关键词: 水源涵养, 景观格局指数, InVEST模型, 地理探测器, 景观异质

Abstract: Investigating landscape pattern influences on water retention functions is critical for constructing regional ecological security. We analyzed the spatiotemporal variations of water retention in Huaihua City from 1985 to 2020 using the InVEST model. The evolution of landscape patterns was assessed with selected indices, and the relationship between water retention and landscape patterns was explored via the Geodetector method. The results showed that the multi-year average water retention depth in Huaihua City was 78.52 mm, with a total water retention volume of 2.166 billion m³, demonstrating a pattern of initial increase, subsequent decline, and gradual recovery. At the landscape level, temporal changes in indices indicated increased concentration, enhanced connectivity, and reduced fragmentation. At the class level, indices exhibited the characteristics of “stable forest land, decreased cropland, and urban expansion”. Geodetector analysis demonstrated that the Shannon Diversity Index (SHDI) had the highest explanatory power (q=0.2344), followed by the Landscape Shape Index (LSI, q=0.2109) and the Largest Patch Index (LPI, q=0.2021), indicating that greater landscape diversity, larger patches, and reduced fragmentation enhance water retention capacity. Interaction detection revealed that the strongest synergistic effect between the Patch Cohesion (COHESION) and SHDI (q=0.4398), followed by COHESION and the Contagion Index (CONTAG, q=0.4218), indicating that higher landscape heterogeneity and patch aggregation amplify the spatial variation of water retention. Our results provide effective technical support for the construction of landscape patterns, economic development, and the maintenance of regional ecological security in Huaihua City.

Key words: water retention, landscape pattern index, InVEST model, Geodetector, landscape heterogeneity