潜在泥石流的界定与判识——以金沙江流域溪洛渡库区为例  

The Definition and Identification of Potential Debris flow Gully——Taking the Xiluodu Reservoir Area as an Example

作　　者：庄建琦[1] 裴来政[2] 丁明涛[3] 陈兴长[1] 

机构地区：[1]中国科学院.水利部成都山地灾害与环境研究所,四川成都610041 [2]中国科学院山地灾害与地表过程重点实验室,四川成都610041 [3]中国科学院研究生院,北京100049

出　　处：《灾害学》2009年第4期1-5,10,共6页Journal of Catastrophology

基　　金：中国科学院知识创新工程重要方向项目(KZCX2-YW-302);"十一五"国家科技支撑计划项目(2006BAC10B04)

摘　　要：选择在建的溪洛渡库区作为研究区域,通过详细的野外考察,确定金沙江流域溪洛渡库区干流共有现代活动泥石流沟57条。利用M atlab构建SOM神经网络模型,依据这57条泥石流样本,选择流域面积、主沟长度、相对高差、沟床比降、平均坡度、相对切割程度、圆状率和侵蚀程度等8个指标,对干流46条沟谷进行预测。预测结果分为:①非泥石流沟有19条,分布在雷波-永善三角台地上;②低危险潜在泥石流沟有14条,分布在库区尾端;③高危险潜在泥石流沟有13条,分布于库区中间位置。预测结果可以为库区生态修复和工程治理提供依据。Taking the Xiluodu reservoir area as the research region, 57 active debris flows gullies in the Xiluodu reservoir area are identified on the basis of the detailed location investigation. By using the Matlab software to construct SOM neural network model and based on the drainage samples of the 57 active debris flows, 8 indexes including area, the length of main gully, gully gradient, average gradient, relative incision degree, circular ratio and erosion degree area are selected to predict the risk degree of 46 gullies along the Jinshajiang river. According the predicted results, the 46 gullies are divided into three grades： ①19 non-debris flow gullies which are distributed on the Leibo-Yongshan triangular platform. ②14 low dangerous debris flow gullies which are distributed on the trail of the reservoir area.③13 high dangerous debris flow gullies which are distributed on the central reservoir area. These predicted results can provide the basis for ecological recovery and engineering control in the reservoir area.

关 键 词：潜在泥石流沟 判识 SOM 溪洛渡库区 金沙江流域 

分 类 号：D642.23[政治法律—政治学]