Mapping supply of and demand for ecosystem services to assess environmental justice in New York City

Abstract

Livability, resilience, and justice in cities are challenged by climate change and the historical legacies that together create disproportionate impacts on human communities. Urban green infrastructure has emerged as an important tool for climate change adaptation and resilience given their capacity to provide ecosystem services such as local temperature regulation, stormwater mitigation, and air purification. However, realizing the benefits of ecosystem services for climate adaptation depend on where they are locally supplied. Few studies have examined the potential spatial mismatches in supply and demand of urban ecosystem services, and even fewer have examined supply-demand mismatches as a potential environmental justice issue, such as when supply-demand mismatches disproportionately overlap with certain socio-demographic groups. We spatially analyzed demand for ecosystem services relevant for climate change adaptation and combined results with recent analysis of the supply of ecosystem services in New York City (NYC). By quantifying the relative mismatch between supply and demand of ecosystem services across the city we were able to identify spatial hot- and coldspots of supply-demand mismatch. Hotspots are spatial clusters of census blocks with a higher mismatch and coldspots are clusters with lower mismatch values than their surrounding blocks. The distribution of mismatch hot- and coldspots was then compared to the spatial distribution of socio-demographic groups. Results reveal distributional environmental injustice of access to the climate-regulating benefits of ecosystem services provided by urban green infrastructure in NYC. Analyses show that areas with lower supply-demand mismatch tend to be populated by a larger proportion of white residents with higher median incomes, and areas with high mismatch values have lower incomes and a higher proportion of people of color. We suggest that urban policy and planning should ensure that investments in "nature-based" solutions such as through urban green infrastructure for climate change adaptation do not reinforce or exacerbate potentially existing environmental injustices.

Keywords: cities; climate change adaptation; regulating ecosystem services; resilience; spatial analysis; urban ecosystem services.

Fig. 1

Red lining map of New York City created by the Home Owners’ Loan Corporation during the 1930s. Residential neighborhoods were given a mortgage security grade that reflected the security of a potential investment made by banks and other mortgage lenders. While grade A refers to low risk areas, grade D refers to areas qualified as “hazardously” risky. Data and description obtained from (Nelson et al. n.d.).

Fig. 2

Ratio of total population classified as white according to the 2010 decennial census, per census block (U.S. Census Bureau n.d. , b). Additionally, the distribution of parks in New York City (Department of Information Technology & Telecommunications (DoITT) n.d.) is included to depict the spatial correlation between race and public green space visually.

Fig. 3

Median income per census block according to the American Community Survey 5‐yr estimates 2013–2017 (U.S. Census Bureau n.d. , b), with the distribution of public green (Department of Information Technology & Telecommunications (DoITT) n.d.) spaces overlapped. Values range from 0 (no income) to 1.0 (maximum estimated income in New York City).

Fig. 4

Map showing the dominant race (race with highest percentage) per census block according to the 2010 decennial census (U.S. Census Bureau n.d. , b).

Fig. 5

Supply, demand, and mismatch maps at the census block level for the ecosystem service (ES) local temperature regulation. A composite figure; the rest of the ecosystem services assessed can be found in Appendix S1: Fig. S2.

Fig. 6

Accumulated mismatch value of the four ecosystem services assessed and close ups. Important spatial nuances can be differentiated. For example, Midtown Manhattan (1) shows relatively low mismatch values because of the low population of census blocks occupied by office buildings, whereas the Upper East and West Side show high mismatch values, despite their wealthy population. In Bronx (2), a clear gradient by which central Bronx presents higher values than the edges of the borough. In Queens (3), the influence of parks in reducing the mismatch value of nearby blocks is visible in blocks like those situated in the center of the image.

Fig. 7

Census blocks classified into hotspots and coldspots according to the Z‐score obtained in the cluster analysis for the ecosystem service (ES) local temperature regulation. Clusters range from C1 (extreme lows, or coldspots) to C5 (extreme highs, or hotspots). Even though break values between different categories were set using a Jenks distribution, values 2.58 and −2.58 were set manually in order to keep a minimum degree of significance (P < 0.01). C3 corresponds to those census blocks that obtained a Z‐score between 2.58 and −2.58, meaning that their P value is >0.01. A composite figure with the rest of the ES assessed can be found in Appendix S1: Fig. S4.

Fig. 8

Average proportion of people of color over the total population and relative income per mismatch cluster for the service local temperature regulation. Clusters range from C1 (extreme lows, or coldspots) to C5 (extreme highs, or hotspots). C3 refers to census blocks that do not belong to a high or low cluster based on statistical significance at P > 0.01. Latin and Greek letters indicate statistical significance across the clusters as per the ANOVA tests carried out. All the statistical significance tests returned P values below 0.001. A composite figure with the rest of the ecosystem services assessed can be found in Appendix S1: Fig. S5.

Fig. 9

Average proportion of disaggregated people of color, per mismatch cluster for the service local temperature regulation. Latin and Greek letters and numbers indicate statistical significance across the clusters as per the ANOVA tests carried out. All the statistical significance tests returned P values below 0.001, except for the proportion of residents being Asian when comparing C2–C3, C2–C4, C4–C5, and C4–C3 (P < 0.01). A composite figure with the rest of the ecosystem services assessed can be found in Appendix S1: Fig. S6.