← Back to Sessions

Application of TCN, UMAP, and XGBoost to Pg and Lg Wave Amplitude to Identify Mining vs. Non Mining and Deep vs. Shallow Events

Description: 

Using the CNRS bulletin from the seismic network in Morocco, we collected 4542 regional waveforms for both mining and non-mining events occurring around the station MD31. The events have occurred within 50 to 250 km of the station. Using this mixed bag of waveforms and criterion that SNR must be greater than 2, we selected 260 mining and 161 non-mining events to conduct this study. We processed these waveforms to construct a meta-data file consisting of the RMS amplitude of the Pg and Lg waves in five frequency bands at 4-6, 6-8, 8-10, 6-12 and 8-16Hz. The mining events occurred during the day and they had characteristically strong Pg and weak Lg waves, especially at high frequencies. We observed that while the Pg and Lg amplitude alone did not separate, the high-frequency Pg/Lg amplitude ratios separated the two population quite well. Mining events are shallow and occurred at a single site. For this reason alone we did not perform any MDAC correction to these amplitudes. The supervised 80:20 and 50:50 training and testing of data and analysis of the impact factor of the attributes using the XGBoost algorithm illustrated a high level of success in identifying the source types. Using a supervised trained 90:10 model, we could further identify additional source types for about 428 events in the area. We are currently applying the TCN and UMAP to the same data set to determine the performance in the classification of these events from Morocco. We are now compiling a similar data set in areas where earthquakes occur at many depths, where deep earthquakes may exhibit strong P and weak-to-moderate level of Lg waves. Our primary focus is to evaluate the performance of the algorithms in identifying “deep vs shallow” and “mining explosions vs. shallow and deep” earthquakes.

Session: Machine Learning for Full Waveform Inversion: From Hybrid to End-to-End Approaches - I

Type: Oral

Date: 5/3/2024

Presentation Time: 05:00 PM (local time)

Presenting Author: Chandan

Student Presenter: No

Invited Presentation: 

Authors