重庆地区降雨型滑坡相关性分析及预报模型

Correlation analysis and prediction model for rainfall-induced landslide in Chongqing area

整理大量重庆地区降雨诱发滑坡资料,对降雨因子、累积降雨因子、前期降雨衰减系数与滑坡发生与否及滑坡发生数量之间进行系统的相关分析。结果表明:(1)当日降雨量对滑坡的诱发作用最显著,当日最大小时降雨与当日降雨量高度相关;(2)据滑坡发生前日期增加,降雨因子与滑坡相关及偏相关系数随之降低,在据滑坡发生前3d之后已变得很小;(3)大范围滑坡一般在暴雨当日或滞后一天产生,仅有短时强降雨很难诱发大范围滑坡,还需要一定前期降雨量为基础;(4)累积降雨因子与滑坡因子间都显著相关,且随着累积降雨日数的增加,其相关系数先上升后下降;(5)重庆地区适宜的前期降雨衰减系数在0.6左右。在此基础上拟合出了有效累积降雨量-滑坡发生概率预报模型,并使用当日最大小时降雨量-滑坡发生概率模型对其进行修正。最后使用降雨诱发滑坡实例检验该修正模型,结果可为重庆地区降雨型滑坡的预测预报提供科学依据。

This paper collates a large number of date about rainfall-induced landslides in Chongqing area,doing systemic correlation analysis among the rainfall factors,cumulative rainfall factor,antecedent rainfall attenuation coefficient and landslide or not and the quantity of landslide occurrence. The results showed that:( 1) The effect of daily rainfall on landslide induced is the most significant,the date of daily maximum hourly rainfall is highly correlated to daily rainfall;( 2) According to the date increase before landslide occurred,the correlation and partial correlation coefficients between rainfall factor and landslide decrease,it has become very small in 3 days and longer before landslide occurred;( 3) Large-scale landslide are generally occurred in the rainstorm day or lag day,only short-time heavy rainfall is difficult to induce large-scale landslide,we also requires a certain antecedent rainfall as a basis;( 4) The cumulative rainfall and landslide factors are significantly correlated,and with the number of accumulated rainy days increasing,the correlation coefficient first increased and then declined;( 5) The suitable antecedent rainfall attenuation coefficient is about 0. 6 in Chongqing area. Fitting out the effective cumulative rainfall and landslide occurrence probability forecasting model on that basis,and using the daily maximum hourly rainfall and landslide occurrence probability model to amend its correction. Finally,testing the modified model by using rainfall-induced landslide examples,the results could provide a scientific basis for rainfall-induced landslide prediction in Chongqing area.