The team plans to test their algorithm in real-time scenarios to refine its accuracy. However, they caution against applying this method in new regions without training it with local seismic data first. (Representative Image: Canva)

Imagine knowing about a major earthquake weeks or even months before it strikes. Scientists at Ludwig Maximilian University of Munich (LMU Munich) and the University of Alaska Fairbanks (UAF) say that this might become a reality thanks to a new AI model.

The study, led by Prof. Társilo Girona of the UAF Geophysical Institute and Dr. Kyriaki Drymoni from LMU Munich, explores how AI could revolutionise earthquake prediction. Their method, detailed in Nature Communications, uses machine learning to identify seismic unrest over large areas.

Algorithm Detects Seismic Signs

Girona and Drymoni developed an algorithm designed to detect abnormal seismic activity. They applied this algorithm to analyse two significant earthquakes: the Ridgecrest earthquakes of 2019, which had magnitudes ranging from 6.4 to 7.1, and the Anchorage earthquake of 2018, which measured 7.1.

They discovered unusual increases in low-magnitude seismic activity three months before these major quakes. This activity covered 15% to 25% of the affected regions. Their findings suggest that low-magnitude tremors often precede major earthquakes.Pore Fluid Pressure Insights

The researchers believe increased pore fluid pressure within faults might explain these precursors. When fluids exert enough pressure, they can trigger fault slips. Drymoni explains that these pressure changes affect the faults' mechanical properties, leading to unusual seismic activity.

Machine Learning’s Role in Earthquake Research

Girona emphasises the transformative role of machine learning in seismic research. Modern seismic networks generate vast amounts of data. Proper analysis of this data could reveal patterns indicating impending earthquakes.

The team plans to test their algorithm in real-time scenarios to refine its accuracy. However, they caution against applying this method in new regions without training it with local seismic data first.

Challenges and Ethical Considerations

Girona warns that while accurate forecasting can save lives and reduce losses, it also brings challenges. False alarms may cause panic and economic disruption, while missed predictions could have dire consequences. Balancing these risks with potential benefits remains crucial.

This innovative approach marks a significant step forward in earthquake forecasting, potentially transforming how we prepare for and respond to seismic events.