16. Create a central repository for monitoring tree planting and tree removals on public and private property, possibly using a Web-based application that is open to the public. 17. Plant all available street rights-of-way sites. 18. Address the loss of remaining intact forested tracts in two ways: (i) determine reasons and preventive policy solutions to recent losses identified in this project, and (ii) prioritize areas for preservation and restoration based on existing zoning plans and maximizing ecological, social and economic benefit. Use this information to promote new development in areas such as previously disturbed sites instead of forested sites. 19. Use UTC data to bolster the D.C. DOT education and outreach campaign. Using ANC, land use types, and parcel data, maps could help D.C. government agencies, non-profits, residents and businesses visualize their role in reaching the District’s canopy goals. 20. Use urban forestry goals and design specifications to enhance other environmental planning initiatives such as Sustainable D.C., wetland restoration projects, open space conservation easements, green infrastructure & low impact development (LID) plans, and energy efficiency programs. 21. Coordinate with large property owners (golf courses, institutions, campuses) through public-private partnerships to reach tree canopy goals and desired ecosystem service targets (greenhouse gas emissions, stormwater infiltration, and water quality).
40
Urban Tree Canopy in the District of Columbia (2006 -2011)
Summary Urban forests are dynamic resources that are under constant pressure from natural processes and human interaction. Effective management of urban forests requires a current tree inventory and established planting programs, where trees may be planted equitably to maximize benefits within a restrictive regulatory and physical framework that may impede goals. While methods and results vary slightly from recent assessments, there is consensus that the District’s canopy cover is between 36 and 37 percent based on 2011 imagery. New development has resulted in large canopy losses but also been offset to some degree by the growth of existing trees and robust tree planting efforts over the past 10-20 years. The following items provide a recap of some of the highlights from this study:
Data from this study indicates a potential slight gain in canopy cover from 2006 to 2011, with seven out of eight Wards gaining or maintaining canopy.
Ward three lost roughly 3% canopy cover over the last several years.
Medium density residential land uses lag behind lower density residential lots in canopy cover (22% vs. 49%, respectively), and have similar space available for enhancing canopy cover (35% and 32%, respectively).
One out of six census blocks has less than 10% canopy cover, many in Wards five, seven, and eight outside of the core downtown area.
36% of all tree canopy in the District is found on Open Urban Land (parks/open spaces).
Technology As with any remote sensing assessment tool, limitations exist with UTC science and technology. This study and those from others interested in advancing the UTC assessment practice strengthen our understanding of the process, outcomes, and potential applications for implementation of UTC goals. There are numerous factors that can significantly influence tree cover accuracy, and canopy cover percent is rounded to the nearest whole number rather than to the 10th of a percent for this reason. Examples include image quality, automated analysis and manual interpretation techniques, GIS boundaries used and canopy comparison methods. On future periodic assessments, techniques for monitoring gains and losses need to be consistent and include strict quality control / quality assurance methods to reliably track change.
41
Urban Tree Canopy in the District of Columbia (2006 -2011)
Further research is needed to identify the recommended time interval and method(s) available to reliably assess tree cover change. The discussion in this report offers communities insight into assessment methods and analysis options. The Future of Canopy in the District While the exact percent of UTC in the District can change almost daily, the accuracy and relative nature of available planting areas mapped and presented in this report have been thoroughly analyzed. Medium density residential, open urban land, and institutional land uses have the greatest potential to positively impact the District’s canopy goal. By drilling-down to the block-level via these land uses (Wards, ANCs), opportunities can be further prioritized in objective ways. The District intends to put the data and tools from this assessment to use, create education/outreach campaigns, forge new partnerships, and conduct interdisciplinary management and implementation planning. Balancing new development with the protection and conservation of environmental values related to forest cover such as wetland habitat, air quality, and climate adaptation related to carbon storage and energy conservation will be an ongoing work item. The economic benefits of urban tree canopy are an incentive to continue in this direction. Our Nation’s capital boasts roughly 37% urban tree canopy, at or above many other eastern U.S. cities (see Figure 27 below). Reaching and sustaining 40% overall canopy is a realistic, attainable, and worthy goal to continue exemplifying what is achievable in our cities. It demonstrates that quality of life and urban forestry goals are interrelated of equal importance to decision-makers and residents in the Nation and our Nation’s capital.
Figure 27: Comparing urban tree canopy cover in the District of Columbia to other northeastern U.S. cities.
42
Urban Tree Canopy in the District of Columbia (2006 -2011)
page left intentionally blank
43
Urban Tree Canopy in the District of Columbia (2006 -2011)
page left intentionally blank
44
Urban Tree Canopy in the District of Columbia (2006 -2011)
Appendix Included below are additional details on the methods and processes used, a list of deliverables, tables not included above, and additional examples of some of the products provided. Land Cover Classification Detailed Procedures Remote sensing and GIS tools were used to map eight land cover classes across the District. A Geographic Object-Based Image Analysis (GEOBIA) was performed using Feature Analyst version 5.0 to map five land cover classes (Tree canopy, non-forest vegetation, impervious surfaces, bare soil, and water) from the 2011 NAIP imagery, and to map complete impervious surfaces from the 2011 leaf-off imagery. Manual Quality Assessment/Quality Control (QA/QC) was then conducted to enhance the accuracy of the 2011 canopy cover. Several vector layers were overlaid with the classification to refine classes. Impervious surfaces mapped through the GEOBIA procedure were intersected with roads and buildings layers provided by D.C. GIS to create sub-categories of impervious areas. The “Traffic Circle”, and “Median” road type classes were extracted from the roads layer and classified as possible plantable area vegetation (see UTC Classification Detailed Procedures below). The mapping of wetlands and water were enhanced by GIS layers provided by D.C. GIS, the National Wetlands Inventory (NWI) dataset, and through manual editing based on 2011 imagery. Wetlands were mapped by selecting wetlands from the “Wetland Types” attribute and manually comparing against the 2011 imagery. Furthermore, the “Wetland Restoration AW” GIS layer provided by D.C. GIS was added to the wetlands land cover class. A water layer was also provided by D.C. GIS, and was modified to enhance quality and accuracy for the final land cover water class. The provided water layer was enhanced through; (a) manually adding water digitized from 2011 leaf-off imagery, (b) adding water beneath bridges and roads, and (c) manually modifying “Lake” and “Freshwater Ponds” polygons from the NWI dataset. Water was mapped as the land cover where wetlands and water overlap. Deriving UTC Metrics from Land Cover Inputs Urban Tree Canopy metrics (Existing UTC, Possible Planting Area, and Unsuitable UTC) were derived by combining the eight (8)-class land cover with selected layers to create a twelve (12) class UTC classification. Existing UTC is identical for both classifications. Possible Planting Area (PPA) is a combination Impervious Areas (IA) and non-forest vegetation where it is biophysically possible and feasible to plant trees. Suitable IA includes all impervious surfaces other than roads, transportation corridors, tennis courts, playgrounds, and buildings (mostly parking lots and sidewalks). Roads, buildings, sidewalks, and transportation corridors were provided by D.C. GIS, and overlaid with the land cover classification. Additional building areas, tennis courts, and playgrounds were manually digitized, and then removed from the suitable impervious class. Suitable vegetation includes all non-forest vegetation, excluding areas where it was considered biophysically not feasible (wetlands) or unacceptable (recreational fields, golf courses, and the National Mall) to plant trees. Recreational fields (baseball, softball, rectangular fields with evidence of sporting activities, and running tracks) were manually identified, (mapped from the
45
Urban Tree Canopy in the District of Columbia (2006 -2011)
2011 NAIP imagery), and then excluded from PPA vegetation areas. For golf courses all 2011 fairways were mapped and excluded. The National Mall contains areas where open spaces are intentionally maintained. Therefore, several areas were excluded from the PPA vegetation UTC metric. Excluded mall areas include vegetated areas within and to the north and west of the Lincoln Memorial, Gorla Tree Plain Park, large open area east of the Vietnam Veterans Memorial, area south of the Reflecting Pool, area around the Washington Monument, The Elipse, north to and including the White House lawn (north and south), south to the field west of the United States Holocaust museum, and the center grassy fields east to and including the Capitol and surrounding lawns. Additionally, the Thomas Jefferson Memorial lawns were excluded in the south. Re-Analysis of 2006 Tree Canopy Assessing canopy cover change down to the individual parcel-and-block-level requires a highly accurate canopy classification, consistent imagery & methods, and strict, time-consuming QA/QC procedures. The District’s previous GIS canopy data layer based on 2006 QuickBird imagery was reviewed alongside the 2011-based canopy layer as a task of this project. It was determined that additional QA/QC would improve the accuracy and reliability of the canopy cover change metrics between the two time periods. Two (2) steps were performed. First, a manual systematic GIS editing process was conducted across the District to remove erroneous polygons. These included polygons of grass, shrub, shadows, building edges and areas where the District’s street tree inventory points were buffered but a vacant tree site existed; thus overestimating canopy in these ‘ghost’ tree polygons. During this manual QA/QC step, some tree canopy was also added that was missing in the initial automated canopy layer. Secondly, an automated classification was performed to extract tree polygons in the eastern half of the District where enough tree canopy was underestimated during the prior classification that an automated process would be needed. After several iterations to fine tune the extraction, the additional classified canopy polygons went through a QA/QC process and were merged with the edited original 2006 tree canopy layer. The net result after removing and adding canopy to the 2006 tree canopy layer was an increase from 13,673 acres to 14,245 acres, or 572 acres (1.5% of the city’s land area). Results showing the data before and after this re-analysis of 2006 tree canopy are shown below. On page 47 and 48 imagery (b) with original 2006 canopy, and (c), original 2006 tree canopy (green) overlaid with updated canopy layer (yellow).
46
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figures 28 a, b, c: Example of underestimation; (a) 2006 satellite imagery, (b) original 2006 UTC, and (c) original 2006 UTC (green) overlaid with updated UTC (yellow).
A
B
C
47
Urban Tree Canopy in the District of Columbia (2006 -2011)
Figures 29 a, b, c: Example of overestimation; (a) 2006 satellite image, (b) original 2006 UTC, and (c) updated UTC (yellow) where green areas were removed due to overestimation.
A
B
C
48
Urban Tree Canopy in the District of Columbia (2006 -2011)
Land Use The 2004 District of Columbia existing land use spatial data contained more than 20 “land use” codes. MWCOG collapsed the use codes into ten broad categories so that they could be easily understood and interpreted by project stakeholders. The 2002 Maryland Department of Planning’s land use / land cover classification scheme (which includes 141 categories) was used as a template. MWCOG provided these ten unique broad land use types for this UTC study. The Maryland Land Use / Land Cover Classification land use descriptions are as follows: 11 Low-density residential - Detached single-family/duplex dwelling units, yards and associated areas. Areas of more than 90 percent single-family/duplex dwelling units, with lot sizes of less than five acres but at least one-half acre (.2 dwelling units/acre to 2 dwelling units/acre). 12 Medium-density residential - Detached single-family/duplex, attached single-unit row housing, yards, and associated areas. Areas of more than 90 percent single-family/duplex units and attached single-unit row housing, with lot sizes of less than one-half acre but at least one-eighth acre (2 dwelling units/acre to 8 dwelling units/acre). 13 High-density residential - Attached single-unit row housing, garden apartments, high-rise apartments/condominiums, mobile home and trailer parks. Areas of more than 90 percent high-density residential units with more than 8 dwelling units per acre. 14 Commercial - Retail and wholesale services. Areas used primarily for the sale of products and services, including associated yards and parking areas. 141 Mixed use – Predominantly commercial/residential areas (COG included this LU type). 15 Industrial - Manufacturing and industrial parks, including associated warehouses, storage yards, research laboratories, and parking areas. 16 Institutional - Elementary and secondary schools, middle schools, junior and senior high schools, public and private colleges and universities, military installations (built-up areas only, including buildings and storage, training, and similar areas), churches, medical and health facilities, correctional facilities, and government offices and facilities that are clearly separable from the surrounding land cover. 18 Open urban land - Urban areas whose use does not require structures, or urban areas where non-conforming uses characterized by open land have become isolated. Included are golf courses, parks, recreation areas (except areas associated with schools or other institutions), cemeteries, and entrapped agricultural and undeveloped land within urban areas, and wooded areas. 50 Water - Rivers, waterways, reservoirs, ponds, bays, estuaries, and ocean. 80 Transportation - Miscellaneous transportation features not elsewhere classified. Includes park and ride facilities and transit stations.
49
Urban Tree Canopy in the District of Columbia (2006 -2011)
Complete List of Deliverables The following products were included on DVD in North American Datum, Maryland State Plane (FIPS 1900) in ArcGIS v10 geodatabase format and are available from D.C. GIS (D.C.gis.D.C..gov/).
Land cover data (12-class UTC, 8 class land cover, comprehensive impervious layer, and re-analyzed 2006 tree canopy land cover layer
2011 4-band NAIP aerial imagery mosaic (raster)
UTC assessment boundaries (GIS vector format, ArcGIS v10 geodatabase)
UTC assessment spreadsheet and Canopy Calculator Tool (MS Excel spreadsheets)
Report summarizing the results, methods, and data (MS Word and Adobe PDF format)
Interactive PDF maps for the District (1) and each neighborhood (37)
CommunityViz urban tree canopy scenario planning tools (.mxd and geodatabase)
UTC GIS Database Format All UTC GIS data layers (Wards, ANCs, land use, subwatersheds, census blocks, parcels, and rights-of-way) were delivered with UTC fields and metrics as shown below.
Figure 30: UTC Assessment GIS Database Field Key
50
Urban Tree Canopy in the District of Columbia (2006 -2011)
Examples of Other Tools Provided in this Study Interactive PDF maps were provided to DDOT UFA for use by stakeholders without GIS software or expertise. Map layers appear in a Table of Contents at right. They can be turned on and off to show UTC results against the imagery backdrop. Several versions were provided for each Ward.
Figure 31: A “low-tech” interactive PDF map of UTC and planting areas A “Canopy Calculator” spreadsheet tool allows a user to see the effects of tree planting on canopy goals, or to determine the number of trees required to meet certain goals. Given the District’s 40% overall UTC goal, this mock-up shows the canopy cover percent that each land use type would need to have relative to the land area of each type. It lists the number of trees needed to reach the goal for each land use type and the total needed (60,278). In the top right corner, average crown spread at maturity (crown diameter in feet) is determined by entering the proposed percent distribution of small, medium and large sized tree species. A mortality
51
Urban Tree Canopy in the District of Columbia (2006 -2011)
rate of 10% was used as an assumption value for the percent of newly planted trees that would not make it to maturity. Note that loss of existing UTC is not represented in the Calculator.
Figure 32: AMEC’s Canopy Calculator, a Plug and Play Spreadsheet Tool to assist in Urban Tree Canopy (UTC) Goal Setting.
52
Urban Tree Canopy in the District of Columbia (2006 -2011)
Assessing Reasons for Tree Canopy Gains and Losses in the District of Columbia, 2006 – 2011 Introduction Recent advances in technological approaches and available data have allowed for highly accurate and precise UTC mapping. Understanding how forests are changing across time and space requires UTC assessments to be conducted and compared over multiple time periods. Comparisons of multiple UTC assessments for a single area are sometimes complicated by differences in mapping inputs (i.e. available imagery and technical approach). Since precise differences in UTC can be measured and reported, change results must be assessed to evaluate how a variety of factors contribute to overall reported UTC change. This analysis examined the reasons for tree canopy change (~355 acres) between UTC assessments mapped in 2006 and 2011. It is important to consider that the reported change is a product of; (1) actual change that has occurred (tree plantings and growth, and mortality and development), (2) differences in how actual tree canopy was mapped and presented, and/or (3) spatial shifts in the data of the classified UTC years. To understand the relative contributions to the overall change value, a statistical assessment of sampled parcels was conducted, categorically summarizing the reasons for measured losses and gains across the District. Methods To assess reasons for gains and losses, 100 parcels from across the District were selected and evaluated for classification quality and the reason for reported change between 2006 and 2011 (selection criteria listed below). Metropolitan Washington Council of Governments (MWCOG) and District Urban Forestry Administration staff performed ground-truthing for selected parcels to verify the reliability of electronic data and to understand specific reasons for canopy gains and losses. Parcel Sampling Strategy: A total of 100 parcels were manually inspected for both classification quality and observed for reported canopy change. The sample parcels were selected using five criteria: 1. The top 10 parcels ranked by total canopy area gain. 2. The top 10 parcels ranked by canopy percent gain. 3. The top 10 parcels ranked by total canopy area lost. 4. The top 10 parcels ranked by canopy percent lost. 5. 60 randomly selected parcels with the following constraints: - Selected parcels were required to be larger than 100 square meters. - Selected parcels were required to have losses/gains greater than 10% of parcel. Additionally, after these 100 parcels were assessed, their locations were visually examined to ensure recognizable trends were not solely data driven (e.g. all changes due to spatial shift were not in one area of the District). This is important since the reasons could be due to a 53
Urban Tree Canopy in the District of Columbia (2006 -2011)
spatial phenomenon (i.e. development, or neighborhood related tree planting program), or simply because imagery is poor in a particular location leading to imprecise classification. Assessment Procedure: Parcels containing both the 2006 and 2011 UTC and change information were examined according to the criteria stated above. For criteria one through four, parcels were sorted either ascending or descending by UTC change in square feet or percent. For criterion five, parcels were selected by: (1) assigning a random number to each parcel, (2) sorting by the random number, and (3) starting at the lowest random value, 60 parcels that are larger than 100 square meters with a gain or loss greater than 10% were evaluated. Selected parcels were evaluated using the 2006 and 2011 UTC data and imagery. Next, the overall reasons for measured gain or loss were recorded. For select parcels, physical addresses were provided to MWCOG and DC Urban Forestry staff to field-verify agreement between electronic sources and conditions on the ground. Gains and Loss Reporting Measured canopy gains and losses resulted from four main reasons. Of the 55 sample parcels demonstrating canopy gain, 33 (60%) were the result of canopy expansion or growth of existing trees, four (7%) were the result of tree planting, seven (13%) were the result of misclassification in either 2006 or 2011, and 11 (20%) were the result of spatial shifts between data in the two assessment years. Of the 45 sample parcels demonstrating canopy loss, 14 (31%) were the result of major development or clearing of existing trees, 14 (31%) were the result of individual tree removal, nine (20%) were the result of misclassification in either 2006 or 2011, and eight (18%) were the result of spatial shifts between data in the two assessment years (A table with descriptions of all 100 parcels is available with other deliverables, including a Geoprocessing Identifier (GID) that can be used to link table results directly to the District’s parcels). The first two categories for gains (expansion and planting) and losses (development and removal) represent parcels demonstrating real forest change while the second two categories (misclassification and spatial shift) represent canopy change as a result of data related observations. Combining the first two categories demonstrates that real canopy gains existed in 67% of parcels while real losses were present in 62% of all sampled parcels. Misclassification and spatial-shifting combined accounted for a greater proportion of parcels demonstrating canopy loss (38%) over canopy gain (33%). These categories represent the most significant cause of canopy gain or loss, but a combination of impacts was present for many parcels. What does this analysis say about the usefulness of measured gains and losses 2006 to 2011? The impact of spatial shift on measured canopy gains or losses is mitigated by comparing time periods across larger geographic scales (such as Wards or the entire District). Ignoring observations resulting from spatial shifts, of the 81 remaining parcels, 28 (35%) demonstrated canopy loss, 37 (46%) demonstrated canopy gain, and 16 (20%) resulted from misclassification. Given the unbiased methods used to select parcels, this analysis supports the original report findings that overall canopy across the District increased between 2006 and 2011. 54
Urban Tree Canopy in the District of Columbia (2006 -2011)
55
Urban Tree Canopy in the District of Columbia (2006 -2011)