【作者】 王尧；

【导师】 蔡运龙；

【作者基本信息】 北京大学 ， 自然地理学， 2011， 博士

【摘要】 在全球环境变化的人文因素计划(IHDP)和国际地圈生物圈计划(IGBP)提出的“土地变化科学(Land Change Science, LCS)框架下,土地退化和其影响因素研究是重要问题之一。中国西南喀斯特地区生态环境脆弱,人口压力和不合理的土地利用方式使得当地以土壤侵蚀为特征的土地退化问题严峻。因此对贵州喀斯特区域土壤侵蚀现状和机理进行研究,确定影响土壤侵蚀的主控因子,并建立流域土壤水蚀预报模型,可以为区域综合治理等宏观政策提供帮助,对当地防止土壤退化和减少洪涝灾害具有重要的现实意义。本文选取贵州省境内乌江流域为研究对象,运用遥感、GIS技术、小波分析方法、神经网络技术、RUSLE模型和多元统计分析方法,分析了乌江流域土壤侵蚀的尺度效应和土壤侵蚀主控因子的影响机制,模拟预测了乌江流域的土壤侵蚀空间格局,以期为喀斯特地区的水土流失研究和治理提供可靠依据。1.土壤侵蚀的尺度效应研究(1)时间尺度:利用1980—2000年乌江流域的年径流模数、降雨和输沙模数数据,通过小波分析方法,分析三者变化的周期特性。结果显示,降雨和径流模数序列具有很好的同步性(主周期为16年、次周期为4年);输沙模数的主周期为11年,次周期为5年。在4年和16年这两个尺度上,分别绘制径流、输沙和降雨3个序列的小波变换系数图,发现在这两个尺度上降雨、径流和输沙曲线变化均不同步;这些不同步与不同时期人类对乌江流域的开发特征有关。(2)空间尺度:将乌江流域内各水文站点控制区的产沙模数与面积进行回归分析,表明乌江流域在产沙模数与流域面积的关系中存在尺度效应,经对数转换后流域输沙模数与面积之间呈二次函数回归关系,输沙模数随流域面积变化的趋势表现为先增大后减小。通过设定汇流面积阈值(150 km2),在乌江流域内生成189个子流域。将这些子流域作为样本分析5个土壤侵蚀影响因子(降雨侵蚀力因子R、土壤可蚀性因子K、坡长与坡度因子LS、植被覆盖与管理因子C和水土保持措施因子P)与流域面积的回归关系。分析表明侵蚀产沙影响因子中,坡度坡向因子随流域面积先增大后减小,是决定输沙模数与流域面积关系的主要因素。选择对各个子流域的拟合误差比较小的1000 km2作为标准面积,使用普通Kriging插值法制作出整个研究区乌江流域的输沙模数图,以消除尺度效应的影响。2.土壤侵蚀主控因子影响机制分析(1)降雨:利用1951—2009年降雨资料,计算乌江流域各气象台站降雨侵蚀力,发现乌江流域多年平均降雨侵蚀力变化范围在1751～4349 MJ?mm/(hm2?h?a)之间,平均为2788 MJ?mm/(hm2?h?a),并呈现出南高北低、东高西低的空间分布格局。季节分布上,降雨侵蚀力主要集中在夏季(5—8月),这4个月的降雨侵蚀力占年均降雨侵蚀力的90.42%。1980—2000年,各气象台站降雨侵蚀力倾向率均为正,表明整个乌江流域降雨侵蚀力呈增加趋势。使用“区域重心”概念,计算不同年份降雨侵蚀力重心的变迁及其与坡度等级重心、土地利用重心、区域几何中心和水文站点的空间位置关系,分析了降雨侵蚀力空间分布对土壤流失的影响。(2)地形:乌江流域的侵蚀类型为水力侵蚀,根据强度分为微度、轻度、中度、强度、极强五个等级,以微度和轻度侵蚀为主。在2000年土壤侵蚀分布图和30米分辨率的DEM数据叠加分析的基础上,选取海拔、坡度、地表起伏度和粗糙度四个要素对乌江流域土壤侵蚀进行地貌特征分析:从海拔角度看,微度侵蚀和轻度侵蚀呈现明显的单峰现象,主要集中在1100～1300 m之间的黔中石质丘陵盆地区,中度及以上强度侵蚀面积集中分布在500～900 m和1300～1700 m高程左右;中度及中度以下土壤侵蚀空间分布的坡度特征为单调下降趋势,强度和极强度土壤侵蚀的面积比重随着坡度的增加呈现先增大后减小的趋势,15°左右存在一个侵蚀临界坡度;乌江流域的土壤侵蚀随地形起伏度的增加表现为先增加后减小的趋势,存在7～16 m临界起伏度;随地表粗糙度的增加均呈减小趋势,土壤侵蚀强度的变化对地形起伏度和地表粗糙度的变化均不敏感。(3)土地利用:通过土壤侵蚀强度与土地利用格局之间耦合关系分析可以得出:不同土地利用类型下的土壤侵蚀程度大小为裸岩石砾地<其他林地<建设用地<有林地<水域<水田<低覆盖度草地<灌木林地<高覆盖度草地<疏林地<中覆盖度草地<旱地。旱地、中覆盖度草地和疏林地是流域发生土壤侵蚀的主要土地利用类型。从不同侵蚀等级的发生区域看,各强度类型侵蚀都集中分布在植被覆盖度50%～60%的地区,侵蚀存在50%～60%的植被覆盖度临界值。喀斯特地区土壤侵蚀发生及强弱受限于土层厚度,与非喀斯特地区随着植被盖度降低,侵蚀强度逐渐增大的规律不同。3.土壤侵蚀模拟与预测基于GIS平台,应用RUSLE模型计算乌江流域20世纪80年代和90年代的年均土壤侵蚀模数,计算结果和以往土壤侵蚀调查估计的结果比较吻合。但由于RUSLE模型不包括重力侵蚀,因此仍与实测输沙模数有出入。潜在土壤侵蚀模数方面,90年代比80年代为高,潜在土壤侵蚀呈增加趋势,其中三岔河流域和马蹄河/印江河流域年均潜在土壤侵蚀模数最高。三种主要土地覆被类型中,林地的土壤保持量最大,耕地次之,草地最少,这与非喀斯特地区在水土保持效果上通常林地>草地>旱地的结论有所不同。构建BP神经网络,基于80年代和90年代土壤侵蚀模数及各影响因子数据,预测乌江流域2001—2010年土壤侵蚀模数。结果显示,21世纪前10年,土壤侵蚀模数大幅降低,流域年均土壤侵蚀模数由90年代的2313 t/(km~2·a)降低为101 t/(km~2·a),年均土壤侵蚀量由115.18×10~6 t/a下降为5.03×10~6 t/a。三岔河流域的水土流失得到了控制,黔西、金沙、息烽、修文、贵阳、平坝、思南、石阡、沿河和松桃等县市应是十二五期间的水土流失重点治理对象。更多还原

【Abstract】 Land degradation and its factors are among the most important research themes in Land Change Sciences. Karst areas in southwest China, with a large population and local non-sustainable land use, have the most serious land degradation in China and even the whole world. Therefore research on soil erosion in Karst areas in China, analyzing its factors, using model to simulate the soil erosion are quite necessary for regional management. In this article, we adopt methods like RS, GIS, wavelet analysis, neural network, RUSLE model and multivariate analysis to analyze soil erosion in Wujiang River Basin (WRB) in Guizhou Province. This article is constituted by three parts:1. Scale effect(1) Time scale. Using wavelet analysis method, we calculated the periodicity of annual runoff modulus, precipitation and sediment transport modulus based on their monitoring data from 1980 to 2000. The results show precipitation and runoff modulus data sequences are quite synchronous, which has a primary period of 16 years and a secondary period of 4 years. Sediment transport modulus is a little different, which has a primary period of 11 years and a secondary period of 5 years. We draw the wavelet transform coefficient figures of these three factors using 16-years-peroid and 4-years-period separately. The figures show asynchronization in both two time scales. This is related to different river basin development in different time.(2) Space scale. Through constructing relationships between sediment yield and drainage area of hydrological stations in WRB, it is found that there is a quadratic function relationship. 189 sub-river basins are generated from the whole Wujiang River Basin by setting the threshold value of convergence area (150 km2). We analyzed the regression relationship between 5 factors of soil erosion and the area of sub-river basin. The conclusion is that these 5 factors will change along with the changing of sub-river basin area. To eliminate the influence of scale effect, we further use 1000 km2 as standard sub-river basin area to calculate the sediment transport modulus map using Kriging interpolation method.2. The effect factors of soil erosion(1) Precipitation. The rainfall erosivity for every meteorological station was calculated on the Wujiang River Basin. The results show that the average annual rainfall erosivity changed from 1751 to 4349 MJ?mm/(hm2?h?a), and the mean of it is 2788 MJ?mm/(hm2?h?a) with the distribution of Southern North high-low, East West high-low. The rainfall erosivity is concentrated in summer (May-August), and which account for 90.42% of the annual rainfall erosivity. During 1980-2000, the trend rate of the average rainfall erosivity was positive. It indicates that the rainfall erosivity on the whole Wujiang River Basin is increasing. By using the conception of region gravity center, the change for the gravity center of rainfall erosivity in the different years is estimated as well as relation to the gravity center of slope, land use, geometric and the spatial location. Also, the effect of the distribution of rainfall erosivity on soil erosion is analyzed.(2) Topographic features. The type of erosion is water erosion on the Wujiang River Basin. According to the strength, the erosion can be divided into five degree which are less slight, slight, moderate, strength and pole-strength. The less slight and slight erosion are the main erosion. Based on the overlay analysis of soil erosion and DEM in 2000, the elevation, slope, amplitude and surface roughness are selected to analyze the geomorphology of soil erosion. From elevation, the less slight and slight erosion appear single peak which focus on the rocky hills basin areas in Guizhou with the elevation of 1100-1300 meters. The moderate and above degree erosion concentrate on areas with the elevation of 500-900 meters and 1300-1700 meters. From the slope, the distribution of the moderate and below degree erosion decrease. The area scale of the strength and pole-strength erosion increase with the surface rolling increase firstly and then decrease; 15°is a threshold. The erosion first increases and later decreases along with the increase of amplitude, with a threshold of 7-16 meters. The rougher the land surfaces, the lighter soil erosion. Soil erosion intensity is not sensitive to amplitude and surface roughness. (3) Land use. Analysis on the relationship between soil erosion and land use suggests that different land use types have different degrees of soil erosion, i.e. barren land < other land < construction land < forest land < water area < paddy field < grassland with low coverage < shrub land < grassland with high coverage < open forest land < grassland with medium coverage < dry land. Among these land use types, dry land, grassland with medium coverage and open forest land are mostly frequently accompanied with soil erosion. All types of soil erosion concentrate on areas with vegetation coverage of 50%-60%. When the vegetation coverage falls to 50%-60%, the most serious soil erosion happens. In Karst areas, the degree of soil erosion is strongly related to soil thickness, which is different with the situation in non-Karst areas where soil erosion intensified along with the degradation of vegetation.3. Soil erosion simulation and predictionUsing GIS platform, we calculated the average annual soil erosion modulus in the 1980s and 1990s. Our results coincide with previous soil erosion investigation. However, difference exists because RUSLE model does not have gravitational erosion included. Potential soil erosion modulus in the 1990s is higher than that in the 1980s, which suggests an increase in potential soil erosion in Wujiang River Basin. San-Cha river basin and Ma-Ti river/Yin-Jiang river have the highest value of potential soil erosion modulus. Among three main land cover types, forest land has the most soil conservation, grassland the least, while farmland in the middle. This is different with non-Karst area where grassland is better than farmland in soil conservation.We use BP Neural Network model to calculate soil erosion modulus in the 2000s based on the data in the 1980s and 1990s. The results suggests soil erosion decreases from 2313 t/(km~2·a) in the 1990s to 101 t/(km~2·a) in the 2000s, the amount of soil erosion from 115.18×10~6 t/a to 5.03×10~6 t/a. Soil erosion in San-Cha river is under control. Qianxi, Jinsha, Xifeng, Xiuwen, Guiyang, Pingban, Sinan, Shiqian, Yanhe and Songtao need strict soil management in 12th Five-Year Plan period.更多还原