贵州省降雨型滑坡预警预报模型2020年应用检验分析

Operational Practicability Test of Rainfall-induced Landslides Forecast Model in Guizhou Province in 2020

基于概率拟合曲线的贵州省降雨型滑坡预警预报模型,利用降水预报数据和实况数据对2020年强降水背景下的27次降雨型滑坡事件进行预警预报,从两个方面检验了该模型的业务实用性。结果表明:利用降水预报数据开展降雨型滑坡预警预报,提前3 d和提前1 d的效果优于提前2 d,其准确率均为63%,具有一定的业务实用性;利用降水实况数据开展降雨型滑坡预警预报,滑坡实际发生的位置与模型预警预报中可能性较大以上的区域吻合度较高,进一步验证该模型具有一定的准确性和实用性;能否准确把握降水强度和落区是影响降雨型滑坡预警预报准确率的关键因素。

Based on the rainfall-induced landslides forecast model with probability fitting curve in Guizhou Province,using the precipitation forecast data and observation data of 27 rainfall-induced landslides under the background of heavy precipitation in 2020,the operational practicability of the model was tested.The results show that the accuracy rate of rainfall-induced landslides forecast model of 3d in advance and 1d in advance are 63%,which is better than that of 2d in advance.The accuracy rate shows that with the forecast precipitation data,the rainfall-induced landslides forecast model has certain operational practicability.The actual location of the landslide is in good agreement with the region that is more likely to occur in the model early warning and prediction.It reconfirms the operational practicability of the rainfall-induced landslides forecast model.The precipitation intensity and falling area are particularly important factors affecting the accuracy of early warning and prediction of the rainfall-induced landslides.