The relationship between landscape pattern and water and soil erosion in hilly red soil region in South China
To prevent and control water and soil erosion in hilly red soil region of South China, landscape pattern indexes dominantly influencing erosion area were extracted by analyzing DEM datasets of Anxi County, then a nonlinear mapping relation between land-scape pattern indexes and erosion area was established by back propagation(BP) model. Finally, new erosion area was simulated by the model with adjusted landscape pattern indexes. The results showed that soil and water erosion mainly occurred at tea plantation and bare field in Xixi River Basin. Indexes including class area ( CA) , number of patches, area-weighted mean shape index and patch fractal dimension were positively correlated with erosion area while landscape connectivity was negatively correlated. The abso-lute value of relative error of mapping relation between erosion area and landscape pattern was less than 5%, indicating a high preci-sion of the BP model. The stimulated erosion area was dropped by 14.14% with adjusted LSI of tea plantation and CA of bare field. To summarize, BP model based on landscape pattern indexes was applicable to indicate and control erosion in hilly red soil region of South China.
landscape pattern indexwater and soil erosionback propagation neural networkhilly red soil region in South China
国家科技期刊平台
NSTL
万方数据
维普
