| Summary: | In the context of global warming, urbanization has exacerbated the rise in land surface temperature(LST). However, existing studies are still insufficient on the dynamics of diurnal LST drivers and their marginal effect at multiple spatial and temporal scales in urban agglomeration. This study integrates geographically weighted regression (GWR) and gradient boosting machine (GBM) with accumulated local effect (ALE) plots to quantify spatial heterogeneity and marginal effect of diurnal LST drivers in the Yangtze River Delta urban agglomeration. Albedo shows a strong and stable effect on LST, with daytime warming and nighttime cooling patterns. Enhanced vegetation index (EVI) and evapotranspiration (ET) are negatively associated with LST in all seasons except winter, and the cooling effect of EVI is approximately four times that of ET. GWR results reveal that albedo has the highest positive daytime coefficient (∼85) and negative nighttime coefficient (∼63), while EVI consistently shows cooling effects (∼21 in daytime, ∼12 at night). ALE plots indicate that increasing albedo from 0.075 to 0.15 leads to + 1.30 °C daytime warming and −1.35 °C nighttime cooling, while during the day, an increase in EVI can cool up to about 2.5 °C; at night, when EVI increases from 0.20 to 0.35, it can cool about 0.5 °C. These findings provide new insights into the diurnal and marginal effect of urban surface parameters, offering valuable guidance for precision urban climate adaptation strategies in rapidly urbanizing regions.
|
|---|
