Skip to content

Machine-learning-based Earthquake Catalog Reveals A Clearer View of the Current Phase of Unrest at Campi Flegrei Caldera

Xing Tan1, Anna Tramelli2, Sergio Gammaldi2, Gregory C. Beroza1, William Ellsworth1, Warner Marzocchi3

  • Affiliations: 1Geophysics Department, Stanford University, Stanford, California, United States 2Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano, Naples, 3Deptartment of Earth, Environmental, and Resources Sciences, University of Naples, Federico II, Naples, Italy 

  • Presentation type: Poster

  • Presentation time: Monday 16:30 - 18:30, Room Poster Hall

  • Poster Board Number: 165

  • Programme No: 3.1.57

  • Theme 3 > Session 1

Abstract

Artificial intelligence provides an opportunity to fully harness the potential of a highly efficient seismic monitoring network, such as the one installed in the Campi Flegrei caldera following the significant increase in seismicity recorded in 2021. Three years of continuous seismic data were reanalyzed using a machine learning system specifically trained to identify VT earthquakes in a highly anthropized area like Campi Flegrei. This approach increased the manually compiled seismic catalog from around 9,173 to over 46,350 earthquakes, resolving a fine-grained image of the caldera structure. The developed tool aims to mitigate seismic and volcanic risks in the region by enabling the automatic detection of potential earthquake migrations or the activation of previously seismic areas, which could indicate magma movements.