Calibrating the SNSN seismometer "ONA" against the SCG:

cal-zacc10Hz-SCG

 USAGE:
              cal-zacc10Hz-SCG [options] [YYMMDD [HH]]

         on Holt to be run from cd /home/hgs/Seismo/gcf
         with path (tcsh):
         set path = ( . /home/hgs/Seism/sac/bin /home/hgs/Seism/bin /home/hgs/bin \
         /home/hgs/perlproc /opt/absoft11.5/bin /usr/local/GMT/bin \
         /usr/local/sbin /usr/local/bin /usr/sbin /usr/bin /sbin /bin )
 
 PURPOSE:

         Calibrate the Guralp seismometer against the SCG
         from three hours of 10 Sps ("Hz") vertical acceleration data ("ZACC").
         YYMMDD and HH are the start date and hour.
         (Specify date in short and tight form!) Default is
         midnight two days before today.
        
         More thorough analysis compared with cal-seis
 
         Advantageous selection of date and time:
         If there is a teleseismic event with surface waves of
         roughly the same order of magnitude as the microseismic noise.
         This will turn out a balanced wide-band average of the scale
         factor. How flat is the Guralp's gain spectrum? Inspect the plot!
         Check out microseismics in
            http://barre.oso.chalmers.se/hgs/SCG/memspectro.html ,
         our short notes from USGS
            http://barre.oso.chalmers.se/hgs/SCG/ens-eqtimes-daily-upd.html ,
         and daily residuals
            http://barre.oso.chalmers.se/hgs/SCG/daily-residual-plots.html
         all from our SCG.

         In order to analyse once the data processing is complete,
         issue
 
           fgrep 'Norm.CHI^2' tmp/tmp?*.fit
         e.g => find a minimum in segment 01
 
           fgrep '(admit)' tmp/tmp-01.fit  
         e.g. => admit  = 1.0653   => next time,
           calc "27.504 * 1.0653"  = 29.300
         => gcf2ts-hourly -cal 29.300 would be better
 
         Actually, the later version of the script creates a quite useful report,
         including chi2, admit, a residual slope, and the scale factor w.r.t. to the
         measurement unit of the Guralp's ADC "counts".
         Read the report:
           cat logs/cal-zacc10Hz-SCG-YYMMDD-HH-lag0.rep
         and find e.g. the highest x-correlation at lag -001
         => rerun xcorr-zacc10Hz-SCG -lag -1 ...
         And it iterates determination of a residual slope and the a-priori error value.

         You may have use of the SCG's short-term power spectrum exhibiting the
         sensors eigenresonance (excitation due to replacing the coldhead):
           splist -n/0300 -f0 -q -pdB -Hz ~/Ttide/SCG/o/G1_b_180628-1s.memsp MEM

         The ascii-file ./gcf+sg/cal-zaccf0-lagl-YYMMDD-HH.tsf contains 10,700 lines
>
  0.66666667   0.219979E+01  0.173556E+01  0.672771E+00
  0.68333333   0.257224E+01  0.256459E+01  0.251502E+00
  0.70000000  -0.609250E+00 -0.586430E+00 -0.805768E-01
...
    Minute     Seismo-ZACC       SCG           Diff
        

 OPTIONS:
 
 -GO          - Don't ask - no interaction except with fatal issues.
 
 -NEW         - Process make-gcf-like-sg
                Default is to use the existing files in likeg/
 
 -PLOT        - After processing, do a separate run for the same date and time to only plot.
                The plots that are produced (inspect the posted png's):
                1. Cross-spectrum gain and phase
                   http://barre.oso.chalmers.se/hgs/4me/calzacc/cal-zacc10Hz-SCG-YYMMDD-HH.png
                2. Power spectrum SCG and Seismo-Z acceleration (GGP-filtered)
                   http://barre.oso.chalmers.se/hgs/4me/calzacc/cal-zacc10Hz-SCG-YYMMDD-HH-psp.png
                3. Time series
                   http://barre.oso.chalmers.se/hgs/4me/calzacc/cal-zacc10Hz-SCG-YYMMDD-HH-ts.png

 -P           - Processing and a concluding plot stage
 
 -lag #l      - Apply a shift l (lag < 0, lead > 0 ) to the 10 Sps data.
                The unit for l is 0.1 s
                If the lag is new, the -NEW option must be given.
                In order to find out the lag, run this script without
                the -lag option and look at the files from segmented processing
                   gcf+sg/ZACC10Hz-SG-??????-s??.xcorr
                and in particular at the full-stroke result
                   gcf+sg/ZACC10Hz-SG-??????-??-all.xcorr
                Use the lag (first column) with the highest cross-
                correlation (you don't need to specify the leading `+´).
                The report will contain an extracted list; however,
                spurious large x-corr coefficients at large lags
                may occur. A majority vote can be used.
 
 -bf #f0      - Bessel filter scaling parameter f0; internal default
                in make-gcf-like-sg with resamp.tse,LIKEG is 0.0.174533
                If this option is used, the resulting files will be
                marked -f0
 
 -err #err    - A priori error [1.0]
 
 -R {<rep>|-} - Write a short report to file rep ,
                `-´ defaults to logs/cal-zacc10Hz-SCG-YYMMDD-HH-lagl.rep
                If argument ends with a `/´, write report
                to rep/cal-zacc10Hz-SCG-YYMMDD-HH-lagl.rep
 
 -r           - Test existence of files and exit
 
 EXAMPLE:
              cal-zacc10Hz-SCG -NEW -lag 1 -GO -R - 210526 21
            gets the gcf-data from Elder (if it exists) and analyses
            without asking.
 
              cal-zacc10Hz-SCG -GO -P -err 1000.0 -lag 1 -R - 210521 18 |\
                tee tmp/cal-zacc10Hz-SCG.log
            shows that the SCG runs into over-saturation in the
            surface waves from the Earthquake at Southern Qinghai, China
            2021-05-21 18 04 14 7.4 us-7000e54r 34.586 N 98.255 E
 
 AFTERMATH:
            Directly after a run for YYMMDD HH ,
              extract-xcorr gcf+sg/ZACC10Hz[f0]-SG-YYMMDD-s??.xcorr
            (lag and start hour are not taken into the file names from segmenting)
            can be piped into xyz2grd to visually inspect the stability
            of the lag of the central maximum of cross-correlation.

            The cross-spectrum can be further processed:
            For gain:
               splist -q -f0 -Hz -A  gcf+sg/sg-zacc-1s.csp CAD ADC >! text-file-gain
            For phase:
               splist -q -f0 -Hz -Pd gcf+sg/sg-zacc-1s.csp CPH APH >! text-file-phase
            You can try and find additional poles and zeros to that flattens the
            Guralp's response spectrum.
 
FILES PRODUCED:
            Example for -lag 0 210512 00 (no -bf):
./plot/cal-zacc10Hz-SCG-210512-00-xcorr.ps
./plot/cal-zacc10Hz-SCG-210512-00-ts.ps
./plot/cal-zacc10Hz-SCG-210512-00-psp.ps
./plot/cal-zacc10Hz-SCG-210512-00.ps
./logs/cal-zacc10Hz-SCG-210512-00-lag0.rep
./gcf+sg/cal-zacc-lag0-210512-00.env
./gcf+sg/ZACC10Hz-SG-210512-s10.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s09.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s08.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s07.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s06.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s05.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s04.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s03.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s02.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s01.xcorr
./gcf+sg/ZACC10Hz-SG-210512-s00.xcorr
./gcf+sg/ZACC10Hz-SG-210512-00-all.xcorr
./gcf+sg/G1_sb_210512-3h-10Hz.mc
             beside fetched gcf-data in ./3u93z2/

             After plotting:
           ./gcf+sg/sg-zacc-1s.csp
           ./gcf+sg/cal-zacc-lag0-210512-00.tsf
           /home/hgs/www/4me/calzacc/cal-zacc10Hz-SCG-210512-00-xcorr.png
           /home/hgs/www/4me/calzacc/cal-zacc10Hz-SCG-210512-00-ts.png
           /home/hgs/www/4me/calzacc/cal-zacc10Hz-SCG-210512-00-psp.png
           /home/hgs/www/4me/calzacc/cal-zacc10Hz-SCG-210512-00.png

.bye