摘要
以泥石流灾害极为发育的白龙江流域为研究区,分析区域泥石流爆发的时空规律,构建泥石流堵江概率预测模型,实现区域泥石流堵江灾害危险预测,为区域泥石流堵江链式灾害的风险防控提供参考.基于白龙江流域泥石流爆发的时空规律,在野外调查和资料搜集的基础上,选取流域面积、流域高差、沟床比降、岩性、滑坡密度、日降雨量＞50 mm年均次数、河谷宽度、河流流量和河谷比降共10个参数作为影响泥石流沟是否会形成堵江灾害的关键因子,基于机器学习模型构建泥石流堵江概率的预测模型,模型的精确度为91.3％,ROC曲线下面积为0.931.研究发现河谷宽度是泥石流堵江灾害最重要的影响因素,其次是流域面积、沟床比降、流域高差和滑坡密度等.
Abstract
Taking the Bailong River Basin,where debris flow disasters are extremely developed,as the study area,the temporal and spatial laws of regional debris flow outbreaks were analyzed,and the proba-bility prediction model of debris flow damming constructed to realize the risk prediction of debris flow damming disasters within the region,hoping to provide a reference for the risk prevention and control of regional debris flow chain disasters.Based on the temporal and spatial pattern of large-scale debris flows in the Bailong River Basin,and on the basis of field investigation and data collection,we selected 10 parameters,including drainage area,drainage height difference,gully bed gradient,lithology,landslide density,annual average times of daily rainfall greater than 50 mm,river width,river discharge and river gradient,as the key factors determining whether debris flow gullies would block the river.Based on the machine learning model,a prediction model for the probability of debris flow blocking the river was con-structed,whose accuracy was 91.3％,and the area under the receiver operating characteristic curve was 0.931.Research has concluded that river width is the most important influencing factor for debris flow blockage disasters,followed by drainage area,gully bed gradient,drainage height difference and land-slide density.
维普
万方数据