Quantifying Causal Drivers of Urban Heat through Geospatial Analytics: Evidence from Three East African Cities

Rapid urbanization in East Africa profoundly transforms landscapes, yet a critical understanding of the causal mechanisms behind associated land surface warming remains limited. This study quantifies the spatiotemporal dynamics and causal drivers of urban expansion and its thermal impacts in three East African cities; Wakiso-Kampala (Uganda), Nairobi (Kenya), and Bujumbura (Burundi) from 1995 to 2024. Using Landsat imagery, we derived land use/land cover (LULC), Land Surface Temperature (LST), the Normalized Difference Vegetation Index (NDVI), and the Normalized Difference Built-up Index (NDBI). A Land Cover Thermal Impact (LCTI) metric was introduced to quantify per-unit-area warming contributions, while advanced computational methods Convergent Cross Mapping (CCM, with causal strength measured by ρ) and its spatial extension, Geographical CCM (GCCM) were applied to distinguish causal links from mere correlation, moving beyond traditional statistical approaches. Results show built-up areas tripled in Wakiso-Kampala and Nairobi and quadrupled in Bujumbura, displacing 35-80% of natural vegetation and croplands. This expansion drove a mean LST increase of 5.1°C, 3.3°C, and 2.7°C, respectively. The LCTI revealed that water bodies provided the most efficient per-unit-area cooling in Wakiso-Kampala (LCTI = -1.49 °C km⁻²) and Bujumbura, while forest gains and bare-land conversion were the primary cooling mechanisms in Nairobi. Crucially, causal analysis revealed an asymmetric relationship: NDBI consistently acted as a driver of LST (ρ up to 0.77), while NDVI exerted a cooling causal influence (ρ down to -0.57). These findings confirm that urban expansion and vegetation loss are fundamental, causal drivers of rising surface temperatures. Our geospatial analytics framework demonstrates how data-driven causal inference can inform climate adaptation strategies in rapidly urbanizing regions. The spatial mapping of causal relationships provides actionable insights for urban planners to prioritize locations for blue-green infrastructure expansion, optimize nature-based cooling interventions, and develop targeted heat mitigation strategies that address the specific thermal dynamics of East African cities.

Keywords: Urban Heat Island, Geospatial Analytics, Causal Analysis, Nature-based Solutions, Climate Adaptation, East Africa