[Shake-dev] Re: [Fwd: Re: Solution: regression distances and plotregr?]

[Bruce Worden]

Hi Pete,

I've attached a revised version of Small_Seg.pm, below.  It will  
complain and die if you try to use grind's -psa flag, but it should  
now handle distance (and uncertainty) correctly for PGA and PGV.

> This has the regression following the data more closely than for  
> deeper
> depths. What do seismologists think of this?

I think what is happening is that as you push the regression up to  
zero depth, it is matching the stations' (incorrect) concept of  
distance (Rjb, which puts the source at the surface) more closely.

> I think I am convinced that plotreg is plotting both data and  
> regression
> values against the Joiner-Boore depth. That is good because it  
> means the data
> and regressions are being compared on an "equal basis", apples to  
> apples. And
> I think that means that when grind does its bias calculation, it is  
> also
> comparing station data and regression values using the same depth  
> measure.

I don't think this is quite correct, because the shape of the curve  
changes with distance, and so putting things at the same (but wrong)  
distance may look more correct, but it will not match the way the  
regression was calibrated.  In addition, in the bias calculation, the  
amplitudes returned by the regression for the station locations will  
include the depth term, but the stations will be at JB distance,  
leading to a systematic (but non-constant) shift in the position of  
the curve.

Anyway, the new code is attached, and it should fix the problem.

Bruce

-------------- next part --------------
A non-text attachment was scrubbed...
Name: Small_Seg.pm
Type: text/x-perl-script
Size: 9417 bytes
Desc: not available
Url : http://geohazards.usgs.gov/pipermail/shake-dev/attachments/20070309/77a334d9/attachment.bin 
-------------- next part --------------