The Legacy of Redlining on Urban Heat in Virginia Cities

Abstract:

Beginning in the 1930s, a practice known as “redlining” shaped access to mortgage lending and by extension, the housing stock and development patterns that characterized the decades leading up to and following World War II. The effects of this racially biased historical practice remain visible in the landscape of urban heat exposure in the United States. Redlined neighborhoods were considered high risk for mortgage lenders leading to systematic disinvestment. As heat waves become longer in duration and more frequent due to climate change, formerly redlined vulnerable communities will continue to be adversely exposed to extreme urban heat in the summer. This study aims to estimate the average change in land surface temperature (LST) between the summers of 1985 to 2019 using Landsat Analysis Ready Data (ARD) raster datasets for cities in Virginia with Home Owners Loan Corporation (HOLC) security grade housing maps. An LST anomaly (δLST) measure was calculated by subtracting the mean LST for the whole city from the mean LST of each HOLC rated polygon. T-test results suggest that HOLC class D areas in all cities are warmer than non-redlined areas with a p-value less than 0.001 at a 95% confidence interval. The difference in δLST between non-redlined and redlined neighborhoods in each city is decreasing over time as non-redlined areas get warmer, but these comparisons are not statistically significant. Initial findings from an analysis of land cover change over time suggest that redlined areas are seeing the largest change in an increase of impervious surfaces. To precisely understand changes over time within these HOLC polygons, we conduct a longitudinal analysis that compares δLST and impervious surface land cover rasters at a pixel-by-pixel level. Our study reveals how δLST and impervious surface patterns within polygons that reflect historical HOLC lending risk determinations have changed through time.