[glumreaper]

There's an important difference between prediction and detection. If you're in a nontrivial earthquake, detection isn't helpful: You already know it's happening.

The challenge is prediction.

There are a couple of startups that appear to be operating in this area with similar(ish) concepts to those described in the link.

precursor-spc.com - space based monitoring of atmospheric conditions, providing space weather forecasts and, apparently, earthquake prediction. Ionoterra - ground based monitoring of the ionosphere

I'm skeptical. If these companies were truly on to something then they would be selling their services to all the governments and would be household names.

[bobthepanda]

Detection is not helpful for those at the epicenter but buys seconds for those farther away, which can significantly reduce loss of life.

As a example, the detection system in Japan is linked to the Shinkansen control system, and trains will start emergency braking immediately. It reduces speed enough that there have been no earthquake fatalities resulting from Shinkansen derailments.

[chakintosh]

The recent earthquake in Morocco traveled all the way to Portugal from the epicenter. I was fortunate to drive through the epicenter area earlier that day because the plan was to stay in a lodge not far from the epicenter, but my kid was acting all fussy in the backseat so I drove straight to Casablanca. And that night, when the earthquake hit us, it felt like the whole house was sitting on top of a fright train, everything was shaking violently and I could have used an early warning to get my child and wife outside in time because had the earthquake been any stronger, I don't think anyone in my building would have had time to evacuate.

When you see the timestamp on the CCTV cameras, you could see that close to the epicenter, the earthquake hit around 23:11 in Marrakech, and in Casablanca, it reached us at around 23:13, I think a 2 minutes headsup is enough time for people get up and haul it outside.