Study On The Characteristics Of Soil Erosion Underlying Surface In Watershed And Its Effect On With Soil And Water Loss On Loess Plateau

Geomorphology, land use/land cover and landscape pattern are the important underlying surface conditions of watershed erosion environment and the dominant factors of regional soil and water loss. It is the important things for the soil erosion research on Loess Plateau to quantify the characteristics of the underlying surface in watershed scientifically and rationally, to establish the index system of underlying surface suited to the Loess Plateau, and to study the coupling relationship between the quantitative parameters and sediment yield. According to the inter-disciplinary theories such as hydrology, sediment transport mechanics, soil and water conservation, ecology, fractal theory, geomorphology and RS/GIS technology etc, by using the integration of theoretical analysis and numerical simulation methods, together with taking typical watersheds on Loess Plateau as study objects, with the data of rainfall process and watershed underlying surface, the characteristics of soil erosion underlying surface in watershed and the relationship with soil and water loss on Loess Plateau was analyzed in the paper. The main conclusions are as follows:(1)Dalihe River Geospatial Database was constructed, and the characteristics of the underlying surface in which was studied. Based on geographic information system (GIS) and remote sensing (RS), combined with RS data extraction technology and the spatial information data overlay analysis, the basic character of soil erosion in Dalihe River, the fractal characters and topographic characters of regional soil erosion spatial distribution, and the topographic characters of land use/land cover spatial distribution was researched, which revealed the initial characteristics of the underlying surface in Dalihe River.(2) The theory and method on describing the topographic FBM characteristics of watershed by using Fractional Brownian Motion theory was established, and the calculation software system is developed. According to the subjective arbitrariness in the fractional calculation process and the lack of effective testing means for fractal characteristics, based on the Fractional Brownian Motion theory, combined with watershed digital elevation model, the topographic feature FBM quantitative model was suggested by the fixed size moving window analysis. By studying the increments'probability distribution, statistical self-similarity, and the fractal character, the FBM characteristics of watershed topographic feature was proved. The realization method of topographic feature FBM quantitative model was discussed, and the model was established by the correlative calculation program compiled and debugged.(3)The spatial variability and the scale effect on watershed topographic FBM characteristics in Dalihe River were studied. Based on the topographic feature FBM quantitative model, the topographic feature FBM fractal dimensions of sub-watersheds on upstream, midstream and downstream and sub-watersheds with different area were calculated. The regression and multiple regression methods were used to establish the equation of FBM fractal dimensions variation along the river and the equation of correlation between the FBM fractal dimensions and the watershed area, which reveal the spatial variability and the scale effect on watershed topographic. And the relationship between the FBM fractal dimensions and the traditional topographic feature quantitative model was discussed as well.(4)The quantitative model for the watershed characteristics of vegetation cover and vegetation pattern is established, and the corresponding calculation program is also developed in this paper. Using the TM/ETM remote sensing image data of many periods as the basic information sources, based on the geometric correction, radiometric correction, image subset and radiometric enhancement, the space distributing features of the NDVI vegetation index in the Dalihe watershed are extracted. Through the data mining models, the quantitative indexes of watershed vegetation cover characteristics based on the mean NDVI vegetation index, and the quantitative models of space distributing features of the vegetation based on the vegetation pattern FBM fractal dimension, are proposed, the responding calculation program is developed under the GIS/RS platform, and then the evaluating indexes for the soil and water conservation function of regional vegetation, which can reflect the soil and water conservation effect of the vegetation and is easy to be obtained by remote sensing technology, is primarily established.(5)To describing the temporal and spatial varying characteristics of the vegetation cover of Dalihe watershed, the TM/ETM remote sensing image data of many years was used to extract the vegetation cover information of many periods in this watershed. Through the calculation of mean NDVI vegetation index, the vegetation pattern FBM fractal dimension, and the NDVI vegetation indexes of different grades and different elevation in the sub-watershed of the upper middle and lower reaches of Dalihe watershed, the temporal and spatial varying regulations and its vertical distribution features of the land vegetation cover in Dalihe watershed are quantitatively analyzed, and the dynamic varying feature and regulations of the vegetation conditions in Dalihe watershed of the d 20 years passed are revealed in this paper.(6)The coupling correlation, between the characteristics of the underlying surface and the sediment yield of the watershed, is discussed, and the runoff and sediment data of 38 rainfalls in the Dalihe Watershed is collected and analyzed in this paper. Using the runoff erosion and sediment power of the single rainstorm in the watershed as the dynamics of erosion and sediment transportation in the process of water erosion, the BP Neural Network Model was used to construct the dynamic coupling relationship between the underlying surface characteristics of geomorphology, land cover and landscape patterns and the rainfall erosion and sediment. By the comparison and selection of the inputting parameter for different underlying surfaces, the rationality and feasibility, using the topographic feature FBM fractal dimension, mean NDVI vegetation index, and the vegetation pattern FBM fractal dimension as the quantitative parameters of the underlying surface characteristics for the rainfall erosion and sediment forecasting model, are expressed in this paper.