ABSTRACT The spatial attenuation effect, a core expression of Tobler's First Law of Geography, strongly shapes the distribution of precipitation. Traditional models use kernel functions to capture spatial association but often ignore terrain's three‐dimensional influence, limiting accuracy in complex areas. This study proposes an improved kernel function incorporating the ACE index between point pairs to better simulate spatial diffusion. Applied to autumn precipitation in Zhejiang Province, China, the model includes slope, terrain undulation, coastal distance, and the principal wind direction coefficient (PWEI), with PWEI increasing R 2 by 13.8%. Random validation confirms superior performance at optimal bandwidths. Standardized regression coefficients show coastal precipitation is driven by land‐sea distribution, while inland areas are mainly influenced by elevation, reflecting clear spatial heterogeneity. Terrain sub‐scenario tests further demonstrate the model's strong transferability. Embedding surface morphology as a continuous factor enhances simulation accuracy and offers a novel framework for spatial modeling in geography.
Zhu et al. (Sun,) studied this question.