Eathquake@home ????

Message boards : BOINC client : Eathquake@home ????
Message board moderation

To post messages, you must log in.

AuthorMessage
Knightrider

Send message
Joined: 29 Aug 05
Posts: 9
United Kingdom
Message 683 - Posted: 8 Oct 2005, 18:45:23 UTC

Hi,

Don't know if this is the right place to ask this, but does anyone know if a BOINC project is/can/could/would be developed to predict earthquakes??

K.
ID: 683 · Report as offensive
Paul D. Buck

Send message
Joined: 29 Aug 05
Posts: 225
Message 684 - Posted: 8 Oct 2005, 19:33:19 UTC

Yes
ID: 684 · Report as offensive
fzb

Send message
Joined: 1 Oct 05
Posts: 2
Message 685 - Posted: 9 Oct 2005, 10:01:59 UTC

given the time a wu needs to return and comparing it with the time you have before a earthquake wents off, i am not sure if that would help. "simulating" the earthqwake though would work with boinc. boinc as a warning system would be to little "real time" though i think. (despite i am not aware of a precise prediction method for earthquakes, but that doesn't mean there cannot be one ;) )
ID: 685 · Report as offensive
Paul D. Buck

Send message
Joined: 29 Aug 05
Posts: 225
Message 686 - Posted: 9 Oct 2005, 13:35:20 UTC

There is some expirimental prediction systems that look at the mini-quakes and what they are doing. So, that processing would be a candidate as it is working months to weeks in advance.
ID: 686 · Report as offensive
Bill Michael

Send message
Joined: 30 Aug 05
Posts: 297
Message 687 - Posted: 9 Oct 2005, 14:00:51 UTC

I would also think that a large part of what is needed is to test various methods of prediction. That can only be done "after the fact". Feed in all the data that was collected _before_ some given earthquake, and run the prediction code with all the different variables - see what sets of variables "predict" the earthquake that DID occur, the closest. Similar to the CPDN approach - first you see how well your code matches the historical data before you try to predict the future. If you start with the weather in 1900 and your model says we were in an ice age by 1950... well, you got it wrong. For an earthquake, if your model predicts "nothing", or it predicts the earthquake but a year too early or too late, or the magnitude is off by a lot, then you throw out that model and try another...

Once you have a proven, working model, if such is ever possible, then it's time to be running it on some supercomputer cluster in real time, not relying on volunteers. But volunteers can help define and refine the model.

ID: 687 · Report as offensive

Message boards : BOINC client : Eathquake@home ????

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.