Groundwater Level Prediction and Anomaly Analysis of Earthquake Precursors Based on Attention-LSTM
Using the 4.7-magnitude earthquake that occurred at 10:28 on April 13,2015 in Jianshui County as an example to analyze the potential of groundwater level changes as earthquake precursors.The groundwater level data of the Tonghai high well and the Ji-anshui well around the earthquake are divided into earthquake active and inactive periods,and an Attention-LSTM model is de-signed to train the divided data.The experimental results show that the model can not only capture the abnormal changes in ground-water level caused by the earthquake,but also has ideal effects on different wells during the same earthquake,indicating that the model has transferability and can be widely used in different wells during the same earthquake,providing insights for using ground-water level as an indicator of earthquake precursors.This approach has certain scientific value and practical significance for the re-search method of analyzing groundwater level changes to detect earthquake precursors,and is expected to provide new ideas for earthquake early warning and disaster prevention and mitigation work.
NETL
NSTL
万方数据
