【作者】 陈学兄；

【导师】 常庆瑞；

【作者基本信息】 西北农林科技大学 ， 土地资源与空间信息技术， 2013， 博士

【摘要】 区域土壤侵蚀与环境研究是土壤侵蚀学科的前沿研究领域之一，主要包括区域土壤侵蚀因子研究、区域土壤侵蚀定量评价和土壤侵蚀及其治理的区域环境效应等三个方面，其中区域土壤侵蚀因子研究是认识土壤侵蚀环境特征和进行土壤侵蚀定量评价的基础，也是土壤侵蚀与环境研究中一个重要的方向。土壤侵蚀是自然因素和人为因素共同综合作用的结果。自然因素是土壤侵蚀发生、发展的潜在条件，主要包括降雨、地形地貌、植被、土壤等，这些自然因素对土壤侵蚀的影响各不相同，但又相互影响、相互制约。本研究以通用土壤流失方程系列模型（USLE、RUSLE、CSLE）为基础，利用RS、GIS、统计学等分析技术，分析提取了影响中国土壤侵蚀的各个自然因子，即降雨侵蚀力因子、土壤可蚀性因子、地形因子（如地形起伏度、地面粗糙度等）、植被覆盖度因子，并对动态因子进行时空动态分析，生成了自然因子系列栅格数据库，编制了中国土壤侵蚀因子图（降雨侵蚀力因子图、土壤可蚀性因子图、地形起伏度图、地面粗糙度图、植被覆盖度图等），在此基础上利用坡度和植被覆盖度（1998—2007年）两个指标对土壤侵蚀的分布做了初步分析，然后将USLE与GIS集成，完成中国水土流失定量评价，编制中国1998年和2007年两个时期的土壤侵蚀强度分级图，根据评价结果图进行土壤侵蚀的时空动态分析，以期为进一步开展区域土壤侵蚀综合治理工作奠定基础。本研究取得的主要成果如下：1.利用1998—2011年的日降雨资料，对中国680多个地面气象站的逐月、月平均、逐年和多年平均降雨侵蚀力进行估算。采用Kriging内插法进行空间插值，采用线性倾向估计、滑动平均、累积距平、变异系数（CV）、趋势系数（r）等方法对降雨侵蚀力的时间变化特征进行分析，并采用相关系数统计检验法对长时间序列的降雨侵蚀力总体变化趋势进行显著性检验，同时对降雨量、降雨侵蚀力的空间分布特征进行比较与分析。中国降雨侵蚀力年内年际变化与降雨量、侵蚀性降雨量年内年际变化趋势一致，年内分布均呈单峰型，集中分布在4～10月；1998—2011年年降雨侵蚀力R值的全距为22249.32MJ·mm/(ha·h·a)，年降雨侵蚀力呈波动下降的趋势，倾向率为-278.29MJ·mm/(ha·h·10a)，且逐年降雨侵蚀力总体变化趋势未通过90%的信度检验水平，其年际间的变化并不显著。1998—2011年间，春、夏、冬三季的降雨侵蚀力均呈下降趋势，而秋季的降雨侵蚀力呈上升趋势；1～12月的趋势系数在-0.213～0.338，各月降雨侵蚀力的变异系数均大于0.1，各季变异程度依次为：夏季＞春季＞秋季＞冬季。降雨量与降雨侵蚀力的空间分布具有从东南向西北梯度递减的特点，南方地区则以高值区为中心向外扩展，呈梯度递减分布。2.依据中国1:100万土壤图，分析土壤理化性质的空间分布特征，利用EPIC模型估算中国土壤可蚀性K值并用张科利的修正公式进行修正，得到中国土壤可蚀性空间分布图。我国土壤可蚀性K值的空间分布变异不大，且具有明显的区域特性，中国土壤可蚀性K值的范围为0.0018—0.089t·ha·h/(ha·MJ·mm)，平均值约为0.0363t·ha·h/(ha·MJ·mm)，中国土壤可蚀性K值主要集中在0.030—0.045t·ha·h/(ha·MJ·mm)之间，其面积占研究区1/2之余。土壤可蚀性高值区主要分布在新疆维吾尔自治区的部分地区、内蒙古高原南部、黄土高原北部、青藏高原北部等地区，而南方地区的K值一般低于均值，只有个别地方稍高。3.选用90m×90m SRTM DEM数据，利用邻域窗口分析法（矩形邻域和圆形邻域）提取中国地形起伏度，对邻域面积与平均地形起伏度进行对数方程拟合，通过统计学检验；运用均值变点分析法计算得出基于90m×90m SRTM DEM数据提取中国地形起伏度的最佳统计单元大小为11×11和R=6两个尺度；完成了中国水土流失地形起伏度分级图的绘制并对地形起伏度特征作了初步分析。中国局部的地形起伏度较大，总体上地形较平缓，以中小起伏度为主，微起伏次之。空间上东西、南北差异明显，大、极大起伏明显集中于西部，而平坦、微小起伏及中起伏明显集中于东部；中大起伏多集中分布于南部，而北部的大部分地区平坦，地形起伏较缓和，以微小起伏为主。均值变点分析法很好地克服了主观因素的影响，是确定最佳统计单元的一种较为理想的方法。4.选用90m×90m SRTM DEM数据，利用三角函数中坡度余弦的倒数对中国地面粗糙度进行了计算，全国地面粗糙度的范围为1—31.4296，平均值约为1.035。东西、南北差异明显，西部天山山脉、青藏高原边缘一带的地面粗糙度较大，明显大于东部地区，而北部大部分地区的地面粗糙度都较小，明显小于南部地区。在盆地、平原地区，地面粗糙度较小，而在天山山脉、横断山脉、秦巴山地等地形较复杂的区域，地面粗糙度较大，这种趋势与地形起伏度特征较相似。5.对11×11窗口的地形起伏度、R=6窗口的地形起伏度和地面粗糙度进行了相关分析，三者之间的相关性都较好，但圆形邻域优于矩形邻域，所以在全国尺度上选择最佳统计单元（R=6）地形起伏度来折算坡度，并在此基础上计算了坡度因子值，最大坡度因子值为20.95，平均坡度因子值为7.80。中国坡度因子的空间分布特征与地形起伏度、地面粗糙度基本一致，东西、南北差异较明显。6.选用1998年4月至2008年7月372景逐旬的SPOT-4/VEGETATION数据（S10），利用MVC法、一元线性回归分析法和差值法分析陕西省和中国近10a来植被的整体变化趋势、年内年际间的变化幅度及其空间分布情况等。陕西省和中国植被覆盖度整体都呈波动上升的趋势，其年际变化趋势大致相同，年内变化呈很强的季节性，春、夏季植被覆盖度呈增加趋势，秋、冬季植被覆盖度呈减少趋势。陕西各地区植被覆盖度变化明显，陕北北部地区植被覆盖度显著增加，特别是在榆林市的东南部和延安市北部地区；桥山、黄龙山林区和秦巴山地林区高度植被覆盖度(＞60%)增加10%～20%。我国植被覆盖空间分布呈现出东北–西南向延伸、东南–西北向更替的规律，东半部地区植被覆盖状况较好；而在半干旱和半湿润区域分界线以西，植被覆盖度较低，尤其在荒漠地带，基本无植被覆盖。7.依据水利部标准（SL190—2007），利用坡度和植被覆盖度两个指标对中国水土流失进行定性评价，编制1998—2007年中国土壤侵蚀强度图。1998—2007年不同土壤侵蚀强度的空间分布与变化趋势大体一致，空间上东西差异明显；土壤侵蚀强度以微度侵蚀为主，轻度侵蚀、中度侵蚀次之。依据USLE模型和土壤侵蚀因子栅格图集，利用ArcGIS软件的栅格计算器计算中国1998年和2007年两个时期的土壤侵蚀模数，得到两个时期的土壤侵蚀强度分级图，对1998年和2007年不同侵蚀类型区土壤侵蚀强度的动态变化进行了分析。1998年和2007年两个时期各侵蚀类型区的最大侵蚀模数和平均侵蚀模数相差都很大，从1998年到2007年土壤侵蚀程度在减弱，2007年北方土石山区的最大侵蚀模数和西北黄土高原区的平均侵蚀模数较1998年有所增加，而其他各侵蚀类型区的年总侵蚀量是减少的。更多还原

【Abstract】 Research on soil erosion and environment at regional scale is one of the frontier researchfield of soil erosion science, which mainly includes regional soil erosion factors research,regional soil erosion quantitative assessment, soil erosion and impacts of soil conservation onregional environment and so on. Research on regional soil erosion factors is the basis forunderstanding the characteristics of eroded soil environment and provides parameters for soilerosion quantitative assessment, and is also an important direction of soil erosion andenvironment research.Soil erosion was caused by both natural factors and human factors. Natural factors arethe potential conditions in the development of soil erosion, mainly including rainfall,landform, vegetation, soil and so on, which affect soil erosion in different ways.This studyanalyzed the natural factors including rainfall, soil erodibility, terrain (relief amplitude andground roughness) and vegetation coverage that affected soil erosion in China using remotesensing(RS), geographic information system(GIS) and statistics based on Universal Soil LossEquation(USLE) model, and dynamically analyzed these dynamic factors. Raster database ofnatural factors was created and maps of China’s soil erosion factors (Rainfall erosivity map,soil erodibility map, relief amplitude map, ground roughness map and vegetation coveragemap) were made. The distribution of soil erosion in China was analyzed using slope factorand vegetation coverage (1998-2007) factor, and then the USLE was combined with GIS toquantificationally evaluate the soil and water loss in China and maps of China’s soil erosionclassification map in1998and2007were made. The temporal and spatial dynamic analysis ofsoil erosion were made on the basis of the evaluation maps, which would provide some basicinformation for the comprehensive control of regional soil erosion.The main achievements of this study are as follows:1. The monthly and yearly rainfall erosivity was calculated based on the daily rainfalldata from680ground weather stations in China. The temporal variation characteristics ofrainfall erosivity was analyzed using spatial interpolation by Kriging method, linear tendencyestimation, moving average, accumulated deviation, coefficient of variation (CV), and trendcoefficient (r). Then the overall trend of rainfall erosivity in a long sequence was tested forsignificance using statistical correlation coefficient test, while the spatial distribution features of rainfall and rainfall erosivity were compared and analyzed. The interannual variation ofrainfall erosivity was consistent with the interannual trend of annual rainfall and erosiverainfall, which was unimodal distribution and mainly distributed between April and October.The range of the R of annual rainfall erosivity between1998and2011was22249.32MJ mm/(ha h a), which was decreasing fluctuatingly, and the tendency rate was-278.29MJ mm/(ha h10a). The overall trend of rainfall erosivity did not pass the significanttest with the level of90%, which meant the interannual variation was not significant. From1998to2011, the rainfall erosivity decreased in spring, summer and winter, while the trendwent up in autumn. The tendency rate was between-0.213and0.338from January toDecember and the coefficient of variation of rainfall erosivity in each month was above0.1.The degree of variation of each season in a descending order was as: summer, spring, autumnand winter. The spatial distribution of rainfall and rainfall erosivity gradiently decreased fromsoutheast to northwest. In southern China, it gradiently decreased from the center where thevalue was high to outside.2. The spatial distribution of the chemical and physical properties of soil was analyzedbased on the soil map of China (1:1,000,000). The K value of China’s soil erodibility wascalculated using EPIC model and revised using Zhang Keli’s revised formula and the map ofChina’s soil erodibility was made. The spatial variation of the K value of China’s soilerodibility was small and had significant regional characteristics. The range of the K value ofChina’s soil erodibility was between0.0018and0.089t ha h/(ha MJ mm) and the mean was0.0363t ha h/(ha MJ mm). The K value focused between0.030and0.045t ha h/(ha MJ mm),and the land area with such K value took over half of the study area. The high value of soilerodibility was mainly distributed in parts of the Xinjiang Uygur Autonomous Region, southof Inner Mongolia Plateau, north of Loess Plateau and north of the Qinghai-Tibet Plateau. Insouthern China, the K value was usually lower than average.3. Based on90m×90m SRTM DEM, the relief amplitude of China was extracted usingthe neighborhood statistics analysis method (rectangular neighborhood and circularneighborhood), the neighborhood area and average relief amplitude were carriedlogarithmically fitting and passed statistical tests; the best statistical unit of90m×90m SRTMDEM was calculated by using the mean change-point analysis method, which was the11×11and R=6. Finally, the map of China’s relief amplitude grade was made and the features ofrelief amplitude were analyzed. The relief amplitude in parts of China was large, the wholerelief amplitude was relatively flat, which was mainly medium-sized, followed by micro-sized.The relief amplitude had significant differences in space on the east-west and north-southdirection, the big and great relief amplitudes were clearly concentrated in the west, while the flat, slight and medium relief amplitudes were mainly distributed in the east, and the mediumand big relief amplitude in the south area, and the relief amplitude of the northern part wassmall. The mean change-point analysis method could overcome the subjective factors, whichwas an ideal method to determine the best statistical unit.4. Based on90m×90m SRTM DEM, China’s ground roughness was calculated by theinverse of the cosine of slope. The national ground roughness was between1and31.4296,and the mean value was1.035. Ground roughness had significant differences in the east-westand north-south direction. In western China, along the edge of the Qinghai-Tibet Plateau andTianshan Mountains, ground roughness was great and significantly greater than that of theeastern region, and the ground roughness in north region was mainly small and significantlysmaller than that of the southern region. In the basins and plains region, the ground roughnesswas small, and in the complex terrain such as Tianshan Mountains, Hengduan Mountains andthe Qinling Mountain area, ground roughness was great, this trend was similar to thecharacteristics of the relief amplitude.5. The correlation analysis was made between the relief amplitude (11×11grid unit ofrectangular neighborhood and R=6grid unit of circular neighborhood) and ground roughness.The results showed good correlation, and the circular neighborhood was better thanrectangular neighborhood, so in the national scale, slope and slope factor value was calculatedby the best statistical unit（R=6）of the relief amplitude, the maximum value was20.95, andthe mean value was7.80. The spatial distribution characteristics of slope, relief amplitude andground roughness were basically consistent, and there were significant differences in theeast-west and north-south direction.6. This study applied maximum value composites (MVC), one-dimensional linearregression and differential methods to analyze spatial distribution and inter-annual and annualchanging patterns of vegetation coverage in the whole China and Shaanxi Province based on372images of SPOT-4/VEGETATION (S10) data recorded from April1998to July2008.The vegetation coverage both in the whole China and Shaanxi Province generally raised withfluctuations and its inter-annual variation was basically the same. However, there was obviousseasonal variation within a year that vegetation coverage increased from spring to summerand then decreased from autumn to winter. Vegetation coverage changed significantly invarious regions of Shaanxi. It increased dramatically in Northern Shaanxi, especially in thesoutheast of Yulin city and in the north of Yan’an city. The high vegetation coverage(>60%) inQiaoshan, forest area of Huanglong Mountain and Qinba Mountain increased by10%~20%during the past10years. Vegetation coverage in whole China spatially extended from thenortheast to the southwest and decreased from the southeast to the northwest. Vegetation coverage was high in eastern China and low in the west of the boundary between the semiaridarea and sub-humid area especially and the lowest (about zero value) in the desert zone.7. According to the criteria for soil erosion made by the Ministry of Water Resources（SL190—2007）, the datasets of slope and vegetation coverage were used to conduct qualitativeassessment of soil and water erosion in China and make soil erosion intensity maps(1998-2007). The results showed the spatial distribution patterns and changing trends of soilerosion intensity were generally consistent from1998to2007. Spatially, there weresignificantly differences between eastern China and western China. Among soil erosionintensity types, micro erosion type occupied first place, slight erosion type came second, andmiddle erosion type was the third. Furthermore, this study used USLE model and atlas of soilerosion factors, by means of the raster calculation module of ArcGIS (version9.3), tocalculate soil erosion modulus in1998and2007to obtain the soil erosion intensity gradingmaps of the two years so as to analyze the dynamic changes of soil erosion intensity indifferent erosion areas. There were significantly differences in terms of the largest erosionmodulus and average erosion modulus in each erosion areas. Comparing with the relatedresults in1998, soil erosion intensity generally weakened and the total erosion amountdecreased in2007, except that the largest erosion modulus of mountainous regions in northernChina and average erosion modulus of loess plateau in northwestern China increased.更多还原