Urban landscape heterogeneity disaggregates the legacy of redlining on land surface temperature

• Integrating redlining data with micro-scale development captures urban landscape heterogeneity. • Urban landscape heterogeneity reveals disaggregated patterns of temperature inequality. • Tree canopy cooling varies by redlining legacy, development level, and urban context. • Non-linear canopy cooling thresholds highlight the need for strategic tree planting. • Tree planting initiatives should consider micro-landscape diversity for climate equity. The lingering effects of redlining are linked to contemporary heat inequities observed across US cities. Residential security maps created by the Home Owners’ Loan Corporation (HOLC) have been widely used to analyze neighborhood-level disparities in land surface temperatures. However, the use of aggregated spatial units often fails to capture internal landscape heterogeneity and the heat vulnerabilities associated with redlining. In this study, we introduced urban landscape heterogeneity by incorporating granular development levels captured at different resolutions within HOLC-graded neighborhoods. This approach combined Landsat-based National Land Cover Database (NLCD) data, Sentinel-based WorldCover land cover data, and HOLC map layers. We examined the role of urban landscape heterogeneity in revealing additional patterns of heat inequities beyond those explained by redlining-based macro spatial units, using grouped boxplots and mixed-effects models across three major cities in the Northeastern US: Boston, Massachusetts; New York, New York; and Philadelphia, Pennsylvania. By accounting for urban landscape heterogeneity, our findings revealed that: (1) the well-documented trend of higher land surface temperatures in lower HOLC grades becomes systematically fragmented, (2) statistical models show improved performance in estimating land surface temperature, and (3) the cooling effect of tree canopy exhibits a varying, non-linear threshold pattern. These results highlight the need to consider micro-scale landscape dimensions to better understand the persistent, unequal distribution of temperatures associated with redlining. Municipal and community-led tree planting initiatives should consider comprehensive landscape characteristics to develop spatially targeted heat mitigation strategies and promote equitable climate outcomes.