基于InVEST模型的青弋江流域土壤侵蚀与影响因素研究

为探究安徽省青弋江流域土壤侵蚀的演变规律和驱动因素,采用InVEST模型对该流域2000—2018年的土壤侵蚀特征开展了研究,量化了不同土地利用类型、海拔、坡度下土壤侵蚀状况,并借助地理探测器对流域土壤侵蚀影响因素进行分析。结果表明：(1)2000年、2010年、2018年该流域平均土壤侵蚀模数分别为15.29,14.14,10.74 t/(hm²·a),侵蚀总量分别为1.08×10⁷,1.00×10⁷,0.76×10⁷ t,呈现逐渐减小特征;(2)流域内土壤侵蚀空间差异显著,呈现“北低南高”的分布格局;(3)不同土地利用类型土壤侵蚀模数大小表现为裸地>草地>林地>耕地>建设用地>水体,全流域林地侵蚀量最大,占总侵蚀量的73.71%;(4)地形因子对流域内土壤侵蚀存在显著影响,坡度是青弋江流域土壤侵蚀主导因子,因子间交互作用对土壤侵蚀的解释力均大于单因子,其中坡度与年降水量和土地利用的协同作用解释力最强,分别达22.93%和22.29%;(5)坡地坡度降缓及增加草地和林...     

基于RUSLE模型的2000-2010年长江三峡库区土壤侵蚀评价

[目的]探究三峡库区土壤侵蚀程度,为区域土壤侵蚀合理化治理提供依据。[方法]采用RUSLE模型,结合多源数据（MODIS-NDVI,DEM,土地利用等）,定量评价长江三峡库区2000-2010年的土壤侵蚀时空分布。[结果]①三峡库区除西部地区外,其它地区土壤侵蚀情况严重;2000-2010年,长江沿岸以剧烈侵蚀和极强度侵蚀为主;②三峡库区土壤侵蚀状况在2000-2010年间得到一定程度的改善,强度以上土壤侵蚀面积由2000年的1.71×10⁶ hm²下降为2010年的6.82×10⁵ hm²;库区内平均侵蚀模数由2000年的36.75 t/（hm²·a）下降为2010年的22.79 t/（hm²·a）;③三峡库区经过多年侵蚀治理,强度、极强度、剧烈侵蚀面积逐渐减少的同时,轻度、中度土壤侵蚀面积在逐渐增加。[结论]研究区需要进一步加强对轻度、中度土壤侵蚀的治理。     

基于RUSLE模型的湟水流域土壤侵蚀时空变化

研究黄河上游土壤侵蚀的时空变化对于维持黄河上游生态系统服务功能、保护黄河上游水塔具有重要意义.以黄河上游典型区域湟水流域为研究区,采用RUSLE模型定量评估了该流域2000—2015年土壤侵蚀的时空变化特征,并分析了有无梯田措施下土壤侵蚀的空间变化,从而量化了梯田建设对防治坡面土壤侵蚀的影响.结果表明:2000—201...     

基于地理探测器的岷江上游地区土壤侵蚀变化

为掌握岷江上游地区土壤侵蚀动态变化规律和驱动力,以修正的土壤侵蚀模型(RUSLE)为基础,实现该地区2000—2018年侵蚀定量评价,按照国家水力侵蚀分级标准将其分为6个等级,以斜率变化模型完成其动态变化规律的分析,借助地理探测器实现其变化驱动力的探索。结果表明:微度和轻度侵蚀占据全域总面积的70%以上;近20 a内,全域土壤侵蚀整体得到了有效遏制,整体发展态势相对良好;植被覆盖度、降水和高程是驱动土壤侵蚀强度空间分布格局形成和改变的主要因素,特别是植被度的驱动作用最明显;各因子间产生交互关系时,其协同作用均比单因子产生的驱动作用更明显。岷江上游地区土壤侵蚀强度分布格局差异显著,侵蚀总体得到有效遏制,植被覆盖是驱动该地区土壤侵蚀强度空间分布格局变化的主要因素。     

基于WebGIS的土壤侵蚀模型的研究及应用

土壤侵蚀是一个世界性的环境问题 ,它使土壤的生产力和用水的质量降低 ,从而引起土壤沉积并增加了发生洪涝灾害的可能性。随着世界范围内土壤侵蚀的日益严重以及水土流失问题的突出 ,土壤侵蚀的研究越来越深入。但目前对土壤侵蚀的工作大多局限在模型的改进和其相关因子的选择上 ,缺乏一个良好的用户界面来完成土壤侵蚀模型各个因子以及土壤侵蚀预测结果的分析与其可视化。通过充分利用 GIS的有关技术 ,建立了一个基于 USL E的网络土壤侵蚀模型 ,并设计了应用界面 ,解决了上述问题。试验表明 :这个基于 WEBGIS的土壤侵蚀模型 ,不仅使土壤侵蚀研究成果通过 GIS的良好界面表现出来 ,而且还可以利用 GIS的有关分析工具 ,反过来促进土壤侵蚀模型的改进。     

基于InVEST模型的低山丘陵区土壤侵蚀变化与驱动因素分析

山区最为严峻的生态威胁为土地不合理利用导致的水土流失。为探究山区水土流失时空分布规律及驱动因素,以迁西县为例,利用GIS与InVEST模型定量估算研究区的土壤侵蚀时空变化特征,并通过地理探测器剖析土壤侵蚀的主要驱动因子,以期为区域社会经济与生态环境协调发展提供决策支撑。结果表明:(1)1990年、2000年、2010年与2020年的土壤侵蚀量分别为1.25×10⁷ t,1.41×10⁷ t,1.77×10⁷ t与2.00×10⁷ t,呈不断上升趋势,且在空间上均呈北高南低的分布格局;(2)土地利用类型是土壤侵蚀的最主要驱动因素,工矿用地与未利用地是区域土壤侵蚀风险较高区域,其他因子作用由强至弱依次为坡度、土壤类型、植被覆盖度、降雨;(3)因子间交互作用结果对土壤侵蚀的解释力均大于单一因子的解释力,其中用地类型与其他因子的交互结果对土壤侵蚀的解释力最强。基于以上分析,文章最后提出水土流失治理的措施:加强工矿用地生态修复整治与未利用土地生态保护,植树造林增强其地表植被覆盖度; 控制板栗经济林扩张规模,恢复林下草皮,>25°坡耕地退耕还林还草。     

土壤侵蚀模型研究进展

国内外土壤侵蚀模型的发展过程,可以大致划分为经验统计模型、物理过程模型与分布式模型三个阶段.就这三个阶段,介绍了国内外土壤侵蚀模型的研究成果.并将地理信息系统(GIS)在土壤侵蚀模型中的应用分为三类,一是以GIS为工具,利用GIS提取模型所需因子,然后按照模型要求利用GIS的图形运算和地图代数运算,最后得到计算结果.二是将GIS与土壤侵蚀模型作为两个不同的系统,考虑结合方法的问题.三是利用GIS开发新的模型或改善已有模型.     

基于GIS的区域土壤侵蚀模型设计

区域土壤侵蚀模型是国家和区域土壤侵蚀调查、水土保持宏观规划的支持工具。借鉴国外区域土壤侵蚀模型,以DEM栅格为空间单元,对区域土壤侵蚀的单元模型进行定量表达,包括降雨、植被截留、入渗,微地形填洼等,并利用GIS功能完成径流传递和汇集部分的计算,建立了区域土壤侵蚀模型。所建立模型在延河流域的试运算,结果接近现实,能反映土壤侵蚀时空分布趋势。模型的建立可为水土保持的宏观决策提供支持。     

基于RUSLE模型的淮河流域土壤侵蚀定量评价

为探究淮河流域的土壤侵蚀状况及其空间分布特征,基于GIS技术和RUSLE模型,通过运用降雨、土壤类型、DEM、遥感影像和土地利用类型等数据确定模型中的参数因子,对2000年至2015年间淮河流域涉及的189个县(市、区)的土壤侵蚀强度及其空间分异特征进行了定量分析.结果表明:淮河流域土壤侵蚀区域主要分布在上游河南和湖北...     

基于RUSLE模型和地理探测器的鄂西南土壤侵蚀脆弱性评价

[目的]水土流失是鄂西南突出的生态问题,合理评估该区域的土壤侵蚀脆弱性,并探究其驱动机制是采取有效水土保持措施的前提。[方法]从人地耦合系统视角分析自然和社会经济因素对鄂西南土壤侵蚀的影响,从暴露度、敏感性和适应能力3个方面构建了鄂西南土壤侵蚀脆弱性评价指标体系,并采用综合赋权法确定指标权重。[结果](1)2010—2020年鄂西南土壤侵蚀脆弱性总体上呈现先降低后升高的趋势,强烈及以上强度的土壤侵蚀脆弱性呈现零星的碎片化分布格局,且主要集中在鄂西南的中南部和东部宜昌市辖区的西部地区;(2)研究区敏感性最高的区域集中在海拔800～1 500 m,敏感性在坡度25°～35°最大,并呈现出向两侧递减的趋势,当坡度>15°时,较高及以上敏感性面积急剧增加;(3)较高及以上土壤侵蚀脆弱性高于土壤侵蚀强度,较低及以下土壤侵蚀脆弱性低于土壤侵蚀强度,土壤侵蚀脆弱性与土壤侵蚀强度存在协同变化趋势;(4)土壤侵蚀脆弱性的分布格局是多因素协同作用造成的,地理探测器的分析表明坡度、植被覆盖度、教育质量和城镇化率对土壤侵蚀脆弱性的解释力较强。[结论]未来需要高度重视对坡度>15°地区植被覆盖的保护...     

基于支持向量机的小流域水蚀预报模型研究

 土壤侵蚀过程复杂,很难直接应用土壤侵蚀预报方程进行定量计算。作为一种新的机器学习算法,支持向量机在样本有限的情况下,采用结构风险最小化准则,把学习问题转化为一个二次规划问题,从而得到唯一的全局最优解。首次尝试将最小二乘支持向量机技术用于土壤侵蚀预测,并与BP神经网络的方法进行了对比,取得了较好的预测精度。     

基于机器学习的紫色土坡面水沙预测模型研究

为了精准预测土壤侵蚀变化、科学合理地预防水土流失的发生，为紫土坡地水土流失防治提供参考，基于典型紫土径流小区43场降雨资料，将雨量、历时和I30为特征指标，采用K-means聚类算法对降雨分类；使用随机森林（Randomforest,RF）算法评估各影响因子对径流深（H）、土壤流失量（S）的重要性，结合通径分析法进行重要因子筛选；将筛选所得关键因子作为模型输入变量，H、S为输出，使用BP神经网络构建预测模型。结果表明：（1）划分出3种雨型，B雨型（短历时、大雨强、小雨量）是主要降雨类型，A 雨型（长历时、中雨强、大雨量）最剧烈。（2）各雨型下特征因子对H、S的重要程度明显不同。A雨型下T对S重要程度最高（31%），各因素对H的重要程度相对均匀；B雨型下各因素对S的影响差异小，F对H的重要程度最高（29%）。在C雨型下，Pr对H、S的重要程度最高（33%、36%）。（3）三种雨型下的H、S均受到Pr的显著影响，B、C雨型下的H、S分别同时受F和Vs、Ph的影响显著。（4）利用BP神经网络对H、S的预测精度均较高，Nash-Suttclife 效率系数均高于0.95，且对H的预测模型精确度高于对S的预测模型。     

Sediment transport rate-based model for rainfall-induced soil erosion

A one-dimensional mathematical model, termed sediment transport rate-based model, is developed for determining rainfall-induced soil erosion and sediment transport. The model is comprised of (1) the kinematic-wave equation for overland flow, (2) a transport rate-based advection equation for rainfall-induced soil erosion and sediment transport, and (3) a semi-Lagrangian algorithm for numerical solution of the soil erosion and sediment transport equation. A series of soil flume experiments under simulated rainfalls were conducted to simulate the overland flow and sediment transport and to test the sediment transport rate-based model. Numerical results of sediment transport rate-based model indicate that (i) hydrographs display an initial rising limb, followed by a constant discharge and then a recession limb; (ii) sediment transport rate graphs exhibit the distributions similar to the hydrographs; and (iii) sediment concentration graphs show a steep-receding limb followed by a constant distribution and a receding tail. The numerically simulated hydrographs, sediment transport rate and concentration distributions are in good agreement with those measured in laboratory experiments, demonstrating the efficacy of the transport rate-based model.     

免耕与土壤侵蚀研究进展

综述了国内外免耕控制土壤侵蚀的机理研究以及对土壤性质的影响,并提出免耕在我国的应用方向。     

Regional assessment of soil erosion using the distributed model SEMMED and remotely sensed data

The soil erosion model for Mediterranean regions (SEMMED) is presented and used to produce regional maps of simulated soil loss for two Mediterranean test sites: one in southern France and one in Sicily. The model demonstrates the integrated use of (1) multi-temporal Landsat Thematic Mapper (TM) images to account for vegetation properties, (2) a digital terrain model in a GIS to account for topographical properties and to assess the transport capacity of overland flow, (3) a digital soil map to assess the spatial distribution of soil properties, and (4) a limited amount of soil physical field data. The principle drawbacks of the model are that it does not account for soil particle detachment by overland flow nor for soil surface crusting. The model is most sensitive to the initial soil moisture storage capacity and the soil detachability index. The main advantages of SEMMED are that it simulates processes at a regional scale and, where possible, it uses available data sources such as remote sensing imagery, digital elevation models (DEM) and (digital) soil databases, which usually are not available for smaller catchment areas. Using SEMMED it is possible to produce regional maps of erosion assessments, which are of more practical use in land use planning and land management than simple extrapolations from small plot experiments.     

A pragmatic approach to modelling soil and water conservation measures with a catchment scale erosion model

To reduce soil erosion, soil and water conservation (SWC) methods are often used. However, no method exists to model beforehand how implementing such measures will affect erosion at catchment scale. A method was developed to simulate the effects of SWC measures with catchment scale erosion models. The method was implemented by applying the LISEM model to an agricultural catchment on the slopes of Mt. Kenya. The method consisted of a field scale calibration based on P-factors, followed by application at catchment scale. This calibration included factors such as saturated conductivity, Manning's n, roughness and slope angle. It was found that using data on P-factors, such models can be calibrated to give acceptable predictions at pixel scale. However, P-factors were also found to vary with land use type and storm size. Besides, more data on the physical effectiveness of SWC measures are needed. At catchment scale, the effect of SWC was found to be different from that at pixel scale. Most SWC were simulated to be more effective at catchment scale, indicating additional infiltration during transport through the catchment to the outlet. However, slope corrections in case of terraces were found to be less effective at this scale. Nevertheless, a simulation for current land use with current SWC measures indicated that these SWC measures decrease runoff by 28% and erosion by 60%.     

基于支持向量机的土壤湿度模拟及预测研究

基于中山大学珠海校区气象观测站日平均风速、日平均气温、日平均空气湿度、日平均水汽压、日平均总辐射量、日平均地表温度、日平均降雨量、日平均蒸发量以及日平均10 cm、20 cm、30 cm土层土壤的含水量,利用支持向量机方法建立气象因子与土壤湿度统计关系,并以此为基础建立土壤湿度模拟与预测模型.结果表明,土壤湿度对气象因子有一定滞后相关性,不同土层土壤湿度对气象因子的滞后相关性不同.研究发现考虑滞后相关性的预测模型在精度上要高于不考虑滞后相关性的预测模型.此外,利用气象因子对地下10 cm的土壤湿度模拟与预测精度较高,而对地下20 cm、30 cm的土壤湿度模拟精度较低.利用地下10 cm与20 cm、20 cm与30 cm的土壤湿度相关性大的特点,可以考虑利用支持向量机方法以10 cm土壤湿度模拟与预测20 cm的土壤湿度,以20 cm的土壤湿度模拟与预测30 cm的土壤湿度,分析结果表明模拟精度较高.     

A susceptibility model for post wildfire soil erosion in a temperate oceanic mountain area of Spain

In the North-West of the Cantabrian Range (North of Spain) the climate is oceanic and vegetation cover is continuous. In those areas where livestock farming prevails, wildfires are common, although small in size, their recurrence makes the phenomenon critical for the conservation of soils.In this study we propose that the structural stability of soil, associated with the type and size of the structural aggregates, may be a useful indicator to assess erosion susceptibility in burnt soils. We have chosen an area of approximately 485 km² over quartzitic lithologies where a high recurrence of wildfires has been noted and which displays several forms of erosion: gullying, rilling and erosion by overland flow. We have measured texture, percentage and degree of structural aggregate stability and the rate of infiltration in soils that have been burnt up to 3 times over the last 20 years and also in unburnt soils. The results obtained enable us to establish connections between wildfires, soil deterioration and macro-aggregate stability.We have used the stability of macro-aggregates as an indicator to elaborate a soil erosion susceptibility model for a large area of 10,600 km² with sharp relief and Atlantic climate. The model was constructed by combining three main factors: soil structural stability, fire intensity and relief. Variables related to soil structural stability and presence of basic cations have been derived from lithology (% Clay and % Silt + Fine Sand). The availability of humified organic material has also been taken into account as an additional variable in the formation of stable macro-aggregates. The expected fire intensity was calculated from the amount of inflammable material and the structure of the different vegetation types. Finally, the influence of relief was analyzed by considering the slope steepness.The resultant cartographic model presents five types of post wildfire soil erosion susceptibility. Those areas with the highest risk correspond to quartzite lithology regions, with long, steep hillsides covered with heaths. Those with the lowest risk correspond to limestone bedrock areas with gentle slopes and herbaceous vegetation. The accuracy of the model is determined by the scale of the original thematic cartography: 1:25000, and the cell size of the Digital Terrain Model is 50 × 50 m.     

Effects of soil erosion on long-term soil productivity in the black soil region of northeastern China

China's northeastern Black Soil Region, one of the country's most important crop production areas, has been seriously affected by soil erosion. This study evaluated the effects of soil erosion on the long-term productivity of this region. We used a modified productivity index (MPI) model (MPI is a number between 0 and 1, with 1 indicating highest productivity) to assess the current effects of soil erosion on soil productivity, as well as to predict long-term change in productivity. Samples from 21 black soil profiles yielded varying MPI values, although most MPI values were indicative of moderate productivity. Organic matter content and available water capacity impact MPI values in the region, whereas soil clay content and pH were less important. Overall, organic matter content and available water capacity of soil profiles decreased consistently as depth of erosion increased. Modeling indicated that MPI in the region will decrease by 0.0052 for each centimeter of topsoil eroded; this rate represents 1% of the current average MPI for the study area. The model predicts a 9.6% productivity reduction over 100 years and a 48.3% reduction over 500 years.