Lotfy Samy

Many precursory signal patterns arise before the initial P-wave arrival. The origin of these patterns can be traced to either instrumental or natural ground effects. Identifying these signals’ origin and nature helps to enhance automatic P-wave identification and earthquake source parameter determination. Several investigations have been accomplished to define the origin of these precursory signals, starting from installing a temporary station with 50 and 200 sampling rates for 17 months, in addition to the original station with a sampling rate of 100 samples per second at the New Abu Dabbab station (NADB). From these two stations, 429 earthquakes were recorded at three sampling rates: 50, 100, and 200 samples per second. Then, a spectral analysis was performed, where it was found that there was a clear relationship between the first pattern and the acquisition sampling rate. This proved that the first pattern (the ramping pattern) is an instrumental artifact that occurs from the finite impulse response (FIR) filter in the analog-to-digital conversion stage of the digitizer. The other type of precursory signals was found to have varying spectra independent of the used sampling rate. This led to the differentiation of the natural ground origin and instrumental artifacts-based signals in front of these earthquakes. By automating the classification of these patterns, the true P-wave arrival can be determined in real-time processing, reducing the error in P-wave arrival timing. The current study also introduces this automatic classification by developing various machine learning (ML) and Convolutional Neural Network (CNN) models to highlight the features characterizing each pattern. The classifier topology distinguishes these patterns, and then the automatic P-wave detector can decide the arrival of the P-wave depending on the type of precursory. In instrumental artifacts, the arrival is taken after the precursory and before it in the case of the natural ground-based pattern. The examined classification topologies are Logistic Regression (LR), Kneighbors Classifier (KNN), Support Vector Machine (SVM), Decision Tree Classifier (DT), Random Forest Classifier (RF), XGB Classifier, Naïve Bayes (NB), Voting Classifier and Convolutional Neural Network (CNN). Out of these methodologies, the CNN outperformed with a classification accuracy of 97.67% because of its capability in feature extraction.

A digitizer is considered one of the fundamental components of an earthquake monitoring system. In this paper, we design and implement a high accuracy seismic digitizer. The implemented digitizer consists of several blocks, i.e., the analog-to-digital converter (ADC), GPS receiver, and microprocessor. Three finite impulse response (FIR) filters are used to decimate the sampling rate of the input seismic data according to user needs. A graphical user interface (GUI) has been designed for enabling the user to monitor the seismic waveform in real time, and process and adjust the parameters of the acquisition unit. The system casing is designed to resist harsh conditions of the environment. The prototype can represent the three component sensors data in the standard MiniSEED format. The digitizer stream seismic data from the remote station to the main center is based on TCP/IP connection. This protocol ensures data transmission without any losses as long as the data still exist in the ring buffer. The prototype was calibrated by real field testing. The prototype digitizer is integrated with the Egyptian National Seismic Network (ENSN), where a commercial instrument is already installed. Case studies shows that, for the same event, the prototype station improves the solution of the ENSN by giving accurate timing and seismic event parameters. Field test results shows that the event arrival time and the amplitude are approximately the same between the prototype digitizer and the calibrated digitizer. Furthermore, the frequency contents are similar between the two digitizers. Therefore, the prototype digitizer captures the main seismic parameters accurately, irrespective of noise existence.

We develop a random forest (RF) model for rapid earthquake location with an aim to assist earthquake early warning (EEW) systems in fast decision making. This system exploits P-wave arrival times at the first five stations recording an earthquake and computes their respective arrival time differences relative to a reference station (i.e., the first recording station). These differential P-wave arrival times and station locations are classified in the RF model to estimate the epicentral location. We train and test the proposed algorithm with an earthquake catalog from Japan. The RF model predicts the earthquake locations with high accuracy, achieving a mean absolute error (MAE) of 2.88 km. As importantly, the proposed RF model can learn from a limited amount of data (i.e., 10% of the dataset) and much fewer (i.e., three) recording stations and still achieve satisfactory results (MAE < 5 km). The algorithm is accurate, generalizable, and rapidly responding, thereby offering a powerful new tool for fast and reliable source-location prediction in EEW.

We develop a deep learning framework based on deep image prior (DIP) and attention networks for 3-D seismic data enhancement. First, the 3-D noisy data are divided into several overlapped patches. Second, the DIP network has a U-NET architecture, where the input patches are encoded to extract the significant latent features, while the decoder tries to reconstruct the input patches using these extracted features. Besides, the attention network is used to scale the extracted features from the encoder and the decoder. Third, the attention network output of the encoder is concatenated with that of the decoder to obtain high-order features and guide the network to extract the most significant information related to the seismic signals and discard the others. Finally, the 3-D seismic data are reconstructed using the output patches obtained by the DIP network. The proposed algorithm is an iterative and unsupervised approach, which does not require labeled data. We evaluate the proposed algorithm using several synthetic and field data examples. As a result, the proposed algorithm shows the ability to enhance the 3-D seismic data by attenuating the random noise and preserving the 3-D seismic signal with minimal signal leakage. Moreover, the proposed algorithm shows good denoising performance when tested using various types of events, e.g., linear, hyperbolic, low and high dominant frequencies, and weak amplitude. In addition, the proposed method outperforms the predictive filtering (PF) and damped rank-reduction (DRR) methods. To further understand the principle of the proposed method inside the DIP network, we analyze the weighting matrices in the encoder and decoder parts in detail. We attribute the denoising ability of the DIP network to the improvement of the extracted basis features from the encoder to the decoder layers through a deep network.

The accurate detection of P-wave arrival time is imperative for determining the hypocenter location of an earthquake. However, precise detection of onset time becomes more difficult when the signal-to-noise ratio (SNR) of the seismic data is low, such as during microearthquakes. In this letter, a stacked denoising autoencoder (SDAE) is proposed to smooth the background noise. The SDAE acts as a denoising filter for the seismic data. In the proposed algorithm, the SDAE is utilized to reduce background noise such that the onset time becomes more clear and sharp. Afterward, a hard decision with one threshold is used to detect the onset time of the event. The proposed algorithm is evaluated on both synthetic and field seismic data. As a result, the proposed algorithm outperforms the short-time average/long-time average and the Akaike information criterion algorithms. The proposed algorithm accurately picks the onset time of 94.1% for 407 field seismic waveforms with a standard deviation error of 0.10 s. In addition, the results indicate that the proposed algorithm can pick arrival times accurately for weak SNR seismic data with SNR higher than -14 dB.