Preprints
https://doi.org/10.48550/arXiv.2112.07551
https://doi.org/10.48550/arXiv.2112.07551
 
30 May 2022
30 May 2022
Status: this preprint is open for discussion.

A study on the effect of input data length on deep learning-based magnitude classifier

Megha Chakraborty1,4, Wei Li1, Johannes Faber1,2, Georg Rümpker1,4, Horst Stoecker1,2,3,5, and Nishtha Srivastava1 Megha Chakraborty et al.
  • 1Frankfurt Institute for Advanced Studies, 60438 Frankfurt am Main, Germany
  • 2Institute for Theoretical Physics, Goethe Universität, 60438 Frankfurt am Main, Germany
  • 3Xidian-FIAS international Joint Research Center, Giersch Science Center, D-60438 Frankfurt am Main, Germany
  • 4Institute of Geosciences, Goethe-University Frankfurt, 60438 Frankfurt am Main, Germany
  • 5GSI Helmholtzzentrum für Schwerionenforschung GmbH, 64291 Darmstadt, Germany

Abstract. The rapid characterisation of earthquake parameters such as its magnitude is at the heart of Earth-quake Early Warning (EEW). In traditional EEW methods the robustness in the estimation of earthquake parameters have been observed to increase with the length of input data. Since time is a crucial factor in EEW applications, in this paper we propose a deep learning based magnitude classifier and, further we investigate the effect of using five different durations of seismic waveform data after first P-wave arrival– 1s, 3s, 10s, 20s and 30s. This is accomplished by testing the performance of the proposed model that combines Convolution and Bidirectional Long-Short Term Memory units to classify waveforms based on their magnitude into three classes– "noise", "low-magnitude events" and "high-magnitude events". Herein, any earthquake signal with magnitude equal to or above 5.0 is labelled as high-magnitude. We show that the variation in the results produced by changing the length of the data, is no more than the inherent randomness in the trained models, due to their initialisation.

Megha Chakraborty et al.

Status: open (until 04 Aug 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2022-4', Filippo Gatti, 13 Jun 2022 reply

Megha Chakraborty et al.

Megha Chakraborty et al.

Viewed

The metrics are limited to the final revised paper since the preprint was posted outside of Copernicus Publications.

Total article views: 104 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
104 0 0 104 1 0
  • HTML: 104
  • PDF: 0
  • XML: 0
  • Total: 104
  • BibTeX: 1
  • EndNote: 0
Views and downloads (calculated since 30 May 2022)
Cumulative views and downloads (calculated since 30 May 2022)

Viewed (geographical distribution)

The metrics are limited to the final revised paper since the preprint was posted outside of Copernicus Publications.

Total article views: 97 (including HTML, PDF, and XML) Thereof 97 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 02 Jul 2022
Download
Short summary
Earthquake magnitude is a crucial parameter in defining its damaging potential, and hence its speedy determination is essential to issue an early warning in regions close to the epicentre. This study summarises our findings in an attempt to apply deep learning-based classifiers to earthquake waveforms, particularly with respect to finding an optimum length of input data. We conclude that the input length has no significant effect on the model accuracy, which varies between 90–94 %.