A Deep Learning Analysis of Seismic Waveforms for the Estimation of Peak Ground Acceleration
Abstract
The identification and rapid estimation of earthquake parameters such as Peak Ground Acceleration are crucial for Earthquake Monitoring and Earthquake Early Warning especially for strong earthquakes. The conjugation of seismic waves with the layers of geological media of varying elastic properties results in a varying PGA from site to site. The local site effects heavily influence the PGA values , for instance if the site is composed of soft-sediments the amplification within the ground motion is more prominent than that of a rocky terrain or very firm sediments. We use deep learning to model these nonlinearities to identify the PGA value of the incoming earthquake signal. We use the global data STanford EArthquake Dataset, STEAD benchmarked for Machine Learning applications. The proposed model architecture may be considered as a first prototype that could be adapted into EEW systems for effective emergency response.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.S42C0170S