I am working the seismic hazard of Oman and would like to have the best ground motion prediction equation for the area. Oman is characterised by its low seismicity, although some moderate earthquakes were recorded. During my scouting, I found some GMPE, which seems reasonable for Oman regarding the tectonic environment, but derived mainly using weak motion data.
I need as conclusive answer as possible.
Conceptually, strong ground motion at a given site is dependent on three components: 1) source(s), 2) propagation, and 3) local site effects. In the absence of local strong ground motion recordings, one can either assume generic source conditions, or one could look to geologically/tectonically similar regions elsewhere in the world. I would suggest starting with the generic assumption.
For the propagation effects, weak motion data has been used to model the propagation of strong ground motions (an example would be studies looking as the difference between the western US and the eastern US. http://www.ce.memphis.edu/7137/PDFs/attenuations/USGS2008/Campbell2003.pdf http://www.ce.memphis.edu/7137/PDFs/Kappa/Atkinson_Boore_2006_Bssa.pdf ). If weak ground motion data is available for Oman, one could use it to come up with a region specific GMPE based on that data rather than just borrowing what someone else did in a similar tectonic environment.
Finally, the local site effects can be reasonably estimated using the average near-surface shear-wave velocity. Or by using correlations between average Vs and near-surface geologic materials and conditions.
The foregoing is a conceptual simplification of the details that can be found in the two papers linked to, and the papers that they cite. These two papers are from 2003 and 2006. There is a steady stream of literature being published regarding GMPE. My familiarity with these two older papers reflects my career taking a turn around that time.
All that said, the uncertainty/variance of strong ground motion predictions, even with observational measurements of previous strong ground shaking, is very large. Using existing, generic strong ground motion prediction equations is likely to get one as close to the truth of some future measured observation as might a very elaborate, region specific equation that was developed without any region specific strong ground motion records. Thus, using a generic GMPE is a good place to start.
Dear Prof. Matthew Mabey
Thanks alot for your kind answer and effective suggestion. I still have three points to be clarified.
1- The role of stress drop, which controls the high frequency part of the ground motion that contributes mostly for the damage, is not clear. Some studies suggest average stress drop across a broad moment magnitude range (self similarity), and other studies show varying stress drop with magnitude. If the later is true, the ground motion derived by weak motion cannot be used for predicting strong motions. Is there any conclusive study resolving this problem?
2- Similarly, Some studies show the dependance of the decay rate on the magnitude, but with no conclusive evidence for all the tectonic environment. I am looking for a definitive conclusion to accept or reject the ground motion prediction equation.
3- I do not know if the large earthquakes have the same rupture process as the small ones.
Professor Ahmed Deif ,
First, let me say that your three points show that you have a firm grasp of the uncertainties that exist in seismology because large earthquakes are relatively rare events and it has only been in recent decades (that is to say within my memory) that we have had high quality, broadband recordings of earthquakes. In the case of Oman, the frequency of large earthquakes is much lower than in places like, for example, Japan. Thus the uncertainties are in many respects even greater. Give the sparse instrumental record in Oman, I was trying to steer you toward using ground motion prediction methods based on global data sets.
I don't think that we yet have a definitive answer to your point #1. One relatively recent publication related to that topic is https://www.scec.org/publication/7345 One of the things I would highlight in this paper is that even for a given recorded event there isn't precise agreement regarding the stress drop. Thus, we should be very suspect of our ability to precisely predict the stress drop in some future earthquake in a region without modern instrumental recordings of large earthquakes.
I am hoping that your question attracts the attention of someone doing active research in the areas that the question deals with. I would like to read some different perspectives.
If one is doing a probabilistic hazard assessment, then one does not need to settle on a single answer to your questions, or any of the many other factors about which there is uncertainty. Instead, the various possibilities can be factored into the overall hazard assessment. e.g.Book Earthquake scenario and probabilistic ground shaking maps fo...
Even in areas with abundant seismicity, we are still refining our understanding of earthquakes and the ground motions they produce. The Tohoku earthquake in Japan is a fairly recent example of the consequences of experts underestimating the hazard because they overestimated our understanding of earthquakes.
We should never let the precision of our instruments overshadow the limitations of our understanding.