5 ms·

How far are we from decent prediction? Many years ago ago I looked quite deeply at stick-slip friction models, the mathematics, dynamics and signals. I'd wage

How far are we from decent prediction? Many years ago ago I looked quite deeply at stick-slip friction models, the mathematics, dynamics and signals. I'd wager, in the time since 2005, there's a lot we could do with machine learning, predictive signal analysis and good historical data. Even 5 minutes warning could save a lot of lives. Any ML + geo-scientists on the case?

daveaiello 4y ago
A seismologist from Berkeley said on KCBS that the ShakeAlert system worked in the case of this earthquake. So many people in the area were warned. I heard this on Audacy from the Philadelphia area.
- daveaiello 4y ago
  Info on how ShakeAlert works: https://twitter.com/USGS_ShakeAlert/status/1584983600378032128 https://twitter.com/USGS_ShakeAlert/status/15849836003780321...
- jofer 4y ago
  ShakeAlert and similar systems are awesome, but not exactly what most folks mean by a prediction, FWIW. They're meant for warnings of seconds, but not for things on the scale of minutes or hours or days. Not to take away from ShakeAlert in any way though! Just clarifying that in terms of another comment I made about earthquake precursor signals. ShakeAlert and similar aren't about precursors, but instead are about "the earthquake has happened, let's tell folks before the damaging surface waves reach them".
  - lazide 4y ago
    Still better than a sharp stick in the eye!
  - fragmede 4y ago
    I mean, warnings of seconds is the best we can do right now but the goal is prediction. it's not a lot but it's enough to stop surgeries and trains.
jofer 4y ago
I really want to write up a long response to this... I used to be closely adjacent to this field, and I do ML work in other domains. ML is mostly irrelevant for this, though. Here's the short version: We're good at predicting where. We're still not good at when. Understanding what segments of a fault zone have built up the most elastic strain is something we're really quite good at these days. Given historical earthquake data, GPS measurements of deformation, and knowledge of the geometry and makeup of the subsurface, we can reasonably accurately model what segments of a fault are most likely to fail next. No ML needed or wanted. The issue with the "when" is what the precursor signals are. The precursor signals we're most likely to be able to reliably measure involve drilling down and putting sensors on both sides of the fault zone and measuring the sonic velocity across the fault zone repeatedly. As rocks or even a fault gouge begin to fail, they have to dilate slightly so that grains can rotate into position to slip (bad explanation, sorry). This dilation causes a drop in sonic velocity. Stresses on the rock mass as a whole also have velocity signals. Lab experiments have observed this fairly reliably. We've actually set things up to measure this in the field as well -- look up SAFOD. There were some likely precursor signals (a bit contentious, but it's fair to say there were interesting signals in the data). However, this type of instrumentation just isn't practical to deploy at scale. Other possible precursor signals involve hydrologic changes (e.g. well waters rising / changes in pore pressure) or Radon spikes, both due to the same dilation. Again, though, these are local effects and difficult to deploy a system to measure. (e.g. you need a deep borehole for the hydrologic changes, and the radon spikes are most reliably observed at depth as well) There is one more controversial category that is easily observable: ionospheric perturbations. GPS data is sensitive to the ionosphere, and one of the things you need to do in geodetic GPS station operations is correct for changes in the ionosphere. We tend to have tons of geodetic GPS stations deployed along fault zones. Folks have noticed some very interesting ionospheric perturbations before very large earthquakes. There have been a lot of papers published on this, but it's still not too widely accepted, last I checked. The proposed mechanism (piezoelectric effect and/or dislocation glide/creep building up charge) and the observed magnitude of change don't match. It also has only been documented as a precursor around very major earthquakes (which are rare) and only in retrospect. At any rate, tons of data don't help if it's all noise. The data that we think should have the most signal is _really_ hard to get. Things are still in the "promising, but a long ways off" state when it comes to predicting "when", and seismologists are _very_ careful to avoid doing work that could be taken as predictions due to past PR SNAFUs.
Enginerrrd 4y ago
Well, I'm not a domain expert here, but I have followed this a bit over the years. Here's my understanding: 1. Prediction actually is getting quite good, just not on the time scales that we're interested in, which would probably be somewhere between weeks and minutes. 2. I don't think that at the moment we have access to input variables with sufficient power to be predictive on those time scales. Right now, I think one of the best predictors of large EQ's on short time scales is a sudden increase in variance of seismic activity. Which is sort of obvious when you think about it... The best predictor of a big slip is when it's already started slipping. 3. There have been good gains on minute time scale warnings simply by using sensors and communication networks that can outpace the speed of the seismic waves. This can actually be incredibly helpful for seismic sensitive systems!

daveaiello 4y ago

A seismologist from Berkeley said on KCBS that the ShakeAlert system worked in the case of this earthquake. So many people in the area were warned. I heard this on Audacy from the Philadelphia area.

daveaiello 4y ago
Info on how ShakeAlert works: https://twitter.com/USGS_ShakeAlert/status/1584983600378032128 https://twitter.com/USGS_ShakeAlert/status/15849836003780321...
jofer 4y ago
ShakeAlert and similar systems are awesome, but not exactly what most folks mean by a prediction, FWIW. They're meant for warnings of seconds, but not for things on the scale of minutes or hours or days. Not to take away from ShakeAlert in any way though! Just clarifying that in terms of another comment I made about earthquake precursor signals. ShakeAlert and similar aren't about precursors, but instead are about "the earthquake has happened, let's tell folks before the damaging surface waves reach them".
- lazide 4y ago
  Still better than a sharp stick in the eye!
- fragmede 4y ago
  I mean, warnings of seconds is the best we can do right now but the goal is prediction. it's not a lot but it's enough to stop surgeries and trains.

jofer 4y ago

I really want to write up a long response to this... I used to be closely adjacent to this field, and I do ML work in other domains. ML is mostly irrelevant for this, though. Here's the short version: We're good at predicting where. We're still not good at when. Understanding what segments of a fault zone have built up the most elastic strain is something we're really quite good at these days. Given historical earthquake data, GPS measurements of deformation, and knowledge of the geometry and makeup of the subsurface, we can reasonably accurately model what segments of a fault are most likely to fail next. No ML needed or wanted. The issue with the "when" is what the precursor signals are. The precursor signals we're most likely to be able to reliably measure involve drilling down and putting sensors on both sides of the fault zone and measuring the sonic velocity across the fault zone repeatedly. As rocks or even a fault gouge begin to fail, they have to dilate slightly so that grains can rotate into position to slip (bad explanation, sorry). This dilation causes a drop in sonic velocity. Stresses on the rock mass as a whole also have velocity signals. Lab experiments have observed this fairly reliably. We've actually set things up to measure this in the field as well -- look up SAFOD. There were some likely precursor signals (a bit contentious, but it's fair to say there were interesting signals in the data). However, this type of instrumentation just isn't practical to deploy at scale. Other possible precursor signals involve hydrologic changes (e.g. well waters rising / changes in pore pressure) or Radon spikes, both due to the same dilation. Again, though, these are local effects and difficult to deploy a system to measure. (e.g. you need a deep borehole for the hydrologic changes, and the radon spikes are most reliably observed at depth as well) There is one more controversial category that is easily observable: ionospheric perturbations. GPS data is sensitive to the ionosphere, and one of the things you need to do in geodetic GPS station operations is correct for changes in the ionosphere. We tend to have tons of geodetic GPS stations deployed along fault zones. Folks have noticed some very interesting ionospheric perturbations before very large earthquakes. There have been a lot of papers published on this, but it's still not too widely accepted, last I checked. The proposed mechanism (piezoelectric effect and/or dislocation glide/creep building up charge) and the observed magnitude of change don't match. It also has only been documented as a precursor around very major earthquakes (which are rare) and only in retrospect. At any rate, tons of data don't help if it's all noise. The data that we think should have the most signal is _really_ hard to get. Things are still in the "promising, but a long ways off" state when it comes to predicting "when", and seismologists are _very_ careful to avoid doing work that could be taken as predictions due to past PR SNAFUs.

Enginerrrd 4y ago

Well, I'm not a domain expert here, but I have followed this a bit over the years. Here's my understanding: 1. Prediction actually is getting quite good, just not on the time scales that we're interested in, which would probably be somewhere between weeks and minutes. 2. I don't think that at the moment we have access to input variables with sufficient power to be predictive on those time scales. Right now, I think one of the best predictors of large EQ's on short time scales is a sudden increase in variance of seismic activity. Which is sort of obvious when you think about it... The best predictor of a big slip is when it's already started slipping. 3. There have been good gains on minute time scale warnings simply by using sensors and communication networks that can outpace the speed of the seismic waves. This can actually be incredibly helpful for seismic sensitive systems!