ABCDEFGHIJKLMNOPQRSTUVWX
1
Smart Surfaces Decision Support Tool (DST) Data Layers
2
Data NameDescriptionData Source
3
Geographies
4
Census tractsAll census tracts within the regional planning areas, including a variety of attributes relating to health, environmental, and economic factors. Health data from CDC Places were incorporated. CDC PLACES is a collaboration between CDC, the Robert Wood Johnson Foundation, and the CDC Foundation. PLACES provides health data for small areas across the country. This allows local health departments and jurisdictions, regardless of population size and rurality, to better understand the burden and geographic distribution of health measures in their areas and assist them in planning public health interventions.

Attributes used from the Justice40 Climate and Economic Justice Screening Tool (CEJST) include rows 6- 10 listed below in this data dictionary (e.g. energy burden, percentile, housing burden, percentile, etc.)

All values not marked here as percentiles are given as percentages out of 100.

For edge cases where census tracts are not fully covered by raster data, the summary statistics apply only to the area that is covered. Census tracts should be inspected prior to interpretation.		United States Census TIGER/Line shapefiles, CDC PLACES 2025, CEJST screening tool
5
Energy burden, percentile. Indicates whether a community faces high energy and air pollution costs relative to the nation. A census tract is considered to have an energy burden if:

1) It ranks at or above the 90th percentile nationally for either household energy costs or fine particulate matter (PM2.5) concentrations;

AND

2) It qualifies as low income, meaning it falls at or above the 65th percentile nationally in the share of households earning ≤200% of the federal poverty level (excluding college students).
CEJST screening tool, accessed August 2024 (archived documentation available)
6
Housing burden, percentile. Identifies communities facing high housing-related costs or poor housing conditions. A census tract is considered to have a housing burden if:

1) It ranks at or above the 90th percentile nationally for any of the following: housing costs, lack of green space, lack of indoor plumbing, presence of lead paint, or historic underinvestment;

AND

2) It qualifies as low income, meaning it falls at or above the 65th percentile nationally in the share of households earning ≤200% of the federal poverty level (excluding college students).		CEJST screening tool, accessed August 2024 (archived documentation available)
7
Areas of persistent poverty. Defined as census tracts maintaining poverty rates of 20% or higher during the three decades from 1989 to 2015-2019, were incorporated from the U.S. Department of Transportation MPDG 2025-2026 Mega Grant program site. DOT MPDG
8
% Low median household income as a percent of area median income, percentile. Represents how a community’s median household income compares to the median income of the surrounding area, adjusted for regional cost of living. Expressed as a percentage of the area median income (AMI), where lower percentages indicate lower relative income levels. In CEJST, this variable is percentile-ranked so that tracts with lower relative incomes correspond to higher burden percentiles.CEJST screening tool, accessed August 2024 (archived documentation available)
9
% Individuals below 200% Federal Poverty Line, imputed and adjusted, percentile. Represents the share of residents living in households with incomes at or below twice the federal poverty level, excluding full-time college and university students. Where census data are missing, values are imputed based on the average of adjacent tracts (or county/state averages when needed). This adjustment ensures more accurate representation of low-income populations in communities with incomplete data.CEJST screening tool, accessed August 2024 (archived documentation available)
10
% Unemployment, percentile. Represents the share of the civilian labor force (individuals ages 16 and older who are working or actively seeking work) that is unemployed, based on U.S. Census American Community Survey data. In CEJST, higher unemployment rates correspond to higher socioeconomic burden percentiles.CEJST screening tool, accessed August 2024 (archived documentation available)
11
Core CityMunicipal boundaries of the ten core Smart Surfaces cities included in the Cities for Smart Surfaces Program - Atlanta, Boston, Charlotte, Columbia SC, Dallas, Jacksonville, New Orleans, Phoenix, Portland OR, and San Antonio.US Census incoporated places, 2020
12
Core city council districtsCity council district boundaries for the ten core cities of Atlanta, Boston, Charlotte, Columbia SC, Dallas, Jacksonville, New Orleans, Phoenix, Portland OR, and San Antonio.Municipal governments and data hubs of core Smart Surfaces cities
13
Municipalities across Census-Defined Urban AreaCensus-designated municipalities that intersect the census-defined urban areas of the ten core Smart Surfaces cities. Geographic boundaries are derived from the U.S. Census Bureau’s 2020 incorporated places dataset and include incorporated places, consolidated cities, and census-designated places (CDPs) that fall within the urban area for each core city. County subdivisions are excluded.

Note: In certain states (e.g., those in New England), many towns are defined as county subdivisions rather than municipalities and are therefore not included in this layer.		US Census incorporated places, 2020
14
CountiesCounties which intersect with the regional planning areas. This layer is for overlay purposes only, dataset coverage for other layers does not extend across all counties included.United States Census TIGER/Line shapefiles, 2020
15
Census-defined urban areasBoundaries of the urban areas associated with the core municipalitiesUS Census urban areas, 2020
16
Regional planning areasRegional boundaries as designated by national and other local entities. For seven of the ten core municipalities, regional planning area boundaries reflect the Metropolitan Planning Organization (MPO) boundary published by the Bureau of Transportation Statistics (BTS) in their National Transportation Atlas Database (NTAD). For the other three cities, regional planning area boundaries were created to encompass all counties included within the following local planning districts:

Columbia, SC : Central Midlands Council of Governments
Charlotte, NC: Centralina Economic Development District
Jacksonville, FL: Northeast Florida Regional Council		U.S. Department of Transportation. (2025). Metropolitan Planning Organizations (MPOs) [Data set]. Bureau of Transportation Statistics, National Transportation Atlas Database.
Source
Accessed December 2024.
17
Overlay data
18
Parcels Land parcels as designated by counties, cities, and other unitary authorities. For edge cases where parcels are not fully covered by raster data, the summary statistics apply only to the area that is covered. Parcels should be inspected prior to interpretation.Regrid, June 2024 release
19
Publicly-owned parcelsLocations of publicly-owned or leased land throughout regional planning areas. A subset of Parcels, above.Derived from Regrid, June 2024 release & City-specific data layers last updated between 2024-2025
20
City-owned parcelsLocations of city-owned or leased land throughout core municipalities, as well as city-owned or leased land outside of municipality boundaries, if applicable. A subset of Publicly-owned parcels, above. Derived from Regrid, June 2024 release & City-specific data layers last updated between 2024-2025
21
K-12 Schools (Public & Private)Locations of K-12 schools, both public and private, in the United States, clipped to areas of interest.Homeland Infrastructure Foundation-Level Data (HIFLD), National Urban Institute, Trust for Public Land (2023)
22
Areas within a 10 minute walk to schoolsUsing the actual road and walking network, these polygons show areas around each K-12 school within a 10-minute walk for an average adult. They do not account for the presence/absence of sidewalks, but do account for non-walkable roads such as major highways and other barriers. They were developed using network analysis tools in ArcGIS Pro.Trust for Public Land, 2025
23
Areas within a 20 minute walk to schoolsUsing the actual road and walking network, these polygons show areas around each K-12 school within a 20-minute walk for an average adult. They do not account for the presence/absence of sidewalks, but do account for non-walkable roads such as major highways and other barriers. They were developed using network analysis tools in ArcGIS Pro.Trust for Public Land, 2025
24
ParksPublic parks and publicly accessible open space, clipped to areas of interest.Trust for Public Land's Parkserve database (accessed July 2024)
25
Areas within a 10 minute walk to parksUsing the actual road and walking network, these polygons show areas around each park within a 10-minute walk for an average adult. They do not account for the presence/absence of sidewalks, but do account for non-walkable roads such as major highways. They take into account park entrances and exits, as well as inaccessible sides of parks due to fencing or other barriers. These polygons were developed using network analysis tools in ArcGIS Pro.Trust for Public Land's Parkserve database (accessed June 2025)
26
Historic redliningThis feature layer is a digital representation of areas drawn on 1930s redlining maps (HOLC, Home Ownership Loan Corporation) that were georeferenced and traced by the University of Richmond Digital Scholarship Lab. This is the third and most up-to-date version of the Mapping Inequality spatial dataset. This dataset includes residential areas (graded A through D) and non-residential areas.

Neighborhoods they deemed "best" and safe investments were given a grade of A and colored green. Those that were deemed "hazardous" were given a grade of "D" and colored red. Non-residential areas were not graded, but simply designated as Commercial, Industrial, or Industrial/Business. 		University of Richmond Digital Scholarship Lab (accessed March 2024)
27
Sidewalks (not available for all cities)Locations of sidewalks throughout core municipalities. Not available for Jacksonville, New Orleans, Phoenix, or suburban municipalities. Cities of Atlanta, Boston, Charlotte, Columbia, Dallas, Portland, and San Antonio (2022-2024)
28
Historically Disadvantaged CommunitiesHighlights Census tracts designated as Historically Disadvantaged Communities by the Climate and Economic Justice Screening Tool (CEJST). CEJST identifies disadvantaged communities through eight categories of disadvantaged status including climate change, energy, health, housing, legacy pollution, transportation, water and wastewater, and workforce development. Census tracts — home to an average 4,000 residents — are identified as disadvantaged if they meet 90th percentile thresholds for indicators within any of the eight categories and are at or above the 65th percentile for low-income. CEJST screening tool, accessed August 2024 (archived documentation available)
29
Tree Equity ScoreTree Equity Score measures how well the benefits of trees are reaching communities living on low-incomes, communities of color and others disproportionately impacted by extreme heat and other environmental hazards.

Tree Equity Score is a nationwide score that highlights inequitable access to trees.
The score is calculated at the neighborhood (Census block group) level.
The score ranges from 0-100. The lower the score, the greater priority for tree planting. A score of 100 means the neighborhood has enough trees.
Tree Equity Score covers every urban Census block group in the United States, including Hawaii and Alaska. Data for Puerto Rico and the U.S. Virgin Islands are due for release in 2024. Detailed methodology.		American Forests (accessed March 2024)
30
Surfaces
31
Land Use / Land CoverWorld Resources Institute (WRI) created the OpenUrban land use/land cover (LULC) data for the 10 U.S. cities participating in the Smart Surfaces Coalition project. The LULC V3 data is primarily derived from Open Street Maps (OSM), along with other ancillary datasets. Public open space, water, roads, buildings, and parking lots are acquired from OSM and converted to rasters with a 1-m spatial resolution. Where there are gaps in the coverage, values are filled in from the European Space Agency (ESA) Worldview dataset. The width of roads is estimated from the number of lanes, assuming an average urban lane width of 10 ft (National Association of City Transportation Officials. The buildings data is supplemented with the Google-Microsoft Open Buildings dataset and roof slope is predicted based on whether the building is in a residential or non-residential area, the average height of the surrounding buildings, and the building footprint size. Additional details are available here.World Resources Institute (WRI) (2025)
32
Albedo (reflectivity)Median albedo was calculated using Sentinel-2 imagery from summer (June, July, August) 2023. Images were filtered for cloud cover using the Cloud Score+ data product (threshold: 0.60) in Google Earth Engine.
Surface albedo (𝛼) was estimated using the narrow-to-broadband conversion coefficients presented in Bonafoni and Sekertekin (2020), as follows:

𝛼 = 𝐵2 × 0.2266 + 𝐵3 × 0.1236 + 𝐵4 × 0.1573 + 𝐵8 × 0.3417 + 𝐵11 × 0.1170 + 𝐵12 × 0.0338

Where:
B2, B3, B4 = visible blue, green, red
B8 = near-infrared (NIR)
B11, B12 = shortwave infrared (SWIR)

Albedo ranges from 0 to 1. Values greater than 1 were capped at 1. In this process the SWIR bands (20-m resolution) are downscaled to match the spatial resolution of the visible bands (10-m). The albedo rasters contain the pixel-wise medians for summer 2023 at a spatial resolution of 10-m.
Bonafoni, S., & Sekertekin, A. (2020). Albedo Retrieval From Sentinel-2 by New Narrow-to-Broadband Conversion Coefficients. IEEE Geoscience and Remote Sensing Letters, 17(9), 1618–1622. https://doi.org/10.1109/LGRS.2020.2967085

Pasquarella, V. J., Brown, C. F., Czerwinski, W., & Rucklidge, W. J. (2023) Comprehensive Quality Assessment of Optical Satellite Imagery Using Weakly Supervised Video Learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 2124-2134).

Copernicus Sentinel-2 data (2023). Retrieved from Google Earth Engine. Accessed 01 July 2024, processed by ESA.
33
Tree canopyTree cover was derived from the WRI-Meta High Resolution Canopy Height dataset, accessed via Google Earth Engine. The Global Canopy Height Maps dataset offers comprehensive insights into tree canopy presence and heights worldwide for the period 2009-2020, with approximately 80% of the source imagery acquired between 2018 and 2020.

To provide a conservative estimate focused on mature canopy that contributes shade, smaller trees and shrubs were excluded from the analysis. Pixels with canopy heights greater than 3 meters were classified as tree-covered (value = 1), while all others were set to 0. This produced a binary map of canopy presence/absence at 1-m spatial resolution.
High Resolution Canopy Height Maps (CHM). Meta and World Resources Institute (WRI) - 2023. Source imagery for CHM © 2016 Maxar. Accessed 01 July 2024.
34
Fractional vegetationFractional vegetation (Fr), representing the fraction of a pixel covered by green vegetation, was calculated from the Normalized Difference Vegetation Index (NDVI). NDVI was derived from cloud-free Sentinel-2 observations across the full calendar year, and pixel-wise 90th percentile composites were used to reduce seasonal noise. Following Carlson and Ripley (1997), fractional vegetation was computed as:

Fr = ((NDVI90 − NDVInonveg) / (NDVIveg− NDVInonveg))^2
where NDVIveg and NDVInonveg represent the NDVI of fully vegetated and completely unvegetated surfaces respectively. Using percentile values from Gao et al. (2020), we took as endmembers the 75th-percentile NDVI90 value of all pixels classified in the Dynamic World vegetation classification dataset (Brown et al. 2022) as either tree, grass, or scrub and shrub for NDVIveg, and for NDVInonveg, the 5th-percentile NDVI90 for pixels classified as built-up in Dynamic World. The pixel-wise Fr raster contains fractional vegetation values derived from these percentile composites at 10-m spatial resolution.
Brown, C.F., S.P. Brumby, B. Guzder-Williams, T. Birch, S.B. Hyde, J. Mazzariello, W. Czerwinski, et al. 2022. “Dynamic World, Near Real-Time Global 10 M Land Use Land Cover Mapping.” Scientific Data 9 (1): 251.

Carlson, T.N., and D.A. Ripley. 1997. “On the Relation between NDVI, Fractional Vegetation Cover, and Leaf Area Index.” Remote Sensing of Environment 62 (3): 241–52. https://doi.org/10.1016/S0034- 4257(97)00104-1.

Gao, L., Wang, X., Johnson, B. A., Tian, Q., Wang, Y., Verrelst, J., Mu, X., & Gu, X. (2020). Remote sensing algorithms for estimation of fractional vegetation cover using pure vegetation index values: A review. ISPRS Journal of Photogrammetry and Remote Sensing, 159, 364–377. https://doi.org/10.1016/j.isprsjprs.2019.11.018
35
Heating and flooding
36
City Cooling Potential (deg F)Air-temperature reductions were modeled by Altostratus, Inc. using a customized, high-resolution WRF-ARW urban climate model (250–500 m). For each city, the model simulated the single hottest month in the hottest recent year (2010–2023), using only meteorologically stable days to isolate the temperature impacts of Smart Surfaces. Baseline and intervention conditions were defined using remotely sensed land cover, high-resolution urban morphology (1–30 m), and location-specific weather inputs.

The cooling layer shown here represents a single peak-afternoon snapshot—a time slice from one modeled day, not an average across all summer days. It reflects direct, localized cooling under conditions when benefits are expected to be greatest.

Smart Surface strategies were modeled using the following citywide deployment thresholds:

Cool roofs: 100% of mapped rooftop area, with aged albedo targets of 0.76 (low-slope) and 0.27 (steep-slope).

Cool pavements: 100% of mapped roads and parking lots, with an aged albedo target of 0.30.

Urban trees: Canopy cover in developed areas increased to each city’s 75th-percentile vegetation benchmark, limited by available plantable space within each grid cell.

Additional information on the general approach and methodology, is available here.		Altostratus, Inc. (2025)
37
City Air Temperature (deg F)Baseline air temperature was modeled by Altostratus, Inc. using a customized, high-resolution version of the WRF-ARW atmospheric model (250–500 m) to simulate urban canopy and boundary-layer dynamics. The results reflect citywide heat exposure under current land use and land cover conditions during the hottest summer month in a recent hot year (2010–2023). The snapshot shown captures modeled air temperature at one of the hottest hours on a particularly hot summer day, offering a high-resolution view of peak afternoon heat. This is a single-point-in-time estimate—not an average across days or years—and serves as a baseline reference for evaluating Smart Surface cooling scenarios. Additional information on the general approach and methodology, is available here. Altostratus, Inc. (2025)
38
Metro Cooling Potential (deg F, 1km resolution)Air temperature reductions were modeled by Altostratus, Inc. using a customized, high-resolution version of the WRF-ARW atmospheric model (1km) to simulate urban canopy and boundary-layer dynamics. Each metro was modeled for the single hottest month in the hottest recent year (2010–2023), using only meteorologically stable days to isolate Smart Surface impacts. Baseline conditions and scenario-based perturbations were informed by remotely sensed land cover, high-resolution urban morphology (1–30 m), and location-specific weather inputs. The results reflect the direct cooling effects of technically feasible, citywide deployment of cool roofs, cool pavements, and urban trees. This snapshot illustrates peak air temperature reduction for an afternoon hour when benefits are expected to be among the greatest. Secondary downwind effects were excluded to provide a conservative estimate of direct, localized cooling. Additional information on the general approach and methodology, is available here.Altostratus, Inc. (2025)
39
Metro Air Temperature (deg F, 1km resolution)Baseline air temperature was modeled by Altostratus, Inc. using a customized, high-resolution version of the WRF-ARW atmospheric model (1k m) to simulate urban canopy and boundary-layer dynamics. The results reflect metrowide heat exposure under current land use and land cover conditions during the hottest summer month in a recent hot year (2010–2023). The snapshot shown captures modeled air temperature at one of the hottest hours on a particularly hot summer day, offering a high-resolution view of peak afternoon heat. This is a single-point-in-time estimate—not an average across days or years—and serves as a baseline reference for evaluating Smart Surface cooling scenarios. Additional information on the general approach and methodology, is available here. Altostratus, Inc. (2025)
40
Summer Land Surface Temperature 2022-2024 (deg F)Median land surface temperature (LST) was calculated from Landsat 8 satellite data over June, July, and August for 2022-2024 using Google Earth Engine.
To ensure sufficient clear-sky observations, only images with <20% cloud cover were selected. Cloud pixels were removed (masked) using the QA band. Because Landsat’s revisit time is 8 days, a 3-year window was used to generate robust summer medians.
LST is originally measured at 100-meter resolution and was downscaled to 30 meters to match other bands. The final raster contains median summer LST values per pixel for the 2022–2024 period.
Landsat 8 data (2022-2024). Retrieved from Google Earth Engine. Accessed 01 October 2024, processed by the U.S. Geological Survey.
41
Pluvial flooding risk screening indexA composite screening index developed by TPL to indicate relative pluvial (rainfall-related) flood susceptibility. Not a hydraulic model or probabilistic flood map. Areas appearing blank reflect missing input data (often city centers with limited SSURGO coverage) and should not be interpreted as low risk.

This screening index integrates five physical datasets: imperviousness, flow accumulation, fractional vegetation, soil infiltration,and sinks. All inputs were reclassified. For impervious surfaces, reclassification was done on an equal-interval basis between the maximum and minimum values across all urban areas. The reclassification range was 0-5. For sinks, sink areas were given a value of 5, and non-sinks were given a value of 0. For Soil Hydrologic Group, Group A was given a value of 1, Group B was given a value of 2, Group C was given a value of 3, and Group D was given a value of 4. Groups A/D, B/D, and C/D were given the same values as A, B, and C accordingly as urban areas are generally hydrologically drained where possible, especially in more densely populated areas. Flow accumulation was reclassified from 1-5 on an equal area basis to differentiate areas of subtly different accumulation, as opposed to highlighting existing water or stream beds. Areas of no data for any dataset were given a value of 0. Note that SSURGO data is absent in some portions of some project urban areas. The reclassified values were combined to create a pluvial flood risk index ranging from 4 to 20. Values were then split by Jenks natural breaks into categories of Low, Moderate, High, and Very High Risk. Areas with higher index values can be considered to have higher pluvial flood risk based on fundamental characteristics. Final resolution for the output index is 30-meter.		Trust for Public Land, USGS Earth
Explorer 1/3 arc-second DEM, Soil Survey Geographic Database (SSURGO) Hydrologic Soil Group, Multi-Resolution Land Characteristics Consortium NLCD imperviousness (2022)
42
Riverine flood riskThis data represents floodways, 100-year flood zones, and 500-year flood zones.Federal Emergency Management Agency. 2022. National Flood Hazard Layer, Version 1.1.1.0. Washington, D.C. Retrieved: July 17, 2022, from https://msc.fema.gov
43
City-Specific Data
44
Jacksonville - Core to Coast TrailA potential alignment for a planned pedestrian and bike trail through JacksonvilleCity of Jacksonville (2024)
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100