汶川震区滑坡堆积体坡面侵蚀量测算方法

Estimation of Erosion from Earthquake Landslides in Wenchuan Area

摘要:

在汶川县选取典型滑坡堆积体，采用三维激光扫描仪实地测量滑坡堆积体形态，用图像分析方法鉴别沟道，用于估算堆积体形成后的土壤侵蚀量。根据滑坡堆积体坡面实测点云数据，采用最小二乘法拟合二次曲线，近似为堆积体坡面沟道各横截面发生侵蚀前的坡面线。结合堆积体发生侵蚀后的坡面沟道实测数据，测算各段沟道侵蚀量后估算出滑坡堆积体坡面侵蚀总量。典型滑坡堆积体坡面侵蚀总量为355m3。采用传统断面法对计算结果进行验证的结果表明，拟合断面法计算的各分段侵蚀沟体积略大于传统断面法计算结果，相对误差为15.6%。5条分段沟道中最大相对误差为27.1%，最小相对误差为7.9%，拟合断面法计算侵蚀沟体积的精度较好。

Abstract:

This paper describes a new method for determining the amount of earthquake-added soil erosion in Wenchuan earthquake area. To study on the method, typical landslides were chosen in Wenchuan County, and the information of typical landslide surface and its surrounding terrain were obtained by on-site measurement. The rill/gulley on the landslides were identified by image analysis, and the amounts of slope erosion from landslides were estimated by numerical calculation. The on-site measurement was carried out by using a 3-D laser scanner (HDS3000, Leica Company). The scanning distance reached as far as 300m, with a distance measurement accuracy of ±4mm at 50m. The scanner was set up in front of the slope of the landslide to be measured. The distances from the scanner to near end and the far end of the landslide were 58.32m and 136.81m respectively. The scan grid was set to 20.0cm horizontal and 20.0cm vertical at 200 m from the scanner. The actual scan grid of the measured landslide was 5.8cm×5.8cm at the near end and 13.7cm×13.7cm at the far end, respectively. Landslides surfaces sections prior to erosion were specified by using least squares curve fitting method, and fitting it with curve of the second degree. The gully volume was estimated by comparing the result of surfaces prior to erosion and on-site data collected with laser scanner. The erosion volume of a typical landslide surface gully from 2008 to 2012 was estimated to be 355m3. The method was verified with the data calculated by traditional cross section area method. The cross sections area and length of gully were calculated by the built in method of the 3-D laser scanner. The result showed that the values calculated by the method suggested were a little bigger than that by the traditional cross section area method. The relative error of the method was 15.6% for the whole gully, with the lowest being 7.9% and the highest being 27.1% for the segmentations. The reasonable accuracy of the method showed that the rill/gully volume of landslide could be estimated reliably by using it in engineering application.