[data-issues] Re: [TA_work_group] metadata problem with TA stations

Chad chad at iris.washington.edu
Thu Oct 5 16:58:22 PDT 2006

Hi Gary,

Thanks for the report.  The formal place to report this at the DMC is the
data-issues mailing list:


Of course since it's a Caltech station and Egill will probably get this.

Regarding a), it is very common for data suppliers creating SEED to list only
the sensor in the instrument identifier of a channel.  The convention used in
Antelope to list the sensor and digitizer is just that.  SEED, unfortunately,
does not get specific about what needs to be listed in the field.

Regarding b), I think there is an issue with how the composite FIR filter
delay is specified and interpreted in SEED, changing the sign of the delay for
stage 3 gives me a response similar to that of other responses for Q330's
using the Linear below 100-40 scheme.

Regarding c), up until Aug. 2006 MLAC was recording at 20 sps according to the
metadata.  Normally all TA stations are 40 sps, but there are some anomalies
especially with the cooperative stations.  I don't think we have any TA
stations recording both 20 and 40 sps, it's either or.  Which page indicated
this should be 40 sps data?


Gary Pavlis wrote:
> There appears to be a problem with the metadata for some of
> the Caltek stations used in the TA.  At least the result I 
> have at the end of a processing chain has some clear problems.
> I'm not sure how to report this formally, but I'm going to 
> cast it to this group since it seems to me to fall into our
> role.
> Attached is a result of data I obtained following this chain:
> 1. I used plain old Wilber II to download the data from the
> event I showed you earlier today.  
> 2.  I ran the file I got through Antelope 4.8 sd2db (seed2db seg faulted on
> me on this file for some unexplained reason so I had to use
> the older program sd2db).  
> 3.  I had some problems with EDW2 so I was poking around the
> response information to try to figure out the type of digitizers
> it had.  I did this through Antelope's dbe program.  I found
> these problems:
> a.  The instrument names in the metadata for Caltek stations
> are incomplete.  They seem tagged only by sensor type.  I was
> interested in knowing the type of instrument to track down the
> oddity that was causing me problems (irregular sample rate on
> data from different B channels on the same stations)  I checked
> in the new nifty web pages Tim pointed us to and they propagate
> this problem.  That is stations like EDW2 list the instrument
> type as CMG-3T.  That is the sensor, but not the instrument
> per se.  When I look in the stage table I see only a generic
> "digitizer" to describe digitizer stage use for the response
> definition.  
> b.  More important is that the response information for two
> stations I see here are not correct:  ARV and EDW2.  Attached
> is what dbe gives me as a plot for the response.  It looks like 
> the wrong anti-alias stage was defined for this channel.
> c.  Station MLAC has an oddity.  For some reason the FARM 
> data has the 20 sps B channel data.  When I looked at the 
> US-TA web site Tim told us about, however, it claims this
> stations should be producing 40 sps B channel data also.  
> Why did I get the 20 sps data instead of 40 sps for this station?
> That's it.  By the way the original problem I was chasing was an
> oddity of sd2db.  For some odd reason it dithered the sample rate
> on one channel of EDW2 and not the others.  When I used their
> psd2db that problem disappeared.  The others noted above did 
> not.  
> ===============================================================
> Professor Gary Pavlis
> Department of Geological Sciences
> Indiana University 
> 1001 East 10th Street
> Bloomington, IN 47405
> Phone:  812-855-5141, FAX:  812-855-7899
> email:  pavlis at indiana.edu
> ===============================================================

More information about the data-issues mailing list