Hybrid CNN-LSTM Model for Surface Displacement Prediction in Accumulation Layer Landslides

The surface displacement of a sudden small-scale accumulation landslide recorded by the global navigation satellite system（GNSS）of a health center in Shixi village, Huangshi City, as well as the rainfall and soil moisture content of this area are used to predict the surface displacement variations of this landslide by using the hybrid model of convolutional neural network（CNN）and long shorterm memory network（LSTM）. Initially, CNN model is used to extract features from time-series data, and ReLU activation function is used to introduce nonlinear transformation to improve the expression ability of the model. Then the extracted features are transferred to the LSTM layer for time series modeling, and the output of the LSTM is mapped to the specified output dimension through the full connection layer to complete the landslide surface displacement prediction. The results show that compared to traditional methods and standalone LSTM networks, the CNN-LSTM hybrid model exhibits significant advantages in predicting surface displacement of accumulation layer landslides, achieving substantially improved prediction accuracy. The results can provide scientific insights for early warning and mitigation of landslide hazards in Huangshi area.