四川省滑坡灾害气象预报方法及其在延伸期-月尺度中的应用

Meteorological Forecast Method of Landslide Disaster in Sichuan Province and lts Application in Extended Period Month Scale

  • 摘要: 分析四川省2014—2018年滑坡灾害与前期降水量之间的关系,构建滑坡灾害气象预报模型,并验证模型在延伸期-月尺度中的应用效果,为四川省延伸期滑坡灾害预报工作提供理论参考。利用历史实测降水数据对四川省滑坡灾害发生的前45 d、35 d、25 d、15 d、10 d和6 d降水量进行统计分析,结果显示,滑坡灾害的发生与其相关性是逐步递减的;在此基础上,结合逻辑回归法分析滑坡灾害发生前2 d降水量的权重关系,得到降水诱发滑坡灾害概率值;再确定滑坡灾害预报模型的易发性背景值与降水诱发滑坡灾害概率值的权重系数;最后,分别计算川东、川西和川西南三个区域临界雨量,优化滑坡灾害预报模型。结果表明:四川省滑坡灾害发生当日、前1 d和前2 d降水量对灾害的影响权重分别为0.048、0.031和0.029;滑坡灾害模型的易发性概率值和降水诱发滑坡灾害概率值的权重参数川东为0.249和0.751,川西为0.655和0.345,川西南为0.501和0.499;分区域临界雨量计算结果有明显差异,川东地区临界雨量值最大,川西南地区次之,川西地区最小。利用CFSv2降水预报产品对滑坡灾害预报模型在延伸期-月尺度中的应用效果进行验证,证实模型对延伸期滑坡灾害整体有较好的预报效果。

     

    Abstract: This paper analyzes the relationship between the landslide disaster and the early precipitation in Sichuan Province from 2014 to 2018, constructs the meteorological forecast model of landslide disaster, and verifies the application effect of the model in the extend-month scale, which provides theoretical reference for the prediction of landslide disaster in Sichuan Province during the extend-month period.Based on the historical measured precipitation data, the precipitation in the first 45-day, 35-day, 25-day, 15-day, 10-day and 6-day of geological disasters in Sichuan Province was statistically analyzed.The results showed that the occurrence of landslide and its correlation is gradually diminishing.On this basis, combined with logistic regression method, the weight relation of the precipitation in the 2-day before the occurrence of landslide disaster was analyzed, and the probability value of precipitation induced landslide disaster was obtained.The weight coefficients of the background value of the landslide disaster prediction model and the probability value of the landslide disaster induced by precipitation are determined again.Finally, the critical rainfall in the three regions of Eastern Sichuan, Western Sichuan and Southwest Sichuan are calculated respectively, and the landslide disaster prediction model is optimized.The results show that the influence weight of precipitation on the day of occurrence, the precipitation in the 1-day and the precipitation in the 2-day before the landslide disaster in Sichuan Province are 0.048, 0.031 and 0.029 respectively.The weight parameters of the probability value of the landslide disaster model and the probability value of the landslide disaster induced by precipitation are 0.249, 0.751 in Eastern Sichuan, 0.655 and 0.345 in Southwest Sichuan, and 0.501 and 0.499 in Western Sichuan.There are obvious differences in the calculation results of regional critical rainfall.The critical rainfall value is the highest in Eastern Sichuan, the second in Southwest Sichuan, and the smallest in Western Sichuan.The application effect of the landslide disaster prediction model in the extend-month scale is verified by using CFSv2 precipitation prediction product, which proves that the model has a good overall prediction effect on the extend-month landslide disaster.

     

/

返回文章
返回