Investigating the 2024 Swarm–Like Activity Offshore Kefalonia Island, Aided by Machine Learning Algorithms

Abstract In late February 2024, a swarm–like seismic activity took place north of Kefalonia Island, in the area of central Ionian Islands. Following a machine-learning aided workflow, we compiled an enhanced, relocated seismic catalog of 2495 low- to moderate magnitude earthquakes during a 2–month period. Spatiotemporal analysis reveals a narrow epicentral distribution of nearly E-W alignment, approximately 5 km long, much longer than the length anticipated by common scaling laws for the aftershock area extension of the stronger earthquakes that did not exceed M4. Seismic activity decays at a rate slower than mainshock-aftershock sequences, providing evidence of swarm-like behavior. Fluid diffusion appears to be the critical driving force behind this sequence, effectively reproducing the spatiotemporal diffusion of the analyzed activity, whereas cascade triggering due to stress changes and transfer by the combined effect of the two relatively strongest earthquakes promote the triggering of most of the weaker earthquakes that follow in the sequence. Our ML-enhanced spatiotemporal analysis, along with the computation of 17 focal mechanisms of the stronger earthquakes using waveform modeling, support the presence of a population of smaller faults that strike obliquely in respect to the Kefalonia Transform Fault Zone (KTFZ) forming a strike slip duplex in the area between them.