Urban Heat Exposure Interactions with Environmental Drivers at the Neighborhood Scale: The Case of Bağcılar and Esenler Districts, Istanbul

Urban heat exposure is an environmental challenge in densely populated metropolitan regions, especially at fine spatial scales where intra-urban thermal disparities arise. This research examines persistent neighborhood-level thermal exposure in two adjacent high-density districts of Istanbul, Bağcılar and Esenler, utilizing Landsat-derived land surface temperature (LST) data from 2016, 2022, and 2024. Environmental drivers of surface temperature were assessed using the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Built-up Index (NDBI), derived consistently for each year. Neighbourhood-level mean LST, NDVI, and NDBI values were calculated using zonal statistics, and multiyear mean values were used to assess the consistency of spatial thermal patterns. Results show that several neighbourhoods in both districts consistently exhibited elevated LST across all three years, indicating spatially persistent thermal hotspots. District-wide mean LST showed interannual variability without a monotonic trend, suggesting spatial heat patterns are more stable than temporal changes. Multiple linear regression identified NDVI as the only significant predictor of LST (p < 0.05), with higher explanatory power in Esenler (R² = 0.81) than in Bağcılar (R² = 0.46). Regression results indicate that a 0.01 increase in NDVI is associated with an approximately 0.45 °C decrease in LST in Bağcılar and ~0.24 °C in Esenler, while NDBI was not significant. These findings highlight vegetation as the strongest predictor of neighbourhood-scale LST among the variables considered.