| Date: | 1999-08-24 16:26:00 | |
| Sender: | Graham Appleby <[Mailed W. Seemueller <seemuell@dgfi.badw-muenchen.de>]> | |
| Subject: | [SLR-Mail] No. 410: message to Eurolas stations | |
| Author: | Graham Appleby | |
| Content: | ******************************************************************************** SLR Electronic Mail 1999-08-24 16:26:00 UTC Message No. 410 ******************************************************************************** Author: Graham Appleby Subject: message to Eurolas stations Dear Colleagues Van Husson has found that normal point data computed by some stations that are using the ´RGO´ software ´NORMPT´ has an error in the normal point RMS field. The program originally calculated the standard error of the mean, but the NP format requires that the RMS be computed. Thus the value as computed by some stations is currently wrong by a factor of root(n). Please see the corrected code below, and make the changes to your version of the software if required. Regards, Graham Appleby ---------------------------- Coding in program ´normpt´ to compute the RMS value of each normal point. Please check that your version of the program computes ´rms´ as shown here. An earlier version of the program computed the ´standard error´ of the normal point from rms/root(n) c c RMS in ps of normal point from fit residuals: c if only 1 raw data point in bin, use overall pass RMS. c sum=0.0 if(n.eq.1)then rms=prms*2000.0 else do 450 k=1,n sum=sum+(resr(k)-resmn)**2 450 continue rms=sqrt(sum/float(n)) endif c c the above value of rms is calculated as given in the c algorithm for the standard error of a single observation c contributing to a normal point. --------------------------------------------------------------- ------------------- . . - ------------- /------------------------- Wolfgang Seemueller ( ) ( ) ***_ ^ -o- Tel. +49 89 23031-109 DGFI, Abt.I | | | | *|* o|_ / / / Fax: +49 89 23031-240 Marstallplatz 8 |o| |o| |o |_) / / email: D-80539 Muenchen | |/| | |___| / seemueller@dgfi.badw-muenchen.de ----------------------------------------------------------------------------- [Mailed From: ”W. Seemueller” ******************************************************************************** |