Application of satellite remote sensing for mapping wind erosion risk and dust emission‐deposition in Inner Mongolia grassland, China
Application of satellite remote sensing for mapping wind erosion risk and dust emission‐deposition in Inner Mongolia grassland, China
复制标题
DOI：
10.1111/j.1744-697x.2011.00235.x
复制
发表时间：
2012-03
影响因子：
1.3
通讯作者：
M. Reiche;R. Funk;Zhuodong Zhang;C. Hoffmann;J. Reiche;M. Wehrhan;Yong Li;M. Sommer
中科院分区：
文献类型：
--
作者：
M. Reiche;R. Funk;Zhuodong Zhang;C. Hoffmann;J. Reiche;M. Wehrhan;Yong Li;M. Sommer
过度放牧导致土地退化以及草原生态系统的沙漠化，进而引发严重的环境和社会问题。中国内蒙古的锡林郭勒草原数百年来一直是沙尘汇聚区，受到过度放牧和风力侵蚀的负面影响极大。本研究的目的是提供一幅空间分辨率为25米的风力侵蚀风险图，以确定实际的沙尘源区和汇聚区。采用综合方法，将植被特征和地表粗糙长度\(z_0\)的实地测量结果与先进星载热发射和反射辐射仪（ASTER）图像数据相结合，进行土地利用分类。为确定不同土地利用类型的特征，2009年4月进行了一次实地观测（地面实况调查）。植被高度与\(z_0\)的相关性（\(R^2 = 0.8\)，\(n = 55\)）为区分“草原”、“无植被”和“其他”三个主要类别提供了依据。将土壤调节植被指数（SAVI）与经过大气校正的ASTER 1、2和3波段（可见光到近红外）的光谱信息相结合，得出总体分类精度（OA）为0.79，卡帕（\(\kappa\)）统计值为0.74。此外，利用数字高程模型（DEM）确定与主风向相关的地形效应，从而对潜在的沙尘沉积区进行定性评估。生成的地图对锡林郭勒草原的空间变异性有更显著的描述，反映了草原当前状态下不同的土地利用强度——轻度、中度和高度退化。风力侵蚀风险图能够识别出特征性的矿物沙尘源、汇聚区和过渡带。
Intensive grazing leads to land degradation and desertification of grassland ecosystems followed by serious environmental and social problems. The Xilingol steppe grassland in Inner Mongolia, China, which has been a sink area for dust for centuries, is strongly affected by the negative effects of overgrazing and wind erosion. The aim of this study is the provision of a wind erosion risk map with a spatial high resolution of 25 m to identify actual source and sink areas. In an integrative approach, field measurements of vegetation features and surface roughness length z0 were combined with Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) image data for a land use classification. To determine the characteristics of the different land use classes, a field observation (ground truth) was performed in April 2009. The correlation of vegetation height and z0 (R2 = 0.8, n = 55) provided the basis for a separation of three main classes, “grassland”, “non-vegetation” and “other”. The integration of the soil-adjusted vegetation index (SAVI) and the spectral information from the atmospheric corrected ASTER bands 1, 2 and 3 (visible to near-infrared) led to a classification of the overall accuracy (OA) of 0.79 with a kappa () statistic of 0.74, respectively. Additionally, a digital elevation model (DEM) was used to identify topographical effects in relation to the main wind direction, which enabled a qualitative estimation of potential dust deposition areas. The generated maps result in a significantly higher description of the spatial variability in the Xilingol steppe grassland reflecting the different land use intensities on the current state of the grassland – less, moderately and highly degraded. The wind erosion risk map enables the identification of characteristic mineral dust sources, sinks and transition zones.