Aug. 30 () –
A new detection method, based on machine learning, allows to identify low-level tectonic upheaval over large areas prior to a major earthquake even months in advance.
“Our article demonstrates that the advanced statistical techniques, Machine learning in particular has the potential to identify precursors to large-magnitude earthquakes by analyzing data sets derived from earthquake catalogs,” he explains. in a statement Professor Társilo Girona, Research Assistant Professor of the UAF Geophysical Institute (University of Alaska Fairbanks).
Girona, a geophysicist and data scientist, studies the precursor activity of volcanic eruptions and earthquakes. Geologist Kyriaki Drymoni of Ludwig-Maximilians-Universität in Munich was a co-author in developing the new detection method, published in Nature Communications.
The authors wrote a computer algorithm to search the data for abnormal seismic activity. Algorithms are a set of computer instructions that teach a program to interpret data, learn from it, and make informed predictions or decisions.
They focused on two major earthquakes: the 2018 magnitude 7.1 Anchorage earthquake and the 2019 Ridgecrest, California, earthquake sequence of magnitudes 6.4–7.1.
They found that approximately three months of abnormal low-magnitude regional seismicity had occurred in about 15% to 25% of south-central Alaska and southern California before each of the two earthquakes studied.
Their research finds that the disruption preceding large earthquakes is mainly reflected in seismic activity with a magnitude less than 1.5.
The Anchorage earthquake occurred on November 30, 2018 at 8:29 a.m., with an epicenter located approximately 16 kilometers north of the city. It caused extensive damage to some roads and highways, and several buildings were damaged.
Using their data-trained program, Girona and Drymoni found that for the Anchorage earthquake, the probability of a major earthquake occurring in 30 days or less rose sharply to about 80% about three months before the November 30 earthquake. The probability then rose to about 85% just a few days before it occurred. They obtained similar probability results for the Ridgecrest earthquake sequence. during a period beginning about 40 days before the start of the earthquake sequence.
Girona and Drymoni propose a geological cause for the low-magnitude precursor activity: a significant increase in pore fluid pressure within a fault.
Pore fluid pressure refers to the pressure of the fluid within a rock. High pore fluid pressures can potentially trigger fault slip if the pressure is too high. It is enough to overcome the frictional resistance between the rock blocks on either side of the fault.
“Increased pore fluid pressure on faults triggered by major earthquakes changes the mechanical properties of the faults, which in turn leads to uneven variations in the regional stress field,” Drymoni said. “We propose that these uneven variations control the anomalous and precursory low-magnitude seismicity“.
Machine learning is having a huge positive impact on earthquake research, Girona said.
“Modern seismic networks produce huge data sets that, when properly analyzed, can offer valuable insights into the precursors to seismic events,” he said. “This is where advances in machine learning and high-performance computing can play a transformative role, allowing researchers to identify significant patterns that could indicate an impending earthquake.”
The authors say their algorithm will be tested in near-real-time situations to identify and address potential challenges to earthquake prediction. The method should not be used in new regions without training the algorithm on the historical seismicity of that area, they add.
Producing reliable earthquake forecasts has a “a deeply important and often controversial dimension,” said Girona.
“Accurate forecasting has the potential to save lives and reduce economic losses by providing early warnings that enable timely evacuations and preparedness,” he said. “However, the uncertainty inherent in earthquake prediction also raises important ethical and practical questions.”
“False alarms can lead to unnecessary panic, economic disruption and loss of public confidence, while wrong predictions can have catastrophic consequences“, he said.
Add Comment