United States Surface Urban Heat Island database

This dataset contains urban and rural LST, DEM, and NDVI data for annual, summer, and winter daytime and nighttime for all census tracts in US urbanized areas, as well as the mean values for the entire urbanized area.

METADATA
DEM: Digital Elevation Model
NDVI: Normalized Difference Vegetation Index
LST: Land Surface Temperature
_urb: Urban values (all urban pixels within urbanized areas)
_rur: Rural reference (Spatial mean of the non-urban, non-water pixels within the region of interest)
Regions of Interest:
_CT: Spatial mean of pixels intersecting the Census Tract clipped to the urbanized area (one value per census tract). This should be equal to the _CT for census tracts that are completely within the urbanized areas (the census tracts with the green dots in the image below)
_all: Spatial mean of all pixels intersecting the urbanized area, as defined by the US census (one value for one urbanized area)
_CT_act: Spatial mean of all available pixels intersecting the Census Tract (one value per census tract) [This should be equal to the previous values I calculated]

For the UHI: The ideal configuration is LST_urb_all-LST_rur for the entire urbanized area and LST_urb_CT_act-LST_rur for individual census tracts within the urbanized areas

For the equity analysis: Either _CT or CT_act can be used if we are only concerned with spatial variation.
Using CT_act leads to mismatch between census data for the tracts crossing the urban boundary and the remotely sensed data.
Using _CT leads to mismatch between the UHI analysis and the equity analysis.