基于ARDL模型的滑坡地下水水位预测

Prediction of the landslide water table based on ARDL model

滑坡稳定性与地下水水位变化有密切的关系,准确预测滑坡地下水位的波动过程,能有效地开展滑坡预警并降低成本。利用中林滑坡的监测数据,建立降雨量和地下水水位的自回归分布滞后(ARDL)模型,能突破隐藏裂隙的优先流难以刻画的局限,确定降雨引起地下水水位变化的滞后时间、影响系数和有效时段。通过统计在不同降雨强度下地下水水位变幅的变异系数,分析影响滑坡地下水水位的有效降雨量阈值,并对ARDL模型进行修正。将模型预测值与实测值进行对比,验证模型对滑坡地下水水位预测的有效性。研究表明,修正后的模型能更好地预测强降雨条件下滑坡地下水水位迅速涨落的过程。

Precipitation is the most important factor to the occurrence of a landslide. There is a close relationship between the water table and the stability of the landslide. Prediction of the landslide water table variation is an effective way of landslide early warning and reduction in cost. Data of rainfall and water table in the Zhonglin landslide were used to establish the ARDL model which can solve the problem of preferential flow in hidden cracks and analyze the influence lag time,the degree of influence and the effective period of rainfall on groundwater. The variation coefficient of water table under different rainfall intensity conditions was used to examine the effective rainfall threshold of the landslide water table. The model was modified by the effective rainfall threshold. Model prediction was compared with real data,which test the usefulness of the model to predict the landslide water table. The research results show that the process of the rapid landslide water table fluctuation under the condition of heavy rainfall can be predicted accurately with the modified model.