Machine Learning aids rapid assessment of aftershocks: Application to the 2022-2023 Peace River earthquake sequence, Alberta, Canada

Jinji Li, Jesus Rojas-Parra, Rebecca O. Salvage, David W. S. Eaton, Kristopher A. Innanen, Wenhan Sun

The integration of machine learning (ML) models has ignited a paradigm shift in seismic analysis, fostering enhanced efficiency in capturing patterns of seismic activity with reduced need for time-consuming user interaction. Here, we investigate automated event detection and extraction of seismic phases using two widely used ML models, EQTransformer and PhaseNet. We applied both models to four weeks of continuous recordings of aftershocks using a temporary array following the November 30, 2022 ML 5.6 earthquake near Peace River, Alberta, Canada. Both tools identified >1000 events over the recording period. The aftershocks are located in close proximity and depth to the ML 5.6 mainshock on November 30, 2023, as well as to disposal operations that were ongoing at the time. Although there are some differences in the temporal and spatial evolution of the detected events by each model, both sets of detections reveal similar patterns of the aftershock distribution that were not identified by the regional network. Our results highlight the advantages of using ML models for rapid detection and assessment of seismicity following felt events, which is important for assessing hazard potential and risk in near real time.