Faint but significant magnetic field changes have been detected days before Californian earthquakes. It’s not quite the holy grail of seismology, but it could illuminate the path – if the work stands up to further testing, it could allow warnings that could greatly reduce lives lost. The researchers who found the signal acknowledge it’s not currently clear enough to enable these warnings, but that could change.
Hurricanes are becoming more destructive, and far more people are living in their paths, yet deaths are falling since weather satellites transformed our capacity to predict them – might we achieve the same thing with earthquakes? Seismologists have put a lot of effort into trying, but so far with little success.
Dan Schneider of QuakeFinder thinks that may be about to change, with Schneider and co-authors reporting on observations of magnetic fields at 125 sensor stations buried along California’s major fault lines.
Between 2005 and 2019, the area experienced 18 earthquakes of magnitude 4.5 or larger. The authors looked for changes in the lead-up to each. It’s not an easy task, because so many other factors can affect the sensors – solar storms are an important example, but even rush hour traffic makes a difference – and these all need to be controlled for.
Although the authors report distinctive patterns, they stress these are not sufficient to justify evacuations. “The statistical signal is of modest size, which means that we can not directly provide a prediction that can be used to alert the public,” the paper states. “This study provides evidence that there is a physical change that can be observed in the days before an earthquake, but further scientific study is needed to understand this process.”
"It's a modest signal," Schneider acknowledged to Eos. "We are not claiming that this signal exists before every earthquake, but it is very intriguing."
Tales of animals (or even people) who could detect earthquakes before they happened are thousands of years old. Scientists have attempted to study this rigorously, but interpreting a subject’s behavior is an obstacle if they’re not able to talk. However, if these reports are true, there must be some physical change that is being detected. Magnetic fields are a plausible explanation, leading to decades of study and the installation of the sensors used here.
However, this is a classic example of the challenge of differentiating signal from noise, made more difficult by the (fortunate) rarity of major quakes to study.
Even if the signal is real, plenty of questions that affect how useful it might be remain unanswered. For example: do other fault lines show the same pattern, or at least a different one we can identify, before earthquakes? We also need to know if the signals prior to the most devastating earthquakes can be distinguished from magnitude 5 events in order to avoid crying wolf.
To produce useful predictions, the authors intend to continue expanding their sample size and addressing more confounding factors. However, they also note we currently don’t understand the mechanism associating magnetic signals with subsequent events. If the physical connection can be explained, it may open the way to better interpretation of signals in the future.
The article is open access at Journal of Geophysical Research: Solid Earth