Research On Coupling Characteristics Between Soil Erosion And Topographic Fractal Feature Of Taihu Lake Basin

The paper was based on Taihu Lake Basin and the basin was divided into hilly area and plain area. Stream networks and contour features were extracted by ArcGIS,the fractal dimension of water and terrain were calculated,and the topographic fractal feature of Taihu Lake Basin was revealed from different angles.The Universal Soil Loss Equation was re-examinated based on Taihu Lake Basin soil erosion characteristics and the factors that influence soil erosion modulus were defined.The artificial neural network was established by expressing the relationship between topographic fractal feature(the fractal dimension of water and terrain),soil erosion modulus and vegetation coverage.Seven typical rivers in plain area of Taihu Lake Basin were selected to analyze the relationship between per meter of river plate natural sediment,per meter of river plate soil erosion and the fractal dimension of water. The coupling relationship between characteristics of soil erosion and the topographic fractal feature was expounded,and provide new ideas about ecology security of Taihu Lake Basin and other basins.The research results show that:(1) Situation of soil erosion mainly in mild and moderate in hilly area affect by rainfall and vegetation cover.In some districts and counties,as a result of long-term unscientific digging,quarrying,growing white tea and mountain vegetables,causing serious soil erosion.Compared with hilly area,soil erosion in plain area is lighter.But well-developed river transport induces bank collapse and great soil erosion in the frequency impact of ship waves.In addition,precipitation,irrigation and other agricultural activities in the plain area lead to a large number of organic or inorganic pollutants into the water body,such as soil,nitrogen,phosphorus and pesticides,causing heavy non-point source pollution,threatening the water quality of the lake.(2) The total fractal dimension of plain area is greater than hilly area,but the fractal dimensions of sub-water systems in plain area are smaller than the sub-basins in hilly area,and the total fractal dimension is greater than part of the basin can be seen from the calculation result of water and terrain fractal dimension in hilly and plain area.(3) The soil erosion BP neural network based on sub-basins' training samples in hilly area has a better convergence,and the test samples are simulated well in linear correlation with measured values.It illustrated that river fractal dimension,terrain fractal dimension and vegetation cover exist correlation with soil erosion modulus.However,serious interference of human activities in hilly area exacerbate exogenous and uncertainties of the influencing factors of soil erosion,and bring difficulties to express the coupling relationship between soil erosion and topographic characteristics by a certain amount.(4) Seven typical rivers in plain area are selected and the fitting with per meter of river plate natural sediment and water fractal dimension is performed under linear function and logarithmic function,R-square are 0.8619 and 0.8607 respectively,indicating a significant correlation between per meter of river plate natural sediment and water fractal dimension.The fitting with per meter of river plate soil erosion and water fractal dimension is performed under linear function and logarithmic function,R-square are 0.7587 and 0.7572 respectively,indicating a significant correlation between per meter of river plate soil erosion and water fractal dimension.Compared with the results from the fitting,the coupling relation between per meter of river plate natural sediment and water fractal dimension is more significant,and the coupling relation between per meter of river plate soil erosion and water fractal dimension is more complicated.