Researchers in the United States say they have taught artificial intelligence to predict where aftershocks will strike following large earthquakes.
Using data from more than 131,000 pairs of earthquakes and aftershocks, the researchers trained a neural network to develop a better understanding of where earthquakes induce stress.
The network then identified a pattern of aftershock locations.
"The application of machine learning to high-quality earthquake datasets is a big step beyond what has been done in the past," said Professor Mark Stirling, the chair of earthquake science at the University of Otago.
"With evolving methods like this, we stand to gain a better understanding of how this method can contribute to the ensemble of existing earthquake forecasting methods."
- Man's huge quake prediction is 'fearmongering' - GeoNet
- GeoNet's earthquake predictions for New Zealand
- British clairvoyant predicts 'serious earthquak' in NZ
The current prediction method is the Coulomb failure stress change, a model which only looks at one kind of stress, but the researchers say their new model will be more accurate as it identifies several different types of stress.
The research was published in journal Nature.
In an accompanying editorial, Stanford physicist Gregory Beroza said while the method was promising, it's still early days.
"The performance [the] artificial neural network is motivating. Until a few years ago, most statistical forecasts of aftershocks were more accurate than were physics-based forecasts, such as that of the authors.
"But there are now cases in which physics-based forecasting performs as well as purely statistical approaches. The time would seem ripe for methods based on artificial intelligence to enter the fray, and [their work] has established this beachhead."