Abstract: The 2014 Ludian, Yunnan M_S6.5 earthquake triggered an unusually large number of co-seismic landslides compared with other events of similar magnitudes in southwestern China. We use landslide-area ratios(LAR)based on a grid(cell size ≈1km²)to delineate the specific distribution of landslides rather than the averaging landslide density that is commonly employed. Results show that the distribution of co-seismic landslides triggered by the 2014 Ludian M_W6.5 earthquake contains LAR values ranging from less than 1.0% to 36.1% with the high values concentrating in river valleys of the study area. Then we examine the correlation between some topographic parameters and the co-seismic landslides in each grid cell and especially focus on the grids with larger LAR(>10%). In addition to examining the correlation between the elevation, local relief and average slope angle(S), we propose a notion of the expected slope degree(ES)to further analyze the correlation between these co-seismic landslides and local topographic features. The expected slope degree can be calculated by dividing the local relief(the difference between the maximum and minimum elevation)by cell length for each grid cell, which is a kind of smooth of the local terrain instead of the average slope. After this procedure, we make a regression analysis on the expected slope degrees and the average slope degrees for all grid cells. Finally, we choose those cells that deviate from the regression line as the unstable units and examine their relation with landslide distribution. Results show that the elevation has a negative relationship with both the average slope degrees and the local relief in the study area, which can be classified as a kind of negative topography implying strong river erosion on the landform. In general, there exists a good positive linear correlation between expected slope degrees and average slope degree for the study area. While larger LAR values(LAR≥10%)lie at the sites that deviate from the regression line, likely representing unstable slopes prone to landsliding. Moreover, it is found that most of these sites with high LAR are just located along river valleys, which permits to predict that mass wasting by future earthquakes or other factors will recur at these places. As a contrast, the analysis results of topographic parameters in the Jiuzhaigou earthquake affected area show different characteristics. Thus, the understanding of the relationship between the geomorphic features and the spatial distribution and scale of landslides is helpful to improve the accuracy of prediction on earthquake-induced landslides and to identify potential large-scale landslides in the early stage.landscape evolution, landslides, adaptive adjustment, expected slope, the Ludian earthquake

摘要： 长期、 缓慢的地貌演化具有阶段性的特点, 构造抬升与侵蚀相互作用引起山坡物质运移, 使地貌单元具有向相对稳定状态转变的趋势。 滑坡作为山坡物质运移的一种主要方式, 在地貌演化过程中起到了重要作用。 2014年鲁甸M_S6.5地震诱发了异常多的滑坡, 可以看作是该区地貌物质在短时间内发生的集中调整过程。 这些滑坡主要沿河流分布, 表明河流侵蚀使河岸地形变陡、 强度降低, 形成发生物质运移的有利条件, 从而增强了地震滑坡的易发性。 文中以SRTM 30m数字高程模型(DEM)为基础, 通过对鲁甸地震滑坡分布区的网格化划分, 对研究区滑坡分布及其与地形特征的关系进行了定量分析。 除计算网格单元内的高程、 高差及算数平均坡度外, 还提出期望坡度的计算方法以对网格单元内的地形进行平滑。 在此基础上, 对该区域地貌特征参数自相关性进行了分析和比较, 以判断地表物质分布是否均衡并寻找其中的分异性单元(滑坡易发区)。 结果表明, 研究区的高程与坡度、 地形高差呈负相关, 反映出显著的河流侵蚀效应; 其中地形特征在分析单元的期望坡度与算数平均坡度这2个不同尺度下表现出很高的一致性, 可能代表着研究区地貌在演化中具有的一种动态稳定特征, 而与此特征不符的地貌单元则是可能发生滑坡进行物质调整的区域, 是地貌自适应调整的一种表现。 2014年鲁甸地震触发的大部分规模较大的滑坡发生在期望坡度与平均坡度差异较大的区域, 这些区域大多位于河谷, 显示河流侵蚀及其所造成的地形特征对滑坡易发性的控制作用。 基于这样的认识, 认为该区未来的物质运移区域仍然受到河流侵蚀的控制, 滑坡易发性高的位置仍将沿河流分布。 作为对比的九寨沟地震震区的地貌参数分析结果则表现出不同的特点, 这种地形地貌分布上的差异性与滑坡空间分布及滑坡规模等之间的关系值得深入探讨。

关键词: 地貌演化, 滑坡, 相关性分析, 期望坡度, 鲁甸地震