Mapping ecosystem services in New York City: Applying a social–ecological approach in urban vacant...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

http://www.elsevier.com/authorsrights

Author's personal copy

Mapping ecosystem services in New York City: Applyinga social–ecological approach in urban vacant land

Timon McPhearson a,n, Peleg Kremer a, Zoé A. Hamstead b

a Tishman Environment and Design Center, The New School, 79 Fifth Avenue 16th Floor, New York, NY 10003, USAb Environmental Policy and Sustainability Management, Milano School of International Affairs, Management and Urban Policy, The New School, 72 FifthAvenue 5th Floor, New York, NY 10011, USA

a r t i c l e i n f o

Article history:Received 26 December 2012Received in revised form30 May 2013Accepted 24 June 2013Available online 7 September 2013

Keywords:Ecosystem servicesStackingSocial–ecologicalUrbanNew York CityVacant land

a b s t r a c t

As urbanization expands city planners and policymakers need to consider how ecological resources canbe strategically developed and managed sustainably to meet the needs of urban populations. Theecosystem services (ES) approach provides a useful framework for assessing the status quo, setting goals,identifying benchmarks and prioritizing approaches to improving ecological functioning for urbansustainability and resilience. However, new tools are required for comprehensively evaluating urban ESfor ecosystem management and to understand how local and regional trends and plans may affect ESprovisioning. We develop an ES assessment methodology that can be used to assess multiple ES of urbangreen space and integrate them with social conditions in urban neighborhoods. Our approach considerssocial–ecological conditions and their spatial patterns across the urban landscape. Our analysis focuseson vacant land in New York City. Results suggest that a combined social–ecological approach to ESassessment yields new tools for monitoring and stacking ES. We find that clusters of vacant lots in areaswith overlapping low ecological value (e.g. low concentration of green space) and high social need for ES(e.g. high population density) are primarily concentrated in three areas of the city – East Harlem, SouthBronx and Central Brooklyn.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

The world is increasingly urban, interconnected and changing(Seto et al., 2011). Over the last few decades there has beenincreasing recognition that human population expansion anddevelopment, especially in cities, is reshaping the ecology of theentire planet (Alberti et al., 2003; Folke et al., 2003; Rockströmet al., 2009a, 2009b). Urban regions create significant dispropor-tionate direct and indirect environmental impacts at the local,regional and global scale that affect local and global sustainability(Grimm et al., 2008, 2000; Seto et al., 2012). Given globalurbanization trends compounded by the effects of climate changeand other global environmental pressures (IPCC, 2011; Rockströmet al., 2009a, 2009b), a critical dynamic that must be understoodfor increasing urban sustainability and resilience is the social–ecological relationships between humans and the urban ecosys-tems in which the majority of people live (Folke, 2006; Pickett andGrove, 2009).

Local and regional urban ecosystems provide important func-tions that benefit urban residents including habitat for biodiversity,

primary productivity, stormwater retention, air pollution removaland heat mitigation (Bolund and Hunhammar, 1999). The ecosys-tem services (ES) approach provides a useful framework forassessing the status quo, setting goals, and identifying benchmarksthat facilitate long-term monitoring and prioritizing approaches toenhance ecological functions in ways that serve urban commu-nities (Daily et al., 2009; Niemelä et al., 2010; Sukhdev et al., 2010).In particular, spatially-explicit tools are needed for decision-makers to consider how social–ecological characteristics constrainsite suitability for restoring or improving the production of crucialES (Chan et al., 2006; De Groot et al., 2010a, 2010b; Seto et al.,2012). Here, we develop an ES assessment methodology, whichconsiders social–ecological conditions and relationships betweenmultiple ES, as well as spatial patterns of these conditions andrelationships across the urban landscape. We present a social–ecological analysis that focuses on New York City (NYC) vacantlots, understudied areas of the city, which by virtue of beingunderdeveloped, hold potential as spaces for transformation toimprove ES and meet social need for ES (Kremer et al., in press).Goals of this study were to: (1) develop a conceptual frameworkfor mapping the spatial patterns of multiple ES of vacant land andsocial need for ES in NYC; (2) offer an empirical example of howthis framework can be applied in the urban context using simpleindicators and available data; (3) exemplify a spatially explicit ES

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/ecoser

Ecosystem Services

2212-0416/$ - see front matter & 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.ecoser.2013.06.005

n Corresponding author. Tel.: +1 212 229 5321x3220.E-mail address: [email protected] (T. McPhearson).

Ecosystem Services 5 (2013) e11–e26


stacking methodology for evaluating the combined ES of a givenlot or patch; (4) investigate how landcover and other environ-mental characteristics relate to the quantity and quality of differ-ent ES provisioning across the urban landscape at the city scale;and (5) identify patterns and clusters of vacant lots exhibitingsimilar social–ecological characteristics (i.e. areas of high socialneed and low ecological values).

2. Urban social–ecological systems

Urban areas are made up of complex combinations of heteroge-neous social–ecological patches (Cadenasso et al., 2007). In coupledsocial–ecological systems such as urban areas, mutual dependenceexists between social communities and ecological processes withinteractions and feedbacks affecting each other over time (Folke,2006; Holling, 2001; Peterson, 2000). The Human EcosystemFramework (Machlis et al., 1997; Pickett et al., 2001) provides auseful theoretical context to integrate natural and institutionalresources, social structure, ecological processes, and spatial pat-terns across the urban landscape. The spatial heterogeneity ofurban systems has been well noted (Jacobs, 1961; Pickett et al.,2007); however, interactions among various drivers of heteroge-neity in urban systems are not well understood (Pickett et al.,2008). Despite the analytical challenges to working in heteroge-neous urban systems, relationships between social and vegetationcharacteristics in urban areas have been identified. For instance,lifestyle behavior, housing age, family size, marriage rates andother demographic characteristics of neighborhood residents havebeen linked to vegetation cover and biodiversity in urban areas(Grove et al., 2006). Additionally, the ways in which social andecological systems are linked have been examined in a number ofcommunities that depend on natural resources for economicproductivity (Agrawal, 2001; Haase et al., 2012, Larondelle andHasse, 2012; Olsson et al., 2004).

Connections, relationships, and feedbacks among social andecological components of urban systems are critical to assess, andcan theoretically be teased apart to explain observed social–ecological system dynamics (Pickett et al., 2004). A significantfeature of the Human Ecosystem Framework is that it points tothe interactions among social and ecological components of urbansystems and suggests the need to assess social–ecological interac-tions in a spatially explicit way (Gottdiener and Hutchison, 2000).Our inability to tightly control and independently manipulatevariables in real-world coupled social–ecological systems limitsunderstanding of relationships in these systems (Walker et al.,2006). However, by combining the insights gained through theorydevelopment with those derived from analysis of case studies, wecan improve our understanding of how social–ecological systemsfunction, and extract generalities about the fundamental processesthat structure the interactions of human societies embedded inecological systems (Walker et al., 2006). We utilize a case study inNYC to investigate the spatially explicit relationship between socialneed for ES and the ecological value of vacant lots as determined bythe current production of ES, in order to provide a real-worldapplication of the conceptual framework we develop for mappingsocial–ecological conditions of urban spaces in the context of ES.While our study does not hypothesize which specific dynamics linksocial communities to ecological processes in NYC neighborhoods, itdoes identify patterns in the configuration of ES and socio-economic indicators of need for ES across the NYC urban landscape.

2.1. Urban ecosystem services

The concept of ES involves the concurrent analysis of thebiophysical and ecological foundation of ecosystems and the ways

human beings use, benefit from, and value these ecosystems(De Groot et al., 2010a, 2010b). Thus, addressing ES inherentlyrequires a social–ecological perspective (Folke, 2006) and multi-disciplinary tools for analysis (Seto et al., 2012). Since theMillennium Ecosystem Assessment (MA, 2005), ES have beenwidely conceptualized as connecting natural resources, humansociety and the economy (TEEB, 2011). For example, urban forestsand other green spaces in cities provide a wide variety ofimportant ES (Akbari, 2002; Grove et al., 2005; McPhearsonet al., 2010; McPhearson, 2011; Nowak et al., 2002; Troy et al.,2007). In the MA framework, stocks of natural resources and a hostof regulating cycles and support mechanisms underlie the socialand economic capacity to support human development and well-being (Müller et al. 2011) As the human population expands,consumption of natural resources moves closer to planetaryboundaries of regeneration rates (Rockström et al., 2009a,2009b) and the ability of the natural environment to supporthuman development and well-being is eroded. In a world operat-ing near or beyond these boundaries, there is growing need for theassessment, evaluation, and monitoring of the capacity of thenatural environment to provide services and support human well-being (MA, 2005).

Ecosystem services can be quantified either as the biophysicalunits of the service provided or the societal value of the service(most often monetary value) (Müller et al. 2011); the choice ofassessment method is often informed by the research goals as wellas data availability and this study is focused on both socialindicators of social need for ES and biophysical indicators of ESproduction. Equitable distribution of resources and social-culturalneed for ES are rarely evaluated. Few studies suggest a connectionbetween socio-economic status and the availability of urban greenspaces and the ecosystem services they provide (for a short reviewsee Pham et al., 2012). Developing methods that are able toaccount for these multiple perspectives is one of the pervasivechallenges in making social–ecological approaches to urban sus-tainability and resilience operational.

2.2. Stacking ecosystem services

Concurrent assessment of multiple ES, often referred to as“stacking” has been a contentious approach in the development ofecosystem valuation methods (Cooley and Olander, 2012). None-theless, the importance of considering multiple ES for the purposeof decision making in the context of green infrastructure planningand resource conservation management has been increasinglyacknowledged (Buckland et al., 2005; Müller et al. 2011; Hepcanand Ozkan 2011; Koniak et al., 2010; Tallis and Polasky, 2009;Weber et al., 2006). Research has only recently begun to providetools and methods for ES stacking. Buckland et al. (2005) offered acomposite indicator analysis for monitoring change in biodiversityand more recently Naidoo et al. (2008) stacked four ES based ontheir biophysical units and spatial distribution, then correlatedaverage ES provisions with a measure of biodiversity. They suggestthat this methodology is effective in identifying areas of enhancedopportunity or need for ES. It is also important to understandtradeoffs and synergies between different ES. For example, it hasbeen demonstrated that optimizing for one service does notnecessarily result in improvement in the other (Haase et al.,2012; Nelson et al., 2008).

In NYC, the effects of climate change including sea level rise,changing heat and precipitation patterns, and storm frequency andintensity are predicted to place increased pressure on local urbanecosystems to provide critical urban ES (New York City Panel onClimate Change, 2009). Only by simultaneously assessing multipleservices provided by urban ecosystems, how they change overtime, and the factors that strengthen or limit their performance,

T. McPhearson et al. / Ecosystem Services 5 (2013) e11–e26e12


will local and regional stakeholders involved in planning andpolicymaking processes be able to effectively harness ES to achieveshort-term public policy objectives (e.g. stormwater management)while enhancing the long-term resilience and adaptive capacity ofurban communities.

2.3. NYC vacant lots as a social–ecological system

Our study involves linking social and ecological heterogeneityin a spatially explicit conceptual framework to investigate a broadset of social–ecological indicators of ES in NYC. We examinedvacant lots, some of the most heterogeneous urban spaces (Tidballand Krasny, 2010) in NYC, in order to examine patterns of socialand ecological components of NYC across space. Vacant lots areunderdeveloped and underused lands in cities (PennsylvaniaHorticulture Society, 1995; Tidball and Krasny, 2010) that maypresent an opportunity to concurrently enhance ecological andsocial conditions in associated neighborhoods. New York City has29,782 land parcels identified as vacant in the city's MapPluto taxlot database (NYC Department of City Planning, 2011), includingboth publically and privately owned vacant land. Although vacantlots are often discussed as an adverse influence on neighborhoods(Kremer et al., in press), we suggest that vacant lots can be viewedas important sources of ecosystem services depending on theirsocial uses and ecological functioning. Given that vacant lots inNYC tend to be dominated by vegetated landcover, these sites canbe considered a valuable source of regulating ES, including carbonsequestration and storage, air pollution removal, localized coolingeffects and stormwater management. Here, we combine an assess-ment of ES production in NYC vacant lots with a spatial analysis ofindicators of social need for ES to highlight potential areas fordesign or management interventions that can improve ES produc-tion in areas where they are needed most.

3. Methods

3.1. A social–ecological approach to ES assessment

The social–ecological approach to assessing indicators forurban ES presented in this study involved a multi-step methodol-ogy, including collecting data from existing sources regardinggreen infrastructure and socio-economic indicators in the city,taking a stratified random sample of 1502 vacant lots across thefive boroughs of NYC, and analyzing landcover and land use inthese lots. We also calculated indicators of the ecological value ofsampled vacant lots using a multiple criteria approach andanalyzed socioeconomic indicators of social need for ES in theneighborhoods surrounding them. We then stacked multiple socialand ecological indicators by scaling and weighting, integratedthem through the social–ecological matrix framework we devel-oped, and finally mapped them.

Vacant lots in NYC were identified through a spatially refer-enced city tax lot database (NYC Department of City Planning,2011). A stratified 5% random sample of lots from each boroughwas selected (using SASs 9.2) based on borough and lot area. Lotswere then grouped into size categories using the Jenks optimiza-tion method (Longley et al., 2011). The Jenks method groups lotsinto area by maximizing between-group differences while mini-mizing within-group differences, allowing for the identification ofnatural groupings. We then compared the median and mean lotarea for each group of the sample with median and mean lot areafor each group of the population. Finding that large lots wereunder-represented, we added 26 large lots to the sample in orderto adequately represent lot area. Table 1 shows the number ofvacant lots in the population, the sample stratified by borough

only, the sample stratified by borough and lot area, and thepercentage of lots sampled in each borough.

3.2. Data collection

To estimate the potential social–ecological conditions of vacantland we analyzed the landcover and land uses of a sample of 1502vacant lots (5% of the total) in all five boroughs of NYC (Manhattan,Bronx, Queens, Brooklyn, and Staten Island). Vacant lot landcoverand land use data were collected by visually inspecting thesampled vacant lots in Google Earth. A GIS vector file was overlaidon Google Earth's digital globe and the following informationextracted: building typology, percent cover of building, pavement,bare soil, coarse vegetation, fine vegetation and water. In addition,using the high resolution zoom-in and Street View functions inGoogle Earth, actual use was determined for each sampled lot.

We modified the High Ecological Resolution Classification forUrban Landscape and Environmental Systems (HERCULES) devel-oped by Cadenasso et al. (2007) to address the fine-scale, hetero-geneous spatial characteristics of NYC landscapes in a way thatbetter characterized their representation in NYC lots. A detaileddescription of the survey procedure is provided elsewhere(Kremer et al., in press) and summarized here. For each landcoverclass, we defined one of the following semi-quantitative percentcover classes: (0) absent; (1) present–10.99%; (2) 11–35.99%;(3) 36–55.99%; (4) 56–75.99%, (5)476%. In evaluating ES basedon landcover classes, we took a conservative approach by estimat-ing area on a range between (a) the lowest value of each percentcover class and (b) the midpoint within the class. These minimumand midpoint landcover percentages were then multiplied by totallot area to obtain area in each landcover class, except for lotshaving only one landcover class. In these cases, 100% of the lot areawas assigned to a single landcover class. This approach likelyunderestimates landcover areas, and therefore potentially under-values ES in the vacant lots.

Building typology was determined based on the followingclasses: (0) No building; (1) Single structures; (2) Multiple occu-pancy structures under 4 stories; (3) High-rises between 4–10stories; (4) Towers greater than 10 stories; (5) Industrial building;(6) Accessory structures (pools, gazebos, decks, sheds, garages).Additionally, an ‘actual use’ category was developed to describehow the surveyed vacant lots are being used in practice. Actual usewas determined by visually assessing vacant lots in Google Earthand Google Street View (for detailed discussion of survey methe-dology and actual use of vacant lots see Kremer et al., in press).

3.3. Assessment of urban ecosystem services

We assessed 12 urban ecosystem services based on the benefitsthey potentially provide to local and/or regional urban residents.Based on a literature review, we assessed urban ES includingregulating services such as local temperature regulation, carbon

Table 1NYC vacant lots sample and population data.

Borough No. of lots (Totalpopulation)

Sample (5%) of lotsstratified by borough

Sample stratified byborough and lot area

Brooklyn 7902 395 397Bronx 4167 208 210Manhattan 1282 64 67Queens 8552 428 430Staten

Island7879 394 398

Total 29,782 1489 1502

T. McPhearson et al. / Ecosystem Services 5 (2013) e11–e26 e13


storage, carbon sequestration, air pollution removal, and runoffmitigation; provisioning services including food production incommunity gardens; supporting services including habitat forbiodiversity; and, cultural services including recreation.

We used a combination of the data collected in our sample,ancillary spatial datasets (such as soil type) and data recorded inthe literature to calculate the total amount of service provided, pervacant lot, for each ecosystem service, based on the appropriatebiophysical unit. Where calculating the total amount of a specificservice was not possible (e.g. runoff mitigation) an index was usedto signify the quality or level of service provided. Place-specificliterature was used to the maximal extent possible. Namely, NYCdata was preferred over US urban data, which was preferred overinternational urban data, and urban data was used exclusively overnon-urban data. We recognize that the ES urban residents mayderive from vacant lots in NYC may accrue at the neighborhoodscale (e.g. recreation, urban cooling, food production), while othersaccrue at citywide scales (e.g. carbon sequestration and storage),and still others may accrue at both scales (e.g. stormwater runoffmitigation). To be as comprehensive as possible in assessingmultiple ES, we treated the production of all services, regardlessof the scale of where the benefit may be accrued, as services in ouranalysis. Interpretation of the results of this approach may belimited due to the different scales at which urban residents maybenefit from different ES. However, our analysis focuses on theproduction of ES and not measured benefits, and all calculated ESare produced at the local scale within vacant lots. Table 2 sum-marizes ES, indicators, and data and literature sources used in theanalysis. Very little information about how people use vacant lots ispublicly-available. Therefore, measures of cultural services providedby vacant lots are limited, although we recognize that dependingon their actual use, social, cultural and health benefits are mostlikely occurring in and around many vacant lots. Below aredetailed descriptions of the methods used in our assessment of ES.

3.3.1. Carbon sequestrationThe carbon sequestration metric was evaluated on an annual

basis for coarse vegetation in vacant lots, using a NYC-basedestimate for trees (Nowak et al., 2002). The estimated netsequestration rate of 0.12 kgC/year/m2 for trees in NYC is basedon approximate tree conditions, sizes and estimated tree mortal-ity. Nowak and Crane (2002) estimate that net sequestration oftrees in most United States cities is 75% of gross sequestration.However, NYC has a relatively high proportion of dead trees andrelatively large tree diameters, yielding a lower net estimate ofonly 54% of gross carbon sequestration on a per area basis. Thecarbon sequestration rate per square meter was multiplied by thearea of coarse vegetation in the sampled vacant lots. Since wecould not identify comparable estimates for grass and herbaceousvegetation in the urban literature, our evaluation of carbonsequestration excluded these landcover types.

3.3.2. Carbon storageCarbon storage was estimated for coarse vegetation, fine

vegetation and soil carbon in vacant lots. Carbon storage for coarsevegetation was evaluated based on Nowak and Crane's (2002)estimate that trees in NYC store 7.3 kgC/m2. This is lower than themedian carbon storage for trees in urban areas, which theyestimate to be 9.25 kgC/m2. The metric used to estimate carbonstorage of fine vegetation was derived from a Chicago-based studythat measured the carbon uptake of grass in two city blocks to be0.18 kgC/m2 (Jo and McPherson, 1995). Given that the fine vegeta-tion category includes shrubs and other types of vegetation, whichtypically have higher carbon storage values than grass, thiscalculation likely underestimates carbon storage of fine vegetationin the assessed NYC vacant lots. Moreover, Kulshreshtha et al. (2000)estimate that a Canada prairie ecosystem stores 0.19 kgC/m2, provid-ing support that 0.18 kgC/m2 is a reasonable estimate for the carbon

Table 2Ecosystem services indicators, data and literature source.

Ecosystem service Indicator Data Literature sources

Provisioning servicesFood production Community garden (yes/no) Sample data –

Regulating servicesCarbon Seq. Coarse veg area� avg. sequestration rate (kg/m2) Sample data Nowak and Crane (2002: 385) and Akbari (2002)Carbon storage Coarse veg area� avg. storage rate (kgC/m2) Sample data Nowak and Crane (2002), Jo and McPherson (1995),

and Pouyat et al. (2002)Fine veg area� avg. storage rate (kgC/m2)Soil area� carbon density/m2 (kgC/m2)

Air pollutionremoval

Sample data Nowak and Crane (2002) and Yang et al. (2008)

SO2 Coarse veg area�pollution removal rate (g/yr)NO2 Fine veg area�pollution removal rate (g/yr)PM10

O3

COLocal tempregulation

Cooling by vegetation (degrees Celsius) Sample data Huang et al. (1987)

Runoff mitigation Runoff coefficient (% precipitation absorbed) Sample data, NYC soilsurvey

USDA (1986), NYC Soil Survey Staff (2005), Trataloset al. (2007), and Whitford et al. (2001)

Support servicesProvision of Habitatfor biodiversity

Ecological priority areas, %green, brown and blue Sample data, NYSDEC,(2009) and TNC (n.d.)

Turner (1989), Whitford et al. (2001)Situated within high priority area – community occurrences,bird conservation, TNC priority ecoregionsConnectivity: proximity to green areas NYC green layers NYS GIS Clearinghouse; NYC Open MapShape: compactness Sample data –

Size: size of total green space Sample data, NYCMapPluto

NYC Department of City Planning (2011)

Cultural servicesRecreation Public access: owned publicly (yes/no) NYC MapPLUTO NYC Department of City Planning (2011)

Actual use: used by the public for recreation Sample data –



storage of grass. The metric for soil carbon is based on Pouyat et al.(2002) estimate of C density of urban soils of 8.2 kgC/m2. Tocalculate soil area, we summed the total area of coarse vegetation,fine vegetation and bare soil. However, since we were unable todetermine whether landcover under tree canopy is pervious orimpervious, we conservatively excluded coarse vegetation in lotsthat had mostly impervious surfaces (paved landcover and build-ings) from the soil carbon estimate.

3.3.3. Air pollution removalAir pollutant removal by coarse and fine vegetation were

estimated using metrics from an urban forest survey of trees inBrooklyn, New York, which took place in 1994 during nonprecipi-tation periods (Nowak et al., 2002), and a survey of pollutionremoval rates of vegetation on green roofs in Chicago over a one-month period (Yang et al., 2008). Both the Brooklyn and Chicago-based studies surveyed removal of SO2, NO2, PM10, and O3, whilethe Brooklyn-based study surveyed CO in addition. No comparablemetrics for CO removal by fine vegetation were found in theliterature. Although Yang et al. provide an estimate for the removalof O3 by fine vegetation, this metric was not used in our assess-ment. Yang et al. estimate O3 removal by fine vegetation at ahigher rate than O3 removal by trees in Brooklyn (Nowak et al.,2002). Since trees provide greater pollution removal for all otherpollutant types, it seems unlikely that fine vegetation removesmore O3 than trees. Since we did not include CO and O3 removal byfine vegetation in the assessment, estimates for removal of thesepollutants by vacant lots is likely conservatively undervalued inour study. See Appendix B.1 for metrics used to estimate airpollution removal rates of coarse and fine vegetation in NYCvacant lots.

3.3.4. Local temperature regulationTrees reduce local air temperature by shading and by evapo-

transpiration, as they release water into the atmosphere. Coolingeffects of trees can be estimated based on the ratio of latententhalpy hours (LEH) to cooling degree days (CDD) in a city(McPherson and Simpson, 1999). This ratio accounts for theplace-specific climate characteristics that are related to theamount of vegetation required to affect temperature. NYC's LEH/CDD ratio is 1.5, which is associated with a 0.05 1C decrease intemperature per percentage increase in tree cover. Given NYC'sclimate, we can estimate that for every additional percent of treecanopy, local temperature is reduced by 0.05 1C. To estimate thetemperature reduction within each lot provided by tree canopy,the percent of the total lot covered by tree canopy (coarsevegetation) was multiplied by 0.05 1C.

3.3.5. Runoff mitigationTo calculate the potential to mitigate runoff by infiltrating

stormwater, we used an example of a large storm event (5 in. rainevent over 24 h) that would create a significant runoff event likelyto overwhelm the city's sewer system. A larger or smaller rainevent would respectively increase or decrease runoff mitigation.Runoff mitigation was calculated as the percent of rain absorbedin a vacant lot during a 24 h, 5 in. rain event. The USDA SoilConservation Service (SCS) runoff curve number (CN) is assignedto each lot based on combinations of hydrological soil group andlandcover according to the methodology developed in USDA TR-55(USDA, 1986) that is widely used to predict runoff from rainfall.Hydrologic Soil Groups were derived from the NYC soil survey(New York City Soil Survey Staff, 2005) through an overlay analysisin ESRI ArcGIS 10. To utilize the sampled data for this purpose,vacant lots were classified into three major usage categories (openspace, impervious area, and residential area) based on the “actual

use” observations and then reclassified according to the USDATR-55 (1986) landcover categories. USDA TR-55 Landcover cate-gories for urban areas include residential, paved, bare soil and openspace cover. The open space category is further divided to poor(grass covero50%), fair (50%ograss covero75%) and good condi-tion (grass cover475%). Next, we assigned SCS CN to the differentcombinations of landcover/hydrologic soil group as shown inTable B.2 in Appendix B. Runoff and potential retention werecalculated based on the methodology outlined by USDA (1986:Chapter 2) for potential maximum retention during a 5 in. rainevent. Similar to Whitford et al. (2001), we use an index of ‘runoffretention coefficient’ that represents the proportion of precipitationthat will be retained during the rain event. Total precipitationretention for each lot was calculated by multiplying the coefficientby lot area and rain event.

3.3.6. Food productionFood production was evaluated based on the identification of a

lot as a community garden. Because the overwhelming majority ofcommunity gardens in the city use their space to produce food(Gittleman et al., 2010), where a vacant lot was identified as acommunity garden, we assumed food is being produced.

3.3.7. Provision of habitat for biodiversityFour indicators including habitat availability, connectivity,

sensitivity, and shape, were used to evaluate the quality of habitatprovisioning for biodiversity of each of the sampled lots. Habitatavailability is calculated as the percent of the total coarse and finevegetation within a given vacant lot. Connectivity was calculatedas the distance to other green areas in the city using the “Near”tool in ESRI ArcGIS 10 with a search buffer of 500 m. Distancebetween green areas is assumed to be important since speciesdispersal is challenging for many species living in discontinuousurban green spaces, placing a biodiversity premium on connectedurban green spaces. A green layer that includes all major knowngreen and open spaces in NYC was created by merging availableGIS green infrastructure layers including federal and state recrea-tion and non-recreation open space, NYS Department of Conserva-tion land, and NYC parks and other open space. GIS layers wereobtained from the NYS GIS Clearinghouse (http://gis.ny.gov/) andthe NYC OpenData (https://nycopendata.socrata.com/) websites.Ecological Priority Area is defined as a binary indicator suggestingwhether a vacant lot is located within a designated “ecologicalpriority area” or not as defined by New York State Department ofEnvironmental Conservation (2009) and The Nature Conservancy(n.d.). Shape is defined as a measure of the relative compactness ofeach lot. Compactness was determined using ArcGIS tool ShapeMetrics Tool (Parent, 2011), which provides a way to measure apolygon's likeness to a circle (Angel et al., 2010). We assumed thatbiodiversity is higher in green spaces that contain both edge andinterior habitat, since different species inhabit edge versus interiorurban habitat (Dowd, 1992; McKinney, 2002). A circle was used asthe ideal shape because it minimizes habitat edge to interior ratios(Smith and Smith, 2006), therefore providing an indicator ofbiodiversity quality that is biased towards including both edgeand non-edge species. Compactness values calculated by theShape Metrics Tool are given on a 0–1 scale, where 1 is the mostcircle-like.

The total area of green, brown and blue landcover in each lotwas aggregated and grouped into 10 bins. Because our sampleincluded a small number of very large lots with wide dispersion,we grouped lots in the highest five percent of green, brown andblue landcover area into one category. The choice of the top 5% forthe group of largest lots was based on a sensitivity test thatindicated this was the lowest percentage that produced a close



approximation to a normal distribution. The remaining data wereclassified into nine categories using Jenks optimization algorithm.Table 3 shows the percent of the distribution of total green, blueand brown land area for our vacant lot sample data.

3.4. Assessment of social indicators

We selected socioeconomic indicators that we assumed reflectsocial need for ES (Pham et al., 2012) in the vicinity of vacant lots.Since vacant lots themselves provide limited insight into thesocioeconomic characteristics of the neighborhood, social indica-tors are assessed based on a neighborhood analysis. The definitionof a neighborhood for the purpose of ES analysis is little studied,but insight can be drawn from the more expansive literature onwalking behaviors and proximity to green space. Suggested dis-tances for a neighborhood analysis range from 300 to 1000 m(Colabianchi et al., 2007; Hörnsten and Fredman, 2000). In thisstudy, we chose a buffer of 500 m as representative of theimmediate neighborhood.

Four indicators were chosen to exemplify social need andcalculated from publically available Census and City data in theneighborhood of the sampled vacant lots – median householdincome, median real estate values, population density, and densityof green and open space. Median household income and realestate values are commonly used indicators in studies of therelationship between socioeconomic status and various environ-mental issues (e.g. locations of contaminated sites or correlationwith green spaces) (Grove et al., 2006; Mennis, 2002; Talen, 2010;Troy et al., 2007). Moreover, recent environmental justice litera-ture suggests that green space availability and the ES provided byurban areas are positively correlated, implying that low income areunderserved by urban green spaces (Pham et al., 2012). Populationdensity is a measure of the quantity of the population thatpotentially benefits from ES (Pham et al., 2012; Talen, 2010), andgreen space density is used to indicate the current level ofavailability and access of these services.

The Model Builder tool in ArcGIS 10 was used to perform aniterative analysis in which a buffer was drawn around each lot andthe social need indicators were evaluated within the buffer. Westructured the indicators in our combined social–ecological assess-ment to evaluate social need for ES. Thus, low household income,low real estate value, and low green space density correspond tohigh social need while low population density correspond to lowsocial need.

3.5. Social–ecological matrix

To combine the social need for urban ES with the ecologicalvalue (as assessed by ES provisioning) of sampled vacant lotsacross the city, all indicators were rescaled from their organic unitinto a standard scale of 1–10, where 10 is the highest occurring

value of an ES and 1 is the lowest, and all other values normalizedto a 1–10 scale. All ES indicators were then grouped and weightedequally. Each ecosystem service group was weighted as one, and ifmore than one indicator was used to evaluate an ecosystemservice, weight was distributed equally between the differentindicators. For example the ES air pollution removal was calculatedby equally weighting the level of pollution removal of eachpollutant. To calculate the provision of habitat for biodiversity allfour indicators (ecological priority areas, connectivity, shape andsize) where weighted equally. Defining the weights for eachassigned indicator is a crucial task in creating an integratedassessment mechanism for ES. To-date there is little agreementon methods and tools for determining weights, but there is agrowing recognition that such methods are essential (Felix et al.,2011; Müller et al. 2011) and will benefit from progress invaluation research. Some ways the issue of weighting ES has beenresolved in the past include expert and public stakeholder engage-ment (Bryan et al., 2011; Calvet-Mir et al., 2012) or weightingsystems based on existing literature (Hepcan and Ozkan 2011). Forthe purpose of demonstrating a social–ecological approach to ESstacking in NYC, we weighted all indicators equally. After socialneed and ecological values were weighted and summed, theresulting number was assigned a High or Low, based on whetherit was above or below the data median, in order to provide a toolfor examining the social–ecological interactions within a particularspatial location. The choice to designate high and low categoriesdata median was made for simplicity, but different thresholdscould be applied according to different standards, such as com-munity objectives or natural breaks in the data. By splittingecological value and social need for ES into binary High or Lowdesignations, we were able to apply a social–ecological matrixapproach, developed coincidentally with overlapping research inStockholm (Ranara et al., 2011), for identifying the social–ecologi-cal condition of space in the context of ES. For simplicity welimited ourselves to binary classes High and Low, though analyzinga larger number of classes is possible. Social need and ecologicalvalues were then classified into four combined groups (High–High,Low–Low, High–Low, Low–High) in a social–ecological matrix asshown in Fig. 5.

4. Results and discussion

4.1. Vacant lots typology

Results of our assessment revealed that the dominant landcovertype in sampled vacant lots was fine vegetation (54%), followed bycoarse vegetation (29%), then paved surface (6%) and building cover(3%), with water accounting for only 1% of total area of the sampledvacant lots. The lots in the top five percent of area (i.e. large lots) ingreen, blue and brown landcover accounted for 77% of total landarea in the sample, as well as more than fifty percent of eachlandcover category. Vacant lot land area was distributed hetero-geneously across the five boroughs, with 63% of the vacant lot landin our sample concentrated in Staten Island, 23% in Queens and lessthan 10% in each of the other three boroughs, Manhattan, Brooklyn,and Bronx. Landcover types were also distributed heterogeneouslyacross the five boroughs. For example, most building cover wasconcentrated in Brooklyn vacant lots, while most bare soil wasconcentrated in Manhattan vacant lots. Notably, there was highvariance in the area of fine and coarse vegetation of sampled vacantlots, but only in Staten Island and Queens. Fig. 1 shows comparativeland area in each landcover type across boroughs. Fig 2. presentsexamples of surveyed vacant lots.

Table 3Lot green, brown and blue landcover area distribution.

Category Number of lots Percent of total

1 324 222 175 123 167 114 255 175 165 116 164 117 82 58 61 49 29 2

10 80 5



4.2. Ecosystem services of vacant lots

Based on the assumptions and methods outlined above, weestimated total carbon and air pollution removal for the samplepopulation of vacant lots, as well as means and medians on a perlot basis. We also estimated cooling effects and stormwatermanagement at the lot level (Appendix A: Table A.1).

Our results indicate that surveyed vacant lots stored between3.01E+07 and 3.78E+07 kgC. Scaled up to the entire city weestimated that vacant lots stored a total of between 5.97E+08and 7.48E+08 kgC. Similarly, sampled vacant lots sequesteredbetween 9.95E+04 and 1.37E+05 kgC/yr, and the total populationof lots was estimated to sequester between 1.97E+06 and 2.72E+06 kgC/yr. For air pollution we found that tree cover in vacantlots absorbed a total of between 3.21E+08 and 4.01E+08 kg of airpollution (including CO, O3, PM10, SO2, and NO2). We also foundthat vacant lots in NYC were cooled between 1.3 and 1.61 1C by

coarse vegetation on a per lot basis. We hypothesize that for localclimate regulation (and potentially other ES), cooling by bothshading and evapotranspiration from trees in vacant lots has aspatial decay function from the tree to the nearby surroundingarea. However, research is unclear on the shape or rate of thedecay function and how it may influence local need for cooling.Our estimate is conservative in that it does not account foradditional cooling effects of trees on surrounding lots and there-fore captures only a highly localized cooling effect.

Since prioritizing green infrastructure (the dominant landcovertype in sampled vacant lots) for stormwater absorption is a majorpriority in NYC (New York City, 2010), we calculated stormwater runoffinfiltration rates and found that 37% of the rain in a 24 h, 5 in. rainevent could be retained by the sampled vacant lots as they currentlyexist. This translates to total stormwater retention for sampled vacantlots of 38 m3 of rain in such a rain event (Appendix A: Table A.1) anddemonstrates the ecological value of vacant lots at both the local and

Fig. 1. Distribution of landcover in vacant lots in the five boroughs of NYC (N¼1502). Area is represented as percentage of the total area within each borough and labels showthe total area in square meters.

Fig. 2. Google Earth Images of surveyed vacant lots. The lots exemplify the four classes of the social-ecological matrix.



regional scale, as every m3 of rain water absorbed is a m3 preventedfrom contributing to stormwater runoff.

In our surveyed lots, 59–74% of the surface was covered withtrees, vegetation or bare soil, providing a diverse habitat matrix forbiodiversity. Additionally, many of the sampled lots (197) arelocated within areas defined as sensitive ecosystems by TheNature Conservancy. We were also interested in the location ofecologically valuable vacant lots in the overall matrix of ecosystempatches in the city. Median distance to other green spaces,indicating support of connectivity between patches, was 454 m.We also evaluated food provisioning of vacant lots since some lotswere active community gardens, as well as public access andownership type as indicators of the potential for vacant lots toprovide cultural services (Appendix A: Table A.3).

Indicators of social need for ES in the neighborhoods aroundsampled vacant lots suggest that lots are mostly located in areas ofhigh population density (median of 12,033 people\sq km) withbelow average city median household income. High populationdensity is assumed to have greater need for ES as a simple functionof increased density (Talen, 2010). Median household income hasbeen shown to indicate need for ES (Pham et al., 2012). Averageannual median household income in the neighborhood around thesurveyed vacant lots is $42,581, while the city's average is $50,285 USCensus (n.d.). Green density, indicating the relative availability ofgreen space in these neighborhoods, was also relatively low with amedian value of 5%, suggesting high social need for green space andassociated ES. Detailed results and summary of statistical descriptorsof ecological and social indicators are presented in Appendix A.

Fig. 3. Spatially explicit mapping of indicators of ecological value in a sample of vacant lots in NYC.



4.3. Mapping social–ecological systems

Since ecological value and social need are calculated andattributed to a specific lot, we were able to map the quality of ESand level of social need for ES spatially across the city. Thisspatially explicit method afforded the opportunity to compareand identify important areas for conservation and priority areasfor urban planning and social–ecological transformation based onthe site-specific conditions of low ecological value and/or highsocial need. The spatially explicit results of the ecological assess-ment for a subset of ES indicators analyzed including runoffmitigation, carbon storage, habitat provisioning, and air pollutionremoval are shown in Fig. 3. Results are mapped as scaledindicators of ES on a scale between 1 and 10. The ES maps

exemplify an important observation regarding ES in the contextof planning and urban governance, namely, that there are tradeoffsand synergies among various ES. Decomposing the stacked indices(see Fig. 7) can facilitate decision-making about which services tomaximize in a particular planning scenario or among competingneighborhood goals for a particular space. In Fig. 3 we show thatareas most significant for habitat or carbon storage are notnecessarily those best suited to maximizing runoff mitigation.This particular result is mainly due to the fact that soil type is amajor factor in the ability to absorb rain, while carbon storage isdriven by vegetated landcover, and habitat is driven by bothlandcover and proximity to other habitats. In a similar manner,when mapping indicators of social need (Fig. 4), it is clear thatdifferent measures of social need have varying spatial distribution.

Fig. 4. Spatially explicit mapping of indicators of social need for ecological services within a radius of 500 m around sampled vacant lots.



While access to green spaces (or the lack of it) is evident acrossthe five boroughs, high social need indicated by the populationdensity and median income indicators are more concentrated inspecific boroughs and in local neighborhoods within the boroughs(Fig. 3).

4.4. Social–ecological matrix of social need and ecological value

While mapping ecological and social indicators separately isheuristically interesting and provides some utility, a more challen-ging task involves evaluating multiple indicators of value withinthe ‘ecological’ category, multiple indicators of need within the‘social’ categories, and the way social–ecological indicators interactacross space.

We recognize the complexities and difficulties in aggregatingmultiple ES; nonetheless, we also recognize the importance ofadvancing ES stacking methods for the purposes of planning anddecision making and suggest a novel, combined social–ecological,multiple criteria ES assessment method. In the resulting social–ecological matrix (Fig. 5) we present the spatially explicit results ofthe methodological steps involved in weighting and grouping thesocial and ecological indicators described above. Four classes ofresults are created by dividing the aggregated social and ecologicalindicators into high and low categories that indicate the level ofsocial need in conjunction with a combined measure of ecologicalvalue (see Fig. 2 for examples of lots with different class config-urations). For example, H–H represents high social need for ES inconjunction with high ecological value, corresponding to thestacked quality level of ES provisioning. Examining these locationsfurther may offer insight into the interaction of areas of high socialneed with their urban ecology and may provide useful insight intothe governance mechanisms relevant to the management of

vacant lots. On the other hand, L–H lots represent low social needand high level of ES. Many of L–H lots are located in Staten Island,characterized by predominantly middle class population withrelatively low population density and more green space than anyother borough. We found H–L lots (high social need in conjunctionwith low level of ES) to be the most interesting category in termsof demonstrating opportunities for social–ecological transforma-tion. Lots with low existing ecological value in neighborhoodscharacterized by high social need could present an importantopportunity for developing vacant land in underserved areas. Forexample, H–L lots could be prioritized for green infrastructuredevelopment such as community gardens or urban pocket-parkswhich could enhance ecological value while meeting some socialneeds for community spaces that improve neighborhood condi-tions. Further study of local social need in H–L lot areas, perhapsby survey of local residents, could help determine appropriatetransformations. If developed appropriately, these lots couldprovide opportunities to improve sustainability and resilience ofthe NYC social–ecological system.

4.5. Identifying clusters

In order to examine patterns at the neighborhood level, weidentified spatial hotspots of vacant lots that shared both proxi-mity to one another and specific social–ecological configurations,such as high social need and low ecological value. Cluster analysisis useful for examining transformation potential at a broader scale,such as the neighborhood level. Using the Getis-Ord Gin HotSpotanalysis tool in ESRI ArcGIS 10, we identified areas that havestatistically significant (po0.01) clustering of high social need andlow ecological values (H–L), and then intersected them to identifyareas of High–Low clustering. Fig. 6 shows that major H–L clusterswere found in the Bronx and Central and South Brooklyn. Theresults of the Getis-Ord Gin statistic are reported through a Z score,which represents the statistical significance of clustering for aspecified distance that is based on Randomization Null Hypothesiscomputation (Longley et al., 2011).

Overlaying the clustering results with a zoning map of NYCreveals that the majority (88%) of lots in the H–L clusters arelocated in residential areas, 60% of which are located within highdensity residential areas (NYC zones R6–R10) and 28% in lowerdensity residential areas (zones R1–R5). Only 22 lots (10%) withinthe H–L cluster are located in manufacturing zones. The findingthat most H–L clusters are located in residential areas is importantbecause social–ecological transformation is more desirable inresidential areas than industrial zones, which are less accessibleto people. This is not surprising given that industrial zones areexpected to have low population density, whereas residentialzones are expected to have higher population density. However,the high social need found near the H–L lots that are located inmanufacturing zones may be due to the inherently low incomeand property values in manufacturing areas. This exemplifies oneof the challenges in this analysis – the results are sensitive to theranking and weighting scheme. If population density wereweighted more heavily in our analysis, lots in manufacturing areaswould most likely disappear from the clusters, which points to theneed for further research on weighting in ES stacking methods.Nonetheless, the clustering results demonstrate the utility of ourES stacking and assessment methods for highlighting vacant lotsmost in need of social–ecological transformations in the land-scape. We envision these transformations, such as turning parkinglots into pocket parks, as ones that can simultaneously improveecological value in low ecological value locations and meet socialneed in areas with the highest need.

Fig. 5. Social–ecological matrix combining social need and ecological value ofvacant lots in NYC.



4.6. Limitations of the research

Limitations of this study include data availability, and concep-tual and methodological issues. The accuracy of ecological indica-tors is only as good as the accuracy of the landcover data and manyancillary data sets and GIS layers available to researchers. Thesedatasets may not be in a central hub or at an appropriate spatialresolution for a given city or neighborhood assessment. Forexample, one of the early barriers for the development of thisstudy was lack of publically available high resolution landcoverdata for NYC, which prompted the use of the randomized surveyusing Google Earth. Given that our landcover estimations wereinexact, based on ranges, the overall ES estimates for specific lotsare also inexact. For this reason, the cluster analysis, finding astatistically significant number of proximate lots with similar

characteristics, may be more meaningful than the ecological valueof any particular lot. Due to the spatial heterogeneity of the urbanlandscape, high spatial resolution is a matter of necessity forestimating ES based on biophysical data. The continued develop-ment of such datasets and their growing availability will makespatially explicit, social–ecological urban systems research morewidely feasible.

Our calculated ecological value of vacant land is spatiallyfocused on the production of ecosystem services, rather thanattempting to assess benefits to particular urban residents. Poten-tial scale mismatches between the production of ES and where thebenefits accrue is an important area in need of further research.Further, we examined social need for ES, but it may also beinstructive to examine need in an even wider social–ecologicalcontext. For example, examining point source pollution emissions

Fig. 6. High–Low social–ecological clusters. (A) Cluster analysis of ecological value; (B) cluster analysis of social need; and (C) combined social–ecological overlay of clusteranalysis. H–L clusters represent areas where social has significant high Z scores and ecological values have significantly low Z scores. These areas can be identified as priorityareas for transformation to simultaneously improve the provisioning of ES and meet social need for ES in areas with high need.



and more detailed surface flow analysis of current stormwaterflow patterns could be important for characterizing the benefit ofstormwater absorption at scales appropriate for informing man-agement and planning.

An important conceptual issue is the aggregation of social andecological indicators in a multiple criteria indicator analysis. Whilewe demonstrated the importance and usefulness of aggregatedsocial–ecological mapping, the methodology inherently requiresthe obfuscation of the details of the specific indicators, inhibitingthe ability to account for tradeoffs and synergies. To overcome thislimitation, the interplay between the aggregated data and parti-culars of specific locations can be developed by decomposing thestacked indices to examine tradeoffs and synergies among multi-ple ES for a particular lots.. The spider diagram in Fig. 7 (utilizing asubset of ecological and social indicators) exemplifies the multi-variate nature of the indicator stacking and provides one possibleway to decompose the indices and compare characteristics ofdifferent lots. Additionally, a comparison between adjacent lots orlots located in places of interest can provide a more detailedassessment. Here, we present lots that fall within the four classesin the social–ecological matrix. It is clear in Fig. 7 that a particularclass membership can be driven by different score compositions.For example, lots 1 and 3 both represent high social need.However, in lot 1 this is driven by population density, while inlot 3 it is driven by the lack of green space.

A significant methodological difficulty involves the stacking ofecological and social indicators by the assignment of weights.While the assignment of weights is likely to remain a normativeprocess, more research is needed to establish best practices fordetermining appropriate weighting mechanisms in various situa-tions. Another methodological issue is the sensitivity of the High–Low quadrant analysis to the definition of the midpoint, because itoffers a coarse classification of the data. Further research isrequired to evaluate this sensitivity and its impact on researchresults given that this coarse breakdown is one of a few ways to

represent spatially and visually the conjunction of social andecological phenomena. It is important to further develop social–ecological indicator stacking for developing dynamic scenariomodeling of change under different urbanization, global change,and development scenarios, which are essential for the develop-ment of the study and practice of urban resilience.

5. Conclusion

Among the many types of urban infrastructure in the city,vacant lots provide a heterogeneous matrix of social–ecologicalpatches that provide multiple ES. Spatially explicit analyses ofsocial–ecological systems enables the quantification of ecologicaland social variables in ways that can be useful for visualization, forstacking ES, and for identifying priority areas for social–ecologicaltransformation. Our goal has been to understand the socio-ecologyof complex urban patches in order to illuminate overlooked placesin the city where policy, planning, or community developmentcould simultaneously meet combined biodiversity habitat, ESprovisioning and social justice goals.

Mapping stacked ES involved linking social and ecologicalheterogeneity in a spatially explicit conceptual framework. Weused a social–ecological matrix approach to ES assessment wherehigh/low social need for urban ES and high/low ecological value asassessed by urban ES provisioning generated for each vacantpatch. Our results indicate that urban land is indeed highlyheterogeneous, in particular with respect to social need for, andspatial variation in, ES. Many vacant land parcels are alreadyproviding significant ES to local residents and at the city scaleincluding stormwater mitigation, air pollution removal, carbonsequestration and storage, food production, space for recreation,and habitat for biodiversity. A social–ecological matrix mapdemonstrates how our methodological approach can illuminateopportunities for social–ecological transformation of areas in the

Fig. 7. Example of multivariate representation of social and ecological indicators for selected vacant lots (one from each quadrant).



urban landscape that are not providing significant ES. Sampledvacant lots in NYC were often located in conjunction with socialdemographic and economic characteristics that indicate a socialneed for ES, suggesting a novel opportunity for transformingvacant land in ways that enhance ES provisioning, especially inhotspot locations with high social need and low ecological value.

A concerted effort by the city, grass-roots organizations andindividuals to convert underutilized vacant land into green infra-structure with combined social–ecological amenities could pro-vide increased resilience to predicted near-term effects of climatechange (New York City Panel on Climate Change, 2009) includingoffsetting predicted increases in stormwater and urban heat.Transforming vacant land has the potential to increase the overallsustainability of the city as well through provisioning increasedgreen space for urban gardening, recreation, habitat for biodiver-sity, carbon and air pollution absorption and other regulating,provisioning, and cultural ES.

Improved knowledge of the linkages between biodiversity, greenspace, and the potential for improving urban ES can be generated bythoughtfully planned future land use and management scenariosusing some of the approaches we have discussed above. Spatiallyexplicit mapping at high resolution allows visualization of stacked ES,which can provide increased awareness at local and regional levels ofthe benefits of biodiversity and ES for urban areas, especially thebroad suite of services provided by green infrastructure. This knowl-edge can help to reinforce the understanding of the co-dependencyof urban social–ecological dynamics and their impacts on the

surrounding landscape, its biodiversity and the interactions betweenurban, peri-urban and rural areas.

Acknowledgments

We would like to thank Thomas Elmqvist, Åsa Gren, ErikAndersson, Steward Pickett, and two anonymous reviewers for

Table A.1Indicators for regulating ES.

Ecosystem service Total Mean Median

Sample (N¼1502) Population (N¼29,782) Per lot (N¼1502) Per lot (N¼1502)

Carbon removalCarbon sequestration (kg/yr/m2) 9.95E+04–1.37E+05 7.27E+05–9.22E+05 66–91 3–6Carbon storage (kg/m2) 3.01E+07–3.78E+07 2.02E+08–2.47E+08 20,061–25,134 1661–2099

Air pollution removalSO2 (g/yr/m2) 2.32E+06–2.86E+06 1.45E+07–1.76E+07 1541–1905 112–145NO2 (g/yr/m2) 6.48E+06–7.76E+06 3.89E+07–4.64E+07 4315–5163 290–390PM10 (g/yr/m2) 4.37E+06–5.45E+06 2.78E+07–3.39E+07 2907–3629 213–281O3 (g/yr/m2) 2.54E+06–3.50E+06 1.85E+07–2.34E+07 1689–2327 88–160CO (g/yr/m2) 4.81E+05–6.62E+05 3.50+06–4.44E+06 320–441 16–30

Air temperature regulationCooling effects (degrees Celsius) – – 1.3–1.6 0.7–1.1

Stormwater managementProportion of rainfall retained – – 37% 33%Volume rainwater retained (m3) – – 38 10

Table A.2Indicators for biodiversity habitat.

Statistic N (Population) Green, blue and brown landarea (m2)

Proportion green, blue andbrown landcover

Compactness (scale: 0–1 where1¼most compact)

Proximity to other greenspaces (m)

Number oflots withsensitivehabitat

TNC DFW

Mean Per lot(N¼1502)

1943–2,405 59–74% 0.68 527 – –

Median Per lot(N¼1502)

156–202 70–96% 0.65 454 – –

Total Sample (1502) 2,918,469–3,611,792 – – – 197 11Population(29,782)

57,814,269–71,549,709 – – – – –

Table A.3Indicators for cultural and provisioning services.

Statistic Green space density City-owned

Foodproduction

Total # of lots (N¼1502) 80 340 26Percent of total (N¼1502) 5% 23% 2%

Table A.4Four indicators of social need for ES (within a 500 m buffer of lot).

Statistic Medianhouseholdincome ($)

Averagepopulationdensity

Medianproperty value($)

Green density (%)

Mean 42,581 12,984 52,254 11Median 40,900 12,033 34,583 5



comments and suggestions that helped shape and improve thedevelopment of this research.

Appendix A. Detailed social–ecological assessment results

The following Tables A.1–A.4 summarize findings of the socialand ecological assessment of services and indicators. Where mean-ingful, values and indicators were scaled up from the samplepopulation to total population. Table A.1 provides total estimatedcarbon and air pollution ecosystem service values at the sample andpopulation level, as well as measures of central tendency for allecosystem services. Habitat indicators are summarized in Table A.2,including area and proportion of area in green, blue, and brownlandcover; compactness; connectivity and sensitivity. Cultural andprovisioning services and indicators, which include total numberand percent of total lots which are publicly-accessible, city-ownedor being used for food production, are shown in Table A.3. (Valuesin Table A.3 are relevant to the sample only.) Social indicators ofneed, summarized in Table A.4, were estimated within a 500-meterbuffer of each sampled vacant lot. These include household income,population density, property value and green density.

Appendix B

Summary of literature values used to estimate: (1) air pollutionremoval services of trees, shrubs and herbaceous landcover;(2) Runoff mitigation index

See Tables B1 and B2.

References

Agrawal, A., 2001. Common property institutions and sustainable governance ofresources. World Development 29 (10), 1649–1672, http://dx.doi.org/10.1016/S0305-750X(01)00063-8.

Akbari, H., 2002. Shade trees reduce building energy use and CO2 emissions frompower plants. Environmental Pollution 116 (Suppl.), S119–S126. (Retrievedfrom)⟨http://www.ncbi.nlm.nih.gov/pubmed/11833899⟩.

Alberti, M., Marzluff, J.M., Shulenberger, E., Bradley, G., Ryan, C., Zumbrunnen, C.,2003. Integrating humans into ecology: opportunities and challenges forstudying urban ecosystems. BioScience 53 (12), 1169, http://dx.doi.org/10.1641/0006–3568(2003)053[1169:IHIEOA]2.0.CO;2.

Angel, S., Parent, J., Civco, D.L., 2010. Ten compactness properties of circles:measuring shape in geography. Canadian Geographer/Le Géographe Canadien54 (4), 441–461, http://dx.doi.org/10.1111/j.1541-0064.2009.00304.x.

Bolund, P., Hunhammar, S., 1999. Ecosystem services in urban areas. EcologicalEconomics 29 (2), 293–301, http://dx.doi.org/10.1016/S0921-8009(99)00013-0.

Bryan, B.A., Raymond, C.M., Crossman, N.D., King, D., 2011. Comparing spatiallyexplicit ecological and social values for natural areas to identify effectiveconservation strategies. Conservation Biology: The Journal of the Societyfor Conservation Biology 25 (1), 172–181, http://dx.doi.org/10.1111/j.1523-1739.2010.01560.x.

Buckland, S.T., Magurran, a E., Green, R.E., Fewster, R.M., 2005. Monitoring changein biodiversity through composite indices. Philosophical transactions ofthe Royal Society of London. Series B, Biological Sciences 360 (1454), 243–254,http://dx.doi.org/10.1098/rstb.2004.1589.

Cadenasso, M.L., Pickett, S.T.A., Schwarz, K., 2007. Spatial heterogeneity in urbanecosystems: reconceptualizing land cover and a framework for classification.Frontiers in Ecology and the Environment 5 (2), 80–88, http://dx.doi.org/10.1890/1540–9295(2007)5[80:SHIUER]2.0.CO;2.

Calvet-Mir, L., Gómez-Baggethun, E., Reyes-García, V., 2012. Beyond food produc-tion: ecosystem services provided by home gardens. A case study in Vall Fosca,Catalan Pyrenees, Northeastern Spain. Ecological Economics 74, 153–160, http://dx.doi.org/10.1016/j.ecolecon.2011.12.011.

Chan, K.M.a, Shaw, M.R., Cameron, D.R., Underwood, E.C., Daily, G.C., 2006.Conservation planning for ecosystem services. PLoS Biology 4 (11), e379,http://dx.doi.org/10.1371/journal.pbio.0040379.

Colabianchi, N., Dowda, M., Pfeiffer, K.A., Porter, D.E., João, M., Almeida, C.A., Pate,R.R., 2007. Towards an understanding of salient neighborhood boundaries:adolescent reports of an easy walking distance and convenient driving distance.International Journal of Behavioral Nutrition and Physical Activity 4, 1–4http://dx.doi.org/10.1186/1479-Received.

Cooley, D., Olander, L., 2012. Stacking ecosystem services payments: risks andsolutions. Environmental Law Reporter, 10150–10165.

Daily, G.C., Polasky, S., Goldstein, J., Kareiva, P.M., Mooney, H.a, Pejchar, L., Ricketts,T.H., et al., 2009. Ecosystem services in decision making: time to deliver.Frontiers in Ecology and the Environment 7 (1), 21–28, http://dx.doi.org/10.1890/080025.

De Groot, R.S., Fisher, B., Christie, M., Aronson, J., Braat, L., Haines-Young, R., Gowdy, J.,et al., 2010a. Integrating the ecological and economic dimensions in biodiversity andecosystem service valuation. In: Kumar, P. (Ed.), The Economics of Ecosystems andBiodiversity TEEB Ecological and Economic Foundations. TEEB, London, UK.

De Groot, R.S., Alkemade, R., Braat, L., Hein, L., Willemen, L., 2010b. Challenges inintegrating the concept of ecosystem services and values in landscape planning,management and decision making. Ecological Complexity 7 (3), 260–272, http://dx.doi.org/10.1016/j.ecocom.2009.10.006.

Dowd, C., 1992. Effect of development on bird species composition of two urbanforested Wetlands in Staten Island, New York. Journal of Field Ornithology 63(4), 455–461.

Felix, M., Louise, W., Rudolf, D.G., 2011. Ecosystem Services at the landscape scale:the need for integrative approaches. Landscape Online, 1–11, http://dx.doi.org/10.3097/LO.201023.

Folke, C., 2006. Resilience: the emergence of a perspective for social–ecologicalsystems analyses. Global Environmental Change 16 (3), 253–267, http://dx.doi.org/10.1016/j.gloenvcha.2006.04.002.

Folke, C., Colding, J., Berkes, F., 2003. Synthesis: building resilience andadaptive capacity in social–ecological systems. In: Berkes, F., Colding, J.,Folke, C. (Eds.), Navigating SocialEcological Systems Building Resilience forComplexity and Change, vol. i. Cambridge University Press, New York, NY,pp. 352–387, http://dx.doi.org/10.1017/CBO9780511541957.020.

Gittleman, M., Librizzi, L., Stone, E., 2010. Community Garden Survey: New YorkCity Results 2009/2010. New York City.

Gottdiener, M., Hutchison, R., 2000. The New Urban Sociology 2nd edition.Stat. Westview Press, Boulder, CO, pp. 374–390 ⟨http://www.amazon.com/dp/0813343186⟩. (Retrieved from).

Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, J.M.,2008. Global change and the ecology of cities. Science 319 (5864), 756–760,http://dx.doi.org/10.1126/science.1150195.

Grimm, N.B., Grove, J.M., Pickett, S.T.A., Redman, C.L., 2000. Integrated approachesto long-term studies of urban ecological systems. BioScience 50 (7), 571–584,http://dx.doi.org/10.1641/0006–3568(2000)050.

Grove, J.M., Burch, W.R., Pickett, S.T.A., 2005. Social Mosaics and Urban CommunityForestry in Baltimore, Maryland Introduction: Rationale for Urban CommunityForestry Continuities from Rural to Urban Community Forestry. In: Lee, R.G.,Field, D.R. (Eds.), Communities and Forests where People Meet the Land.Oregon State University Press, Corvallis, OR, pp. 248–273.

Grove, J., Troy, A., O'Neil-Dunne, J., Burch, W.R., Cadenasso, M., Pickett, S.T.A., 2006.Characterization of households and its implications for the vegetation of urbanecosystems. Ecosystems 9 (4), 578–597.

Table B.1Summary of literature values used to estimate air pollution removal services oftrees, shrubs and herbaceous landcover.

Pollutant Trees Vegetation

SO2 1.32 g/m2yr Nowak and Crane(2002)

Short grass: 0.65 g/m2yr (Yanget al., 2008)

NO2 2.54 g/m2yr Nowak and Crane(2002)


PM10 2.73 g/m2yr Nowak and Crane(2002)


O3 3.06 g/m2yr Nowak and Crane(2002)

CO 0.58 g/m2yr Nowak and Crane(2002)

Table B.2Summary of SCS curve numbers by landcover.

No. Landcover\Hydrologic soil group A B C D

1 Residential 77 85 90 922 Open space poor (450% cover) 68 79 86 893 Open space fair (50–75% cover) 49 69 79 844 Open space good (o75% cover) 39 61 74 805 Paved 98 98 98 986 Bare soil 72 82 87 89



Haase, D., Schwarz, N., Strohbach, M., Kroll, F., Seppelt, R., 2012. Synergies, trade-offs, and losses of ecosystem services in urban regions: an integrated multiscaleframework applied to the Leipzig-Halle Region, Germany. Ecology and Society17 (3), http://dx.doi.org/10.5751/ES-04853–170322. (art22).

Hepcan, C.C., Ozkan, M.B., 2011. Establishing ecological networks for habitatconservation in the case of Çeşme-Urla Peninsula, Turkey. EnvironmentalMonitoring and Assessment 174 (1–4), 157–170.

Holling, C.S., 2001. Understanding the complexity of economic, ecological, and socialsystems. Ecosystems 4 (5), 390–405, http://dx.doi.org/10.1007/s10021-00.

Hörnsten, L., Fredman, P., 2000. On the distance to recreational forests in Sweden.Landscape and Urban Planning 51 (1), 1–10, http://dx.doi.org/10.1016/S0169-2046(00)00097-9.

Huang, Y.J., Akbari, H., Taha, H., Rosenfeld, A.H., 1987. The potential of vegetation inreducing summer cooling loads in residential buildings. Journal of Climate andApplied Meteorology, 26 (9), 1103–1116.

IPCC, 2011. Managing the Risks of Extreme Events and Disasters to Advance ClimateChange Adaptation. In: Field, C.B., Barros, V., Stocker, T.F., Qin, D., Dokken, D.,Ebi, K.L., Mastrandrea, M.D., et al. (Eds.), Cambridge University Press,pp. 1–29.

Jacobs, J., 1961. The Death and Life of Great American Cities. Random House, NewYork, NY.

Jo, H., McPherson, G., 1995. Carbon storage and flux in urban residential green-space. Journal of Environmental Management 45, 109–133.

Koniak, G., Noy-Meir, I., Perevolotsky, A., 2010. Modelling dynamics of ecosystemservices basket in Mediterranean landscapes: a tool for rational management.Landscape Ecology 26 (1), 109–124, http://dx.doi.org/10.1007/s10980-010-9540-8.

Kremer, P., Hamstead, Z.A., McPhearson, T. A social–ecological assessment of vacantlots in New York City. Landscape and Urban Planning. /http://dx.doi.org/10.1016/j.landurbplan.2013.05.003S, in press.

Kulshreshtha, S. N., Lac, S., Johnston, S., Kinar, C., 2000. Carbon sequestration inprotected areas of Canada: An economic valuation. (No. 549). Canada ParksCouncil; Economics Framework Project.

Larondelle, N., Hasse, D., 2012. Valuing post-mining landscapes using an ecosystemservices approach—an example from Germany. Ecological Indicators 18,567–574.

Longley, P.A., Goodchild, M.F., Maguire, D.J., Rhind, D.W., 2011. Geographic Informa-tion Systems and Science, Third edition, Hoboken, NJ.

Machlis, G., Force, J.E., Burch, W., 1997. The human ecosystem. Part I: The humanecosystem as an organizing concept in ecosystem management. Society NaturalResources 10 (4), 347–367, http://dx.doi.org/10.1080/08941929709381034.

McKinney, M.L., 2002. Urbanization, biodiversity, and conservation. BioScience 52(10), 883–890, http://dx.doi.org/10.1641/0006–3568(2002)052.

McPhearson, P.T., 2011. Toward a Sustainable New York City: GreeningThrough Urban Forest Restoration. In: Slavin, M.I. (Ed.), The Triple BottomLine Sustainability Principles Practice and Perspective in America's Cities. ;Island Press/Center for Resource Economics, Washington, D.C.,pp. 181–204 ⟨http://www.amazon.com/Sustainability-Americas-Cities-Creating-Metropolis/dp/1597267422⟩. (Retrieved from).

McPhearson, P.T., Feller, M., Felson, A., Karty, R., Lu, J.W.T., Palmer, M.I., Wenskus, T.,2010. Assessing the Effects of the Urban Forest Restoration Effort of Million-TreesNYC on the Structure and Functioning of New York City Ecosystems. Citiesand the Environment 3 (1), 1–21.

McPherson, E., Gregory, Simpson, James R., 1999. Carbon dioxide reduction throughurban forestry: Guidelines for professional and volunteer tree planters. GeneralTechnical Report. PSW-GTR-171. Pacific Southwest Research Station, ForestService, U.S. Department of Agriculture, Albany, CA. 237 p.

Mennis, J., 2002. Using geographic information systems to create and analyzestatistical surfaces of population and risk for environmental justice analysis.Social Science Quarterly 83 (1), 281–297, http://dx.doi.org/10.1111/1540–6237.00083.

Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being:Synthesis. World Health. Island Press, Washington, DC.

Müller, F., Willemen, L., DeGroot, R., 2011. Ecosystem services at the landscapescale: the need for integrative approaches. Landscape Online, 1–11. http://dx.doi.org/10.3097/LO.201023.

Naidoo, R., Balmford, a, Costanza, R., Fisher, B., Green, R.E., Lehner, B.,Malcolm, T.R., et al., 2008. Global mapping of ecosystem services and con-servation priorities. Proceedings of the National Academy of Sciences of theUnited States of America 105 (28), 9495–9500, http://dx.doi.org/10.1073/pnas.0707823105.

Nelson, E., Polasky, S., Lewis, D.J., Plantinga, A.J., Lonsdorf, E., White, D., Bael, D.,et al., 2008. Efficiency of incentives to jointly increase carbon sequestration andspecies conservation on a landscape. Proceedings of the National Academy ofSciences of the United States of America 105 (28), 9471–9476, http://dx.doi.org/10.1073/pnas.0706178105.

New York City, 2010. NYC Green Infrastructure Plan: A Sustainable Strategy forClean Waterways. New York, NY.

New York City Panel on Climate Change, 2009. Climate Risk Information: New YorkCity Panel on Climate Change.

New York City Soil Survey Staff, 2005. New York City Reconnaissance Soil Survey.New York, NY.

Niemelä, J., Saarela, S.-R., Söderman, T., Kopperoinen, L., Yli-Pelkonen, V., Väre, S., Kotze,D.J., 2010. Using the ecosystem services approach for better planning and conserva-tion of urban green spaces: a Finland case study. Biodiversity & Conservation 19(11), 3225–3243, http://dx.doi.org/10.1007/s10531-010-9888-8.

Nowak, D.J., Crane, D.E., 2002. Carbon storage and sequestration by urban trees inthe USA. Environmental Pollution 116, 381–389.

NYC Department of City Planning, 2011. MapPLUTOTM.Nowak, D.J., Crane, D.E., Stevens, J.C., Ibarra, M., 2002. Brooklyn's Urban Forest.NYSDEC, 2009. Natural Heritage Community Occurrences. Retrieved from: ⟨http://

www.dec.ny.gov/geodata/DiscoveryServlet⟩.Olsson, P., Folke, C., Berkes, F., 2004. Adaptive comanagement for building resilience

in social–ecological systems. Environmental Management 34 (1), 75–90, http://dx.doi.org/10.1007/s00267-003-0101-7.

Parent, J., 2011. Shape Metrics. ArcGIS Resource Center.Pennsylvania Horticulture Society, 1995. Urban Vacant Land.pdf.Peterson, G., 2000. Political ecology and ecological resilience: an integration of

human and ecological dynamics. Ecological Economics 35, 323–336.Pham, T.T.H., Apparicio, P., Séguin, A.M., Landry, S., Gagnon, M., 2012. Spatial

distribution of vegetation in Montreal: an uneven distribution or environmen-tal inequity? Landscape and Urban Planning 107 (3), 214–224, http://dx.doi.org/10.1016/j.landurbplan.2012.06.002.

Pickett, S.T.A., Belt, K.T., Galvin, M.F., Groffman, P.M., Morgan, J., Outen, D.C., Pouyat,R.V, et al., 2007. Watersheds in Baltimore, Maryland: understanding andapplication of integrated ecological and social processes. Journal of Contem-porary Water Research & Education 136, 44–55.

Pickett, S.T.A., Cadenasso, M.L., Grove, J.M., Groffman, P.M., Band, L.E., Boone, C.G.,Burch, W.R., et al., 2008. Beyond urban legends: an emerging framework ofurban ecology, as illustrated by the Baltimore Ecosystem study. AmericanInstitute of Biological Sciences BioScience 58 (2), 139–150.

Pickett, S.T.A., Cadenasso, M.L., Grove, M.J., 2004. Resilient cities: meaning, models,and metaphor for integrating the ecological, socio-economic, and planningrealms. Landscape and Urban Planning 69, 369–384.

Pickett, S.T.A., Grove, J.M., 2009. Urban ecosystems: what would Tansley do? UrbanEcosystems 12 (1), 1–8.

Pickett, S.T.A., Grove, J.M., Nilon, C.H., Pouyat, R.V., Zipperer, W.C., Costanza, R.,2001. Urban ecological systems: linking terrestrial ecological, physical, andsocioeconomic components of metropolitan areas. (Statistical Data Included).Annual Review of Ecology and Systematics, 32.

Pouyat, R., Groffman, P., Yesilonis, I., Hernandez, L., 2002. Soil carbon pools andfluxes in urban ecosystems. Environmental Pollution 116 (Suppl), S107–S118.(Retrieved from)⟨http://www.ncbi.nlm.nih.gov/pubmed/11833898⟩.

Ranara, J., Ståhle, A., Zetterberg, A., Mörtberg, U., Liljeros, F., Nagendra, H., Telenius,A., K.G.T, T.E., 2011. Mapping of Social–ecological Values in Stockholm:Implications for Ecosystem Services, and Urban Transition Planning. Interna-tional Association of Landscape Ecologists, Laufen, Germany.

Rockström, J., Steffen, W., Noone, K., Lambin, E., Lenton, T.M., Scheffer, M., Folke, C.,et al., 2009a. Planetary boundaries: exploring the safe operating space forhumanity. Ecology and Society 14 (2), 32.

Rockström, J., Steffen, W., Noone, K., Persson, A., Chapin, F.S., Lambin, E.F.,Lenton, T.M., et al., 2009b. A safe operating space for humanity. Nature 461(7263), 472–475.

Seto, K.C., Fragkias, M., Güneralp, B., Reilly, M.K., 2011. A meta-analysis of globalurban land expansion. PloS one 6 (8), e23777, http://dx.doi.org/10.1371/Citation.

Seto, K.C., Reenberg, a., Boone, C.G., Fragkias, M., Haase, D., Langanke, T.,Marcotullio, P., et al., 2012. Urban land teleconnections and sustainability.Proceedings of the National Academy of Sciences, 1–6, http://dx.doi.org/10.1073/pnas.1117622109.

Smith, R.L., Smith, T.M., 2006. Elements of Ecology. Benjamin/Cummmings, NewYork.

Sukhdev, P., Wittmer, H., Schröter-Schlaack, C., Nesshöver, C., Bishop, J., Ten, P.,Brink, H.G., et al., 2010. The Economics of Ecosystems and Biodiversity:Mainstreaming the Economics of Nature. A Synthesis of the Approach, Conclu-sions and Recommendations of TEEB. Environment.

Talen, E., 2010. The spatial logic of parks. Journal of Urban Design 15 (4), 473–491,http://dx.doi.org/10.1080/13574809.2010.502335.

Tallis, H., Polasky, S., 2009. Mapping and valuing ecosystem services as an approachfor conservation and natural-resource management. Annals of the New YorkAcademy of Sciences 1162, 265–283, http://dx.doi.org/10.1111/j.1749-6632.2009.04152.x.

TEEB, 2011. The Economics of Ecosystem and Biodiversity: TEEB Manual for Cities:Econsystem Services in Urban Management. Retrieved from: ⟨http://www.teebweb.org/Portals/25/Documents/TEEB_Manual_for_Cities_Ecosystem_Services_for_Urban_managment___FINAL_2011.pdf⟩.

The Nature Conservancy, n.d.. TNC Maps. Ecoregional Portfolio.Tidball, K.G., Krasny, M.E., 2010. Urban environmental education from a

conceptual framework for civic ecology education. Cities and the Environment3 (1), 1–20.

Troy, A.R., Grove, J.M., O'Neil-Dunne, J.P.M., Pickett, S.T.A., Cadenasso, M.L., 2007.Predicting opportunities for greening and patterns of vegetation on privateurban lands. Environmental management 40 (3), 394–412, http://dx.doi.org/10.1007/s00267-006-0112-2.

Tratalos, J., Fuller, R.A., Warren, P.H., Davies, R.G., Gaston, K.J., 2007. Urban form,biodiversity potential and ecosystem services. Landscape and Urban Planning83, 308–317.

Turner, M., 1989. Landscape ecology: the effect of pattern on process. AnnualReview of Ecology and Systematics (1), 171–197. http://dx.doi.org/10.1146/annurev.ecolsys.20.1.171.

US Census, n.d.. New York (City) Quickfacts from the US Census Bureau.USDA, 1986. Urban Hydrology for Small Watersheds TR-55.



Walker, B.H., Anderies, J.M., Kinzig, A.P., Ryan, P., 2006. Exploring resiliencein social–ecological systems through comparative studies and theory develop-ment: introduction to the special issue. In: Walker, B H (Ed.), Ecology andSociety, 11; 2006, p. 12.

Weber, T., Sloan, A., Wolf, J., 2006. Maryland's Green Infrastructure Assessment:development of a comprehensive approach to land conservation. Landscapeand Urban Planning 77 (1–2), 94–110, http://dx.doi.org/10.1016/j.landurbplan.2005.02.002.

Whitford, V., Ennos, A.R., Handley, J.F., 2001. City form and natural processindicators for the ecological performance of urban areas and their applicationto Merseyside, UK. Landscape and Urban Planning 57, 91–103.

Yang, J., Yu, Q., Gong, P., 2008. Quantifying air pollution removal by green roofs inChicago. Atmospheric Environment 42 (31), 7266–7273, http://dx.doi.org/10.1016/j.atmosenv.2008.07.003.


Mapping ecosystem services in New York City: Applying a social–ecological approach in urban vacant...

Documents

Transcript of Mapping ecosystem services in New York City: Applying a social–ecological approach in urban vacant...