Insights into the 2021 Arkalochori (Crete Island, southern Greece) foreshock swarm through an enhanced deep-learning seismic catalog

On 27 September 2021, a damaging earthquake of M w 6.0 struck central Crete (Greece), close to Arkalochori village; the strongest earthquake ever recorded in this area during the instrumental period. Remarkably, the mainshock was preceded by an intense foreshock swarm that lasted almost four months. However, the lack of local seismological stations severely limited the event-detection capability of conventional methods, rendering an in-depth analysis of the foreshocks' evolution and statistical properties unfeasible. To enrich the seismic catalog, we employed a deep-learning model to detect and pick additional seismic phases automatically. A Bayesian Gaussian Mixture Model was then applied to associate the acquired P- and S-wave picks with specific event origins, followed by single-event location with a local 1D velocity model and double-difference relocation. A new, deep-learning catalog was constructed, with about ten times more events than in the previously published seismic catalogs. The evolution of the swarm presented a complex spatiotemporal pattern, with patches activated at different times. A major M w 4.9 foreshock occurred on 24 July 2021, triggering its own sub-sequence. In addition, the b -value spatiotemporal variations were investigated as possible stress state indicators in the seismogenic area. The analysis revealed persistent lower b -values than the background during the foreshock swarm, with the lowest just before the occurrence of the major M w 4.9 foreshock. Furthermore, the b -values spatial distribution highlighted that both the M w 4.9 foreshock and the M w 6.0 mainshock occurred in low b -value zones, suggesting an increased likelihood of occurrence and a critical accumulation of stress in the epicentral areas of the strongest events. • A deep-learning catalog was constructed for the 2021 Arkalochori foreshock swarm. • The new enhanced catalog provided a large dataset, suitable for statistical analysis. • The b-value of the foreshock swarm is significantly lower than the background value. • The b-values show a global minimum before the major foreshock of 24 July 2021. • A low b-value area is mapped at the locations of the major foreshock and mainshock.

Institutions

Publication Details

Journal
Tectonophysics
Published
2025-12-19
DOI
https://doi.org/10.1016/j.tecto.2025.231045
Primary Topic
earthquake and tectonic studies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Insights into the 2021 Arkalochori (Crete Island, southern Greece) foreshock swarm through an enhanced deep-learning seismic catalog

Tectonophysics
earthquake and tectonic studies
article

Insights into the 2021 Arkalochori (Crete Island, southern Greece) foreshock swarm through an enhanced deep-learning seismic catalog

article en

Abstract

On 27 September 2021, a damaging earthquake of M w 6.0 struck central Crete (Greece), close to Arkalochori village; the strongest earthquake ever recorded in this area during the instrumental period. Remarkably, the mainshock was preceded by an intense foreshock swarm that lasted almost four months. However, the lack of local seismological stations severely limited the event-detection capability of conventional methods, rendering an in-depth analysis of the foreshocks' evolution and statistical properties unfeasible. To enrich the seismic catalog, we employed a deep-learning model to detect and pick additional seismic phases automatically. A Bayesian Gaussian Mixture Model was then applied to associate the acquired P- and S-wave picks with specific event origins, followed by single-event location with a local 1D velocity model and double-difference relocation. A new, deep-learning catalog was constructed, with about ten times more events than in the previously published seismic catalogs. The evolution of the swarm presented a complex spatiotemporal pattern, with patches activated at different times. A major M w 4.9 foreshock occurred on 24 July 2021, triggering its own sub-sequence. In addition, the b -value spatiotemporal variations were investigated as possible stress state indicators in the seismogenic area. The analysis revealed persistent lower b -values than the background during the foreshock swarm, with the lowest just before the occurrence of the major M w 4.9 foreshock. Furthermore, the b -values spatial distribution highlighted that both the M w 4.9 foreshock and the M w 6.0 mainshock occurred in low b -value zones, suggesting an increased likelihood of occurrence and a critical accumulation of stress in the epicentral areas of the strongest events. • A deep-learning catalog was constructed for the 2021 Arkalochori foreshock swarm. • The new enhanced catalog provided a large dataset, suitable for statistical analysis. • The b-value of the foreshock swarm is significantly lower than the background value. • The b-values show a global minimum before the major foreshock of 24 July 2021. • A low b-value area is mapped at the locations of the major foreshock and mainshock.

TectonophysicsVol. 921
National and Kapodistrian University of Athens (GR), Mediterranean University (ME), National Observatory of Athens (GR)
Sustainable cities and communities
Openalex Percentile: Top 31%
earthquake and tectonic studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.