Introduction

Seisan can be used to create response files in GSE format and waveforms in sac files. This response addresses the practical issue of how to convert the GSE response information to a format for use with sac or gsac

The documentation for the GSE2.1 format is found at
http://www.seismo.ethz.ch/autodrm/autodrm_doc.html
and then clicking on the documetation for the provisional_GSE2.1.pdf file.

Files

The files sent to me are the following:
20080529_1600z.sac which is the waveform file in Intel binary format
C214_B__Z.2007-01-01-0000_GSE which is the GSE response information
C214_B__Z.2007-01-01-0000_SEI which is some file created by seisan which I do not need or understand.

Instrument response

Let us look at the GSE response file:

CAL2 C214  B_Z      CMG-6    0.12E-01      1.   40.00000 2007/01/01 00:00                       
PAZ2  1 V  0.20000000E-05                 2   3 Laplace transform                               
 -0.14660765E+00  0.14956973E+00
 -0.14660765E+00 -0.14956973E+00
  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00
DIG2  2  0.66475240E+07    40.00000 CD24E1                                                      

According the GSE documentation, the CAL2 line (page 76) indicates that the system response is 0.12E-01 coutns/nanometer at a period of 1.0 seconds.

The PAZ2 entries indicate that the normalization constant is 0.2E-5, that there are two poles and 3 zeros in the response. The input to the stage is nanometers and the output is Volts.

The DIG2 entry indicates that there are 0.664E+7 counts output for each Volt of input to this stage.

This information can be converted to the sac pole-zereo file given next.

The final pole-zero file in sac format is sac.pz which is also listed here:

* ****
* NETWORK   (KNETWK): 
* STATION   (KSTNM ): 
* COMPONENT (KCMPNM): 
* LOCATION  (KHOLE ):
* START             : 
* END               :
* INPUT UNIT        : METER
* OUTPUT UNIT       : COUNT
* LATITUDE  (DEG)   : 
* LONGITUDE (DEG)   :
* ELEVATION (M)     :
* DEPTH     (M)     :
* DIP       (DEG)   :
* AZIMUTH   (DEG)   :
* INSTRUMENT COMMENT:
* CHANNEL_FLAG      : 
* GSE 0.6647E+7 counts/nm at 40 sec
* constant is the product of the GSE scale 0.20000000E-05
* and the digitizer constant 0.66475240E+07 = 13.295 counts/nanometer
* This is equivalent to 1.3296E+10 for counts/meter
* ****
ZEROS 3
  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00
POLES 2
 -0.14660765E+00  0.14956973E+00
 -0.14660765E+00 -0.14956973E+00
CONSTANT 1.3296E+10

To test this file, use gsac to create a unit area impulse (Note that if you use sac, then you must divide by delta to create a unit area impulse)

GSAC> fg impulse delta 0.025 npts 8192
GSAC> transfer from none to polezero subtype sac.pz
GSAC> echo This is the displacement sensitivity in counts/meter
GSAC> title on l t S M Text "Displacement sensitivity in counts/meter"
GSAC> fft
GSAC> psp
GSAC> fg impulse delta 0.025 npts 8192
GSAC> transfer from vel to polezero subtype sac.pz
GSAC> echo This is the velocity sensitivity in counts/meter/sec
GSAC> title on l t S M Text "Velocity sensitivity in counts/meter/sec"
GSAC> fft
GSAC> psp

The result of these commands are two plots:

Displacment sensitivity in counts/meter

Velocity sensitivity in counts/meter/sec

The student provided the following information about the system:

The sac file was generated from a Guralp GCF format and the response file was produced with RESP program of Seisan using the following parameters:

Seismometer natural period = 30s

Seismometer damping ratio = 0.7

Sensor loaded generator constant = 2000 V/m/s

Recording media gain = 419430.4 Count/V

Amplifier gain = 24 DB

Digitiser sample rate = 40 Sample/s

 

Our seismometer is CMG-EDU based on CMG-6 technology while the digitizer model is CD24E1 

This means that the response of the system is as follows

This means that the response of the system is as follows
1 meter/sec -> Seismometer -> 2000 Volts (in the passband, e.g., 1 Hz)
1 Volt -> amplifier -> 15.85 Volts
1 Volt -> digitizer -> 419430.4 Counts.

Therefore the system outputs 1.3295E+10 counts/meter/second at 1 Hz. To obtain the displacment sensitivity at 1 Hz, we recall that 1 meter/sec at 1 Hz is equivalent to 1/6.2831853 meters. Thus the sensitivity is 8.3535E+10 counts/meter or 83.535 counts/nanometer or equivalently 0.01197 nanometers/count which is the number given on the CAL2 line.

Now that the polezero response used is correct since we see these values on the plots.

Getting ready for do_mft

Removing the instrument response:

Because modern broadband instruments have a flat response to ground velocity in the passband, I prefer to reduce reduce the gorund motion to units of meters/sec in the following manner for this data file:

GSAC> r 20080529_1600z.sac 
GSAC> transfer from polezero subtype sac.pz to vel freqlimits 0.004 0.005 10 20
GSAC> w vel.sac

In order to study group velocities you must place the event coordinates and origin time and the station coordinates into the sac file. As an example I will use the Iceland event for this time period and I will assume that your station is at latitude 0 and longitude 0: (you must use your actual station and event information)

GSAC> r vel.sac
GSAC> ch stla 0 stlo 0
GSAC> ch evla 63.92 evlo -21.17
GSAC> ch ocal 2008 05 29 15 46 6 200
GSAC> ch lcalda true lovrok true
GSAC> wh 

note that the ch ocal is a gsac command. For Sac you must use ch o gmt 2009 150 15 46 6 200. The lcalda and lovrok header variables must be set so that otehr programs can use the computed distances. To see if the header was properly set we can list the header contents:

GSAC> lh
vel.sac (0):
         NPTS               144000            B                    0
            E             3599.975        DELTA                0.025
       DEPMAX         2.584191e-06       DEPMIN        -2.779191e-06
       DEPMEN         6.959091e-09       NZYEAR                 2008
       NZJDAY                  150       NZHOUR                   16
        NZMIN                    0        NZSEC                    0
       NZMSEC                    0       KZDATE   May 29 (150), 2008
       KZTIME         16:00:00.000            O               -833.8
        KSTNM             C214           KCMPNM             B_Z
         STLA                    0         STLO                    0
         EVLA                63.92         EVLO               -21.17
         DIST             7301.042           AZ             156.6483
          BAZ             349.9096        GCARC             65.65807
        USER1                  0.1        USER2                  200
        CMPAZ                    0       CMPINC                    0
        NVHDR                    6       IFTYPE             ITIME
       LPSPOL                FALSE       LOVROK                 TRUE
       LCALDA                 TRUE        LHDR5                FALSE

do_mft

When you are studying teleseismic surface waves, you are note interested in frequencies greater than 1 Hz. If you resample, the time series will be smaller, and the computations will be much faster since the length of the Fast Fourier transforms will be shorter. To do this I lowpass filter and then interpolate:

GSAC> r vel.sac
GSAC> lp c 1 n 3
GSAC> interpolate delta 0.25 
GSAC> w dvel.sac
GSAC> quit