【作者】 沈中原；

【导师】 李占斌；

【作者基本信息】 西安理工大学 ， 水文学及水资源， 2009， 博士

【摘要】 地貌形态、土地覆盖、景观格局是流域侵蚀环境的重要下垫面条件和区域水土流失的主导因素。如何科学合理的对流域下垫面特征进行量化,建立适合于描述黄土高原地区侵蚀环境的下垫面指标体系,研究各种量化参数与流域侵蚀产沙的作用关系与耦合机制,是黄土高原流域水土流失规律研究的重要内容。本研究以黄土高原典型流域大理河为研究对象,采用地貌学、景观生态学、水文学、土壤侵蚀学、分形理论、数理统计学及RS／GIS技术等多学科交叉,对流域土壤侵蚀下垫面特征及与水土流失的关系进行了研究。主要结论如下：(1)构建了大理河流域地理空间数据库,阐明了流域土壤侵蚀空间分布的下垫面特征。以RS／GIS为平台,结合遥感数据成图与空间信息叠加分析,对大理河流域土壤侵蚀的基本特征、土壤侵蚀空间分布的分形特征、地形特征以及土地利用／土地覆被的地形分异性进行了研究,初步揭示了流域土壤侵蚀空间分布的下垫面特征及其复杂性。(2)建立了流域地貌特征分形布朗运动量化模型,开发了相关的计算程序。针对以往地貌分形计算过程主观随意性较大和分形特征缺乏有效检验手段的问题,根据分形布朗运动理论,结合流域DEM特点,提出了基于移动窗口法的流域地貌分形布朗运动量化模型。阐明了地貌特征FBM分形维数对流域地貌表面形态的表征意义,并对流域地貌形态的分形布朗运动特征进行了检验,探讨了流域地貌特征分形布朗运动量化的实现方法,编写并调试了相关计算程序。(3)揭示了大理河流域地貌分形布朗运动特征的空间变异性,探讨了地貌特征的尺度转换的途径,阐明了地貌特征FBM分形维数与传统地貌量化参数的关系。以流域地貌特征分形布朗运动量化模型为基础,对大理河流域上中下游各子流域地貌分形布朗运动特征进行计算,运用回归与多元回归的方法,建立了地貌特征FBM分形维数的沿程变化方程及其与流域面积的相关关系方程,揭示了流域地貌形态的空间变异特征及其变化规律。对地貌特征FBM分形维数与传统地貌量化参数的关系进行了探讨,阐明了地貌特征FBM分形维数对流域地貌特征的整体性、综合性表征。(4)建立了流域植被覆盖及植被格局特征量化模型,并用以评价区域植被的水土保持效应。以多期多源的TM/ETM影像为基础信息源,在完成几何校正、辐射校正、切割裁剪和图像增强等处理过程的基础上,提取了大理河流域各期NDVI植被指数的空间分布特征。通过数据挖掘建模提出了基于均值化NDVI植被指数的流域植被覆盖特征量化指标和基于植被格局FBM分形维数的植被空间分布特征量化模型,并在GIS/RS平台下进行相应的计算程序开发,初步建立了反映植被水土保持作用、易于遥感提取、适合区域研究的植被水土保持功能评价指标。(5)研究了大理河流域植被覆盖的时空变化特征。利用多年的TM/ETM影像资料,提取了流域多期植被覆盖信息,通过对大理河流域上中下游各子流域均值化NDVI植被指数、植被格局FBM分形维数以及不同高程上不同等级NDVI植被指数的计算,定量的分析了大理河流域地表植被覆盖随时间和空间的变化规律及其垂直分布特征,揭示20年来大理河流域植被状况的变化特征和动态变化规律。(6)揭示了流域下垫面特征与流域侵蚀产沙的耦合关系。收集并分析处理大理河流域38场降雨的径流泥沙资料,以流域次降雨径流侵蚀功率为流域水蚀过程侵蚀输沙动力,构建了地貌形态、植被覆盖、植被格局等流域下垫面特征与降雨侵蚀产沙之间的动态耦合关系。通过对不同下垫面输入参数的比选,阐明了流域地貌特征FBM分形维数、均值化NDVI植被指数和植被格局FBM分形维数作为流域降雨侵蚀产沙预报模型下垫面特征量化参数的合理性与可行性。更多还原

【Abstract】 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.更多还原