河北省强降水对道桥影响的致灾阈值研究

Research on Disaster Threshold of the Influence of Heavy Rainfall on Roads and Bridges in Hebei Province

强降水诱发的滑坡、泥石流、河水暴涨等会导致道路损坏、桥涵冲垮,造成人员伤亡和经济损失。采用河北省1984—2014年142个国家地面气象观测站的逐日降水资料和各县(市、区)的历史道桥损失灾情数据,根据最优分割法,将历史灾情案例划分为轻、中、重、特重四个等级,以5日累积降水量和大雨日数为致灾指标,基于逻辑回归法确定不同灾情等级的降水致灾阈值,并对不同等级案例的准确率进行统计分析。结果表明:当大雨日数为1、2、3、4天时,特重等级的5日累积降水量阈值分别为338.4 mm、300.4 mm、262.5 mm、224.6 mm;重度等级的5日累积降水量阈值分别为190.0 mm、184.0 mm、171.8 mm、160.5 mm。所有案例中,致灾阈值标准对特重和轻度等级灾情的预测准确率较高,分别为69.2%和67.3%;山区案例的预测准确率明显高于平原的准确率,有更好的评估效果;0天和3天及以上的大雨日数的灾情等级预测准确率高,大雨日数为1~2天时准确率偏低,预报难度较大。

Landslides, mud-rock flows and river surges induced by severe precipitation can result in road damage and collapsed bridge culverts, causing casualties and economic losses.Using the precipitation data of 142 national meteorological stations in Hebei Province and historical disaster data from 1984 to 2014,the historical disaster cases are divided into four levels: light, medium, heavy, and special according to the optimal segmentation method.The cumulative precipitation over five days and the number of heavy rain days are taken as disaster-causing indexes.Based on logistic regression method, the precipitation disaster thresholds of different disaster levels are determined, and the accuracy of cases in different levels is statistically analyzed.The results indicate that when the number of heavy rain days is 1,2,3,4 days, the five-day cumulative precipitation thresholds of the special heavy grades are 338.4 mm, 300.4 mm, 262.5 mm, 224.6 mm, respectively;the five-day cumulative precipitation thresholds of the heavy grades are 190.0 mm, 184.0 mm, 171.8 mm, 160.5 mm, respectively.In all cases, the hazard threshold standard has a higher prediction accuracy for severe and mild-level disasters, which are 69.2% and 67.3%,respectively.The accuracy of mountainous cases is significantly higher than that of plain cases.The prediction accuracy of days with heavy rain of 0 and 3 days or more is high, while the accuracy of days with heavy rain of 1—2 days is low, which makes it difficult to make prediction.