云南省通建高速公路沿线滑坡降雨阈值估计

Estimation of Landslide Rainfall Threshold along Tong(hai)-Jian(shui) Expressway in Yunnan Province

山区高速公路沿线每年雨季常发生滑坡,严重威胁高速公路的安全运营。目前,降雨阈值估计已成为降雨诱发滑坡气象预警最常使用的方法。然而,有些研究区数据不全(如历史滑坡以及对应的降雨事件),经验统计方法往往受到限制;同时,经验统计方法通常提供的是一个区域阈值,对于高速公路这种线形工程适用性不强。针对以上问题,以云南省通建(通海县-建水县)高速公路为研究对象,利用数值模型(TRIGRS)和最小二乘支持向量机(LSSVM)开发了基于物理模型的降雨阈值计算方法;最终,获得了研究区降雨阈值的空间分布。结果表明,与简单的经验阈值相比,通过数值模型给出的降雨阈值空间分布,不仅能反映研究区不同位置地形条件、土壤厚度等关键指标的差异对降雨阈值的影响,也能给出在不同降雨强度下最容易发生滑坡的区域,可为区域滑坡气象预警提供依据。

Landslides often occur along the highway in mountainous areas during the rainy season,which severely threaten the safety of the highway.Rainfall threshold estimation has become the most commonly used method for meteorological warning of rainfall-induced landslides.However,the data of some study areas are incomplete(such as historical landslides and corresponding rainfall events),and the application of empirical statistical methods is limited.At the same time,empirical statistical methods usually provide a regional threshold,which is not applicable to linear engineering such as highway.To solve the above problems,in this paper relying on the Tong(hai)-Jian(shui)Expressway in Yunnan Province,a physics-based rainfall threshold calculation method is developed by using numerical model(TRIGRS)and least squares support vector machine(LSSVM).Finally,the spatial distribution of rainfall threshold in the study area is obtained.The results show that,compared with a simple empirical threshold,the spatial distribution of rainfall thresholds given by the numerical model can not only reflect the impact of the difference in key indicators such as topographic conditions and soil thickness in different locations in the study area on the rainfall threshold,but also give the areas where landslides are most likely to occur under different rainfall intensity,which can provide basis for regional landslide meteorological warning.