NGA Attenuation

[moe333]

Hi all,

I am trying to estimate the +1 std. deviation of PGA for a M6.9, Vs=800 m/s (estimated), reverse fault (dip=75 degrees) at Rrup=2km (Rjb=0). The site is on the hanging wall of the fault. I am coming up with PGA=1.46g for these parameters using the new NGA attenuation relationship from Abrahamson & Silva, and this is about the same value as I have calculated using their 1997 relationship. These values are significantly higher than other NGA authors predictions.

One reason I am asking is the NGA relationships were supposedly predciting lower ground motion than the 1997 realtionships. If my calculations are correct, I'm wondering if this may be due to larger uncertainty for the +1 std. deviation value than with previous relationships?

If anyone has this relationship programmed I'd be interested to know what value they come up with for these parameters.

Thanks

 

[muuddfun]

Moe

OSHPOD, was still not in full aceptance of the Abrahamson and Silva NGA, see uploaded file. Though I think mostly because it was low in most curcumstances. But it depends on the period on how much the NGA's have changed from before. I am not sure about how much the standard deviation is affecting it, but that may be part of what you are seeing, especially with a reverse fault. What fault are you looking at.

 

[moe333]

muddfun,

It appears OSHPOD is not in full acceptance because it was not used in the USGS hazard mapping. I'm not sure why the the USGS didn't use it, maybe because of low values at some periods as you say. However, it's not low for my case, it's much higher than the others.

FYI, I have gotten some feedback that indicates the numbers I calculated are correct.

Thanks for the info.

 

[howardoark]

It's my understanding that the NGA's are lower for strike slip faults in the near fault zone. I don't think they should be different for thrust faults. I forget the reasoning behind that, but it's got to be based on observed accelerations for recent earthquakes.

The various relationships have different definitions of distance to the fault, so that might be the issue.

DSA makes you adjust the NGAs if you want to use them.

We generally use at least three attenuation relations and then average them.