Early moment detection that marks the start of massive tremor events

[jim mcnamara]

Popular science: https://phys.org/news/2019-05-scientists-early-moment-megaquake.html
https://advances.sciencemag.org/content/5/5/eaav2032
Image courtesy University of Oregon.

15 seconds into a massive event, there is data to support the concept of a warning of a coming massive event. Data is based on GPS ground-level readings for more than 20 years. The red X marks the moment.

 

[256bits]

15 seconds lead time.
How much difference would that make to sound the alarm?

 

[Nik_2213]

"15 seconds lead time."
That's back-to-front, 256 ? If this work is validated, it means that you can spot a 'Great Quake' while its still barely a 'temblor'. The difference between a super-cell warning and sirens...

 

[berkeman]

15 seconds lead time.
How much difference would that make to sound the alarm?

It can help a lot of things. That's why earthquake early warning systems and phone apps are starting to be used in the SF Bay Area:

https://sanfrancisco.cbslocal.com/tag/earthquake-warning-system/
Some of the features that are in the works when an early warning alert is sent out:

• BART and other commuter trains are stopped (or slowed as much as possible before the quake hits them)
• Elevators go to the nearest floor and open their doors
• Operating rooms are alerted so they can back scalpels and other instruments out of the patient before the hard shaking starts
• Airplanes on final approach can be told to abort
• (other creative actions are taken -- to be determined)...

Of course, part of an effective early warning system is to avoid false alarms. So, many detector sites must all indicate that an earthquake over some threshold magnitude is happening, before the alert can be sent out. It's a hard tradeoff between providing more seconds of warning and being absolutely sure that it is a real earthquake (and not just a heavy semi truck rolling past one of the sensors)...

 

[phinds]

15 seconds lead time.
How much difference would that make to sound the alarm?

You are looking at it backwards. It does not say that you get 15 seconds warning, it says you get a warning 15 seconds after the first measureable onset. How much longer it might be to the full quake is not clear since the graph is unfortunately cut off at 30 seconds. It could be less than 15 seconds or it could be more.

 

[256bits]

Thanks guys
I missed/misinterpreted the "15 seconds INTO the massive event"

As per @berkeman seconds can make a big difference - informative explanation how seconds count.

 

[davenn]

If this work is validated, it means that you can spot a 'Great Quake' while its still barely a 'temblor'.

hmmm, can't quite see how that comment makes sense ??

 

[davenn]

You are looking at it backwards. It does not say that you get 15 seconds warning, it says you get a warning 15 seconds after the first measureable onset. How much longer it might be to the full quake is not clear since the graph is unfortunately cut off at 30 seconds. It could be less than 15 seconds or it could be more.

yup

Thanks guys
I missed/misinterpreted the "15 seconds INTO the massive event"

As per @berkeman seconds can make a big difference - informative explanation how seconds count.

no probs,
but 15 sec is irrelevant if you are within 100km (60 miles) of the event as you have already been hit

EDIT
Actually, I will take that back ... as the radio link signals from the GPS sites travel faster than the
seismic waves and even after the delays due to signal processing, there's still a good chance
that you would get 10 - 12 sec of warning D

 

[pinball1970]

hmmm, can't quite see how that comment makes sense ??

Tremor?

 

[pinball1970]

yup

no probs,
but 15 sec is irrelevant if you are within 100km (60 miles) of the event as you have already been hit

EDIT
Actually, I will take that back ... as the radio link signals from the GPS sites travel faster than the
seismic waves and even after the delays due to signal processing, there's still a good chance
that you would get 10 - 12 sec of warningD

I think I could do a lot in 10 seconds in terms of getting to a secure part of a building or getting away from a potential hazard. Also those at risk from tsunami could have 10 extra seconds to think about higher ground.

 

[roj2003]

The analytical paper is here - https://advances.sciencemag.org/content/5/5/eaav2032
It seems the authors still have uncertainties in available monitoring for off-shore slips but on-shore is 'good'.

 

[mfb]

I‘m not sure how much this particular study will help the earthquake warning systems. If you are close to a magnitude 7 or 9 quake then your actions won’t depend on that difference - you protect yourself as good as you can anyway. If you are far away then the warning systems can use the information from the full event.

These warning systems are very useful. Japan has a very advanced one. It gives you a magnitude estimate and then a countdown until it will hit your location. I‘m not sure what it does if the time is too short to give both.

 

[berkeman]

It gives you a magnitude estimate and then a countdown until it will hit your location.

Yeah, that has to be a bit unnerving to see the time counting down just a few seconds until the big shaking starts where you are...

https://i.kinja-img.com/gawker-media/image/upload/t_original/772947448289464239.jpg

 

[mfb]

Much better than no estimate. If you have 5 seconds you duck under whatever is next to you, if you have 30 seconds you have time to go to a safer place.

 

[pinball1970]

The analytical paper is here - https://advances.sciencemag.org/content/5/5/eaav2032
It seems the authors still have uncertainties in available monitoring for off-shore slips but on-shore is 'good'.

So not much use for potential tsunamis in risk areas?