Spatial And Temporal Evolution Simulation Of Anti-erodibility Of Underlying Surface Of Watershed

Soil erosion in Loess Plateau has been the main problem of watershed management in Yellow River; in recent decades, the Soil and Water Conservation work in Loess Plateau has made remarkable achievements, however, the watershed soil erosion problems have not been fundamentally improved, and the further Research and Management of Soil and Water Conservation work still need to be processed.Loess Plateau has features like varied terrain, frequent raining, intensive rainfall, harsh natural environment, combined with unreasonable land use in recent years which causes serious soil erosion, it has become one of the most serious soil erosive regions. Therefore, studying on the Watershed Anti-erodibility of Underlying Surface, modeling and predicting the future changing trend of it, and exploring the effect mechanism of spatial and temporal evolution of the soil erosion, has great significance on the implementing effective strategies and measures of Soil and Water Conservation in Loess Plateau to develop and of great importance.On the basis of utilizing the GIS and RS technology, combining with theories of Watershed Anti-erodibility of Underlying Surface, cellular automata and neural network, after applying topographic data, remote sensing image data of two periods in 1990 and 2004 in Watershed Chabagou of the Loess Plateau and calculating the surface resists force, this paper studies and realizes prediction of the Watershed Anti-erodibility of Underlying Surface in this region in 2018.The main research content and conclusions are as follows:(1) On the basis of studying the soil erosion models domestic and overseas, as well as applying the cellular automata and neural network in the field of geography, this paper processes a spatial and temporal evolution simulation on the selected cellular automaton model based on neural network theory to analyze the Watershed Anti-erodibility of Underlying Surface.(2) After analyzing and studying the impacting factors of development and variation of the Watershed Anti-erodibility of Underlying Surface, the basic principles of cellular automata model and neural network technology, this paper clarifies the configuration of the model, and uses learning vector quantization neural network instead of CA in the transformation rules to simulate and predict changes in land surface resist force.(3) By applying and combining the C #. NET and ArcGIS Engine methods, and integrating GIS, CA models and LVQ neural network, this paper develops and implements a spatial and temporal evolution simulation system of the Watershed Anti-erodibility of Underlying Surface.(4) By using Chabagou watershed as a study area, applying its data on Watershed Anti-erodibility of Underlying Surface in 1990 and 2004, this paper validates the LVQ neural network, and predicts the data of the Watershed Anti-erodibility of Underlying Surface in 2018 based on the validated system. Predicted results show that from 1990 to 2018, the Watershed Anti-erodibility of Underlying Surface in Chabagou watershed showing a gradual increasing trend; this result also indicates that the current soil and water conservation plan being implemented is scientific and reasonable, and enhancing the intensity of further soil erosion management will accelerate the improvement of the soil erosion situation.In summary, the application of neural network-based cellular automata model to the dynamic changes simulation of Watershed Anti-erodibility of Underlying Surface is feasible, and simulation results will provide a scientific basis for soil and water conservation, as well as ecological environment management.