Early Earthquake Detection Using Batch Normalization Graph Convolutional Neural Network (BNGCNN)

Abstract

Earthquake is a major hazard to humans, buildings, and infrastructure. Early warning systems should detect an earthquake and issue a warning with earthquake information such as location, magnitude, and depth. Earthquake detection from raw waveform data using deep learning models such as graph neural networks (GNN) is becoming an important research area. The multilayered structure of the GNN with a number of epochs takes more training time. It is also hard to train the model with saturating nonlinearities. The batch normalization technique is applied to each mini-batch to reduce epochs in training and obtain a steady distribution of activation values. It improves model training and prediction accuracy. This study proposes a deep learning model batch normalization graph convolutional neural network (BNGCNN) for early earthquake detection. It consists of two main components: CNN and GNN. Input to the CNN model is multi-station and three-component waveform data with magnitude ≥3.0 were collected from January 2000 to January 2015 for Southern California. The extracted features of CNN are appended with location information and input to GNN model for earthquake detection. After hyperparameter tuning of the BNGCNN, when testing and evaluating the model on the Southern California dataset, our method shows promising results to the baseline model GNN by obtaining a low error rate to predict the magnitude, depth, and location of an earthquake.

An Early Warning System for Earthquake Prediction from Seismic Data Using Batch Normalized Graph Convolutional Neural Network with Attention Mechanism (BNGCNNATT)

“…Figure 1c represents a layered architecture similar to the model architecture proposed by Ref. [20], where the fourth layer is the batch normalization layer. Overfitting is prevented through batch normalization and dropout regularization, which ensure that the input distribution does not differ between layers.…”

“…Figure 1e shows the proposed CNN block that consists of five layers: convolutional, batch normalization, activation, pooling, and attention layers. The proposed architecture is the combination of Figure 1c [16] and Figure 1d [20]. Experiments on two seismic datasets show that our proposed model outperforms the baseline models of GCNN given in Figure 1a-c as it uses both batch normalization and attention layers to achieve high accuracy.…”

“…Batch size refers to the number of examples given at a time to the model for processing in a single iteration. Small batch sizes improve training by requiring less memory and time [20]. Batch normalization is a process that makes a model faster and more stable by adding an extra layer to the model.…”

Multi-task multi-station earthquake monitoring: An all-in-one seismic Phase picking, Location, and Association Network (PLAN)

“…Several studies have shown that GNNs have the potential to deal with irregularly spaced stations for phase association and event localization [14,16,[35][36][37][38][39]. Here, we build a graph-based network (Fig.…”

“…A standard earthquake monitoring workflow involves a series of steps to detect and characterize earthquakes, including phase picking, association, and event location [1][2][3]. Phase picking, a conceptually simple task which is akin to detection problems in computer vision, has recently been improved through deep learning [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18], where convolutional neural networks (CNNs) [19] are typically used. After the phase picking, traditional [20][21][22][23] and deeplearning-based [24][25][26] phase association algorithms have been used to link seismic phases at multiple stations from the same events.…”