基于LSTM的地下水水位预测及地震前兆异常分析

LSTM-based Groundwater Level Prediction and Anomaly Analysis of Earthquake Precursors

由于在地震之后的地下水水位会受其影响而发生异常变化,通过水位变化的预测研究对于分析可能的地震前兆、次生灾害的减轻等有重要的理论以及现实意义。本研究首先选取了2015年4月13日10时28分在云南省红河哈尼族彝族自治州建水县发生的4.7级地震,通过获得该地震周边观测井的地下水水位数据,利用地震活跃期以及非活跃期的理论将数据集进行划分,然后使用LSTM模型进行训练,最后发现模型预测值与真实值之间的差异。实验结果表明,该模型可以发现因为地震而导致的地下水水位异常变化,对研究地下水水位异常变化作为地震前兆具有一定启示意义。

Due to the influence of earthquakes,groundwater levels can undergo abnormal changes.The prediction study of water level change has important theoretical and practical significance for the analysis of possible earthquake precursors and the mitigation of secondary disasters.In this study,the earthquake selected occurred on April 13,2015,at 10:28 a.m.,in Jianshui County,Honghe Hani and Yi Autonomous Prefecture,Yunnan Province.By obtaining groundwater level data from observation wells in the vicinity of the earthquake,the dataset is divided based on the theoretical concepts of earthquake active and non-active periods.Subsequently,an LSTM model is employed for training.The differences between the predicted values and the actual values are observed.Experimental results indicate that the model can identify abnormal changes in groundwater levels caused by earthquakes,providing valuable insights into studying groundwater level anomalies as earthquake precursors.