AI Takes On Earthquake Prediction (quantamagazine.org) 13
After successfully predicting laboratory earthquakes, a team of geophysicists has applied a machine learning algorithm to quakes in the Pacific Northwest. From a report: In May of last year, after a 13-month slumber, the ground beneath Washington's Puget Sound rumbled to life. The quake began more than 20 miles below the Olympic mountains and, over the course of a few weeks, drifted northwest, reaching Canada's Vancouver Island. It then briefly reversed course, migrating back across the U.S. border before going silent again. All told, the monthlong earthquake likely released enough energy to register as a magnitude 6. By the time it was done, the southern tip of Vancouver Island had been thrust a centimeter or so closer to the Pacific Ocean. Because the quake was so spread out in time and space, however, it's likely that no one felt it. These kinds of phantom earthquakes, which occur deeper underground than conventional, fast earthquakes, are known as "slow slips." They occur roughly once a year in the Pacific Northwest, along a stretch of fault where the Juan de Fuca plate is slowly wedging itself beneath the North American plate.
More than a dozen slow slips have been detected by the region's sprawling network of seismic stations since 2003. And for the past year and a half, these events have been the focus of a new effort at earthquake prediction by the geophysicist Paul Johnson. Johnson's team is among a handful of groups that are using machine learning to try to demystify earthquake physics and tease out the warning signs of impending quakes. Two years ago, using pattern-finding algorithms similar to those behind recent advances in image and speech recognition and other forms of artificial intelligence, he and his collaborators successfully predicted temblors in a model laboratory system -- a feat that has since been duplicated by researchers in Europe. Now, in a paper posted this week on the scientific preprint site arxiv.org, Johnson and his team report that they've tested their algorithm on slow slip quakes in the Pacific Northwest. The paper has yet to undergo peer review, but outside experts say the results are tantalizing. According to Johnson, they indicate that the algorithm can predict the start of a slow slip earthquake to "within a few days -- and possibly better."
More than a dozen slow slips have been detected by the region's sprawling network of seismic stations since 2003. And for the past year and a half, these events have been the focus of a new effort at earthquake prediction by the geophysicist Paul Johnson. Johnson's team is among a handful of groups that are using machine learning to try to demystify earthquake physics and tease out the warning signs of impending quakes. Two years ago, using pattern-finding algorithms similar to those behind recent advances in image and speech recognition and other forms of artificial intelligence, he and his collaborators successfully predicted temblors in a model laboratory system -- a feat that has since been duplicated by researchers in Europe. Now, in a paper posted this week on the scientific preprint site arxiv.org, Johnson and his team report that they've tested their algorithm on slow slip quakes in the Pacific Northwest. The paper has yet to undergo peer review, but outside experts say the results are tantalizing. According to Johnson, they indicate that the algorithm can predict the start of a slow slip earthquake to "within a few days -- and possibly better."
Re:Meh... (Score:5, Informative)
I've heard that the Big One would send California off into the Pacific in the next 30 years.
You were misinformed. San Andreas is a transform fault, so it doesn't "open up", it slides laterally. The coastline is moving north. The inland areas are heading south. Both at 20-30mm per year.
There is no chance of California drifting into Pacific to be become another Hawaii.
Re:Meh... (Score:4, Funny)
Thanks. You just totally ruined Superman (1978) for me.
Re: (Score:2)
Re: (Score:2)
But Luthor had one nuclear warhead. Continental-scale geological forces would be no match.
Re: Meh... (Score:1)
AI and AGW (Score:2)
Re: (Score:2)
If you cast your mind back to the climate modelling that was the second big Distributed Internet Computing Project (after SETI@home [wikipedia.org]), you'll remember that stage one of the project [wikipedia.org] was to find which ranges of "free parameters" produced credible models. They did this by assigning users sets of
Re: (Score:2)
The kind of analysis is stupid, all it does is make the AI able to forecast the kind of quake that was and has very little to do with what will be. Each and every quake is a unqiue event, subject to the strength of the meeting rock faces and a which point the surfaces will break down, one or the other or both and allow slippage to occur, that slippage defined by the strengths of nearbye rock faces associated with the fault and their ability to take up the shifted loads, depending upon how much stress they a
Re: (Score:2)
Yes, of course. No geologist has ever thought differently. I don't think the geologists doing this work were looking for anything other than a process for classifying "slow slip" earthquakes in the early months of their movement. They know that both (1) all earthq
Re: (Score:1)
Has anyone ever run all the available climate data through a ML algorithm to create a model of past or present? AI for earthquakes is cute but we don't have enough information to predict them but my guess is we do have enough climate data to get a reasonable output. Any links or info on this concept?
It will not work.
1. You do not want ALL climate data. Some datasets like tree rings, ocean temperature, etc are extremely good. Measured data at actual met stations is utter shite. It is "polished by hand" in violation of all rules and laws of science to compensate for other antropogenic factors. Exhibit A - the data point for Sofia, Bulgaria in the most commonly used IPCC data set. It was an open field in the 50es when the met moved to 8th kilometer site, in the 60-es the physicists next door built a re