Thread: Difference in 101.3b vs 101.4 with gcp rotation

Started: 2011-07-06 22:19:25
Last activity: 2011-07-06 22:19:25
Topics: SAC Help
Andrew Schaeffer
2011-07-06 22:19:25
Hello,

I am trying to resolve a difference between the SAC versions 101.3b and
101.4.

*In 101.3b*:
I am performing "rotate to gcp" on a north and east component seismogram,
and writing out the radial and transverse components:
r s1.dn s1.de
rotate to gcp
->If I list the header (lh) at this point, the KCMPNM header value is now
empty (ie doesn't show up).
write s1.dr s1.dt
quit

If I start sac again, and read in the dr and dt components, they now have
the KCMPNM header value, with "Q" for both components.

If I repeat the same procedure in 101.4, the KCMPNM header value remains
empty, even after writing the seismograms and reloading them.
Can anyone confirm whether this is the expected behaviour (ie the difference
between the two versions), or what their version is doing?

many thanks,
Andrew

  • Brian Savage
    2011-07-06 18:08:45
    Andrew,

    What you are seeing is correct.

    In previous versions of SAC, 101.3 and earlier:
    If the station name (kstnm) is not defined, it would default to "sta"
    If the channel name (kcmpnm) is not defined, it would default to "Q"
    The stations and channel names were only changed on read.
    It was added a long while ago to "Disallow undefined kstnm and kcmpnm".

    We removed this in 101.4. It was considered to be confusing behavior.
    From the HISTORY file:
    * BUG FIX: Stations no longer named 'sta' and channels 'Q' by default

    If you need this, or have scripts/programs that rely on it, you can change the header values to a something you expect.
    SAC> ch kcmpnm Q
    or if you know they are radial transverse, you could use
    SAC> ch kcmpnm BHR

    For your processing:
    On the initial read, the fields kcmpnm for both files are probably set to BHN and BHE
    The ROTATE command sets the kcmpnm fileds to undefined (-12345.0)
    You write the files out. The kcmpnm is still set as "-12345.0"
    You read the files back in. On this read the kcmpnm is set to "Q" as 101.3 did not allow a undefined kcmpnm field.

    If this change caused havoc, we apologize.

    I hope this helps.

    Brian Savage
    savage<at>uri.edu

    Example run using 101.4:

    SAC> fg seismo
    SAC> ch cmpinc 90
    SAC> ch cmpaz 0
    SAC> ch kcmpnm BHN
    SAC> w tmp.n
    SAC> ch cmpaz 90
    SAC> ch kcmpnm BHE
    SAC> w tmp.e
    SAC> read tmp.n tmp.e
    SAC> lh cmpaz cmpinc az baz kcmpnm columns 2


    FILE: tmp.n - 1
    -----------

    cmpaz = 0.000000e+00 cmpinc = 9.000000e+01
    az = 8.814721e+01 baz = 2.718528e+02
    kcmpnm = BHN

    FILE: tmp.e - 2
    -----------

    cmpaz = 9.000000e+01 cmpinc = 9.000000e+01
    az = 8.814721e+01 baz = 2.718528e+02
    kcmpnm = BHE
    SAC> rotate
    SAC> lh cmpaz cmpinc az baz kcmpnm columns 2


    FILE: tmp.n - 1
    -----------

    cmpaz = 9.185278e+01 cmpinc = 9.000000e+01
    az = 8.814721e+01 baz = 2.718528e+02

    FILE: tmp.e - 2
    -----------

    cmpaz = 1.818528e+02 cmpinc = 9.000000e+01
    az = 8.814721e+01 baz = 2.718528e+02
    SAC> evaluate ( ( &1,baz + 180 ) - &1,cmpaz )
    3.6000000e+02


    On Jul 6, 2011, at 10:19 AM, Andrew Schaeffer wrote:

    Hello,

    I am trying to resolve a difference between the SAC versions 101.3b and 101.4.

    In 101.3b:
    I am performing "rotate to gcp" on a north and east component seismogram, and writing out the radial and transverse components:
    r s1.dn s1.de
    rotate to gcp
    ->If I list the header (lh) at this point, the KCMPNM header value is now empty (ie doesn't show up).
    write s1.dr s1.dt
    quit

    If I start sac again, and read in the dr and dt components, they now have the KCMPNM header value, with "Q" for both components.

    If I repeat the same procedure in 101.4, the KCMPNM header value remains empty, even after writing the seismograms and reloading them.
    Can anyone confirm whether this is the expected behaviour (ie the difference between the two versions), or what their version is doing?

    many thanks,
    Andrew
    _______________________________________________
    sac-help mailing list
    sac-help<at>iris.washington.edu
    http://www.iris.washington.edu/mailman/listinfo/sac-help


03:35:55 v.01697673