International Federation of Digital Seismograph Networks

Thread: SEED band code clarification requested

Started: Jan. 8, 2009, 10:06 a.m.
Last activity: Jan. 8, 2009, 10:06 a.m.
Tim Knight
Jan. 8, 2009, 10:06 a.m.
forwarded message:
From: Doug Neuhauser <doug<at>>
Date: December 24, 2008 12:36:07 PM PST
Subject: SEED band code clarification requested

The SEED manual (Appendix A) provides some guidelines for band code
assignment based on sample rate, and for higher sampling rate, based
on corner frequency. These have recently been extended to define
band codes for higher sample rates based on the same corner frequency

Band Band type Sample rate Corner period
code (sec)
B Broad band >= 10 to < 80 >= 10 sec
H High Broad Band >= 80 to < 250 >= 10 sec
C ... >= 250 to < 1000 >= 10 sec
F ... >= 1000 to < 5000 >= 10 sec

S Short Period >= 10 to < 80 < 10 sec
E Extreme Short Period >= 80 to < 250 < 10 sec
D ... >= 250 to < 1000 < 10 sec
G ... >= 1000 to < 5000 < 10 sec

I and others in the SEED community would like some clarification and
on what the term "Corner period (sec)" refers to. Specifically, it
is not
clear into which of the above "corner period" categories data from
accelerometers and other geophysical instruments such as electric
and magnetic
field sensors should be classified.

For traditional accelerometers:
If the "corner period in seconds" refers to the lower corner of the
passband of the measured observation (eg acceleration for an
then I would infer that accelerometers should be classified along with
"broadband" sensors in the first group of band codes where the
corner period
of >= 10 seconds, since the response of acceleromters is flat to ~
DC (0 Hz ~=
infinity seconds). However, if the corner period in seconds refers
to the
natural period of the actual sensor, most accelerometers have a
natural period
of several Hz (< 1 second), from which I would infer that they
should be
classified with "short period" on the second group of band codes
with a corner
period of < 10 seconds. If the corner period is based on the
of the physical sensor, and not of the property of the data produced
by the
sensor, it is unclear how this may change for non-traditional
sensors such as
silicon-based or fluid sensors.

It sould be extremely helpful to have additional guidelines to help
the SEED community standardize on channel names. Currently, I have
seen both E and H bandcodes uses for 100 Hz accelerometer data,
and the same confusion exists for higher sample rates.

For other geophysical data channels such as electric and magentic
sensors, the same question arises. For both electric and magnetic
field data
channels, the response of data channels are often flat to close to
DC, so I
would infer that they should be classified along with the broad band
However, if the classification is based on the senor
characteristics, it is is
not clear what characteristics of the actual sensor you would use for

- Doug N

Doug Neuhauser University of California, Berkeley
doug<at> Berkeley Seismological Laboratory
Office: 510-642-0931 215 McCone Hall # 4760
Fax: 510-643-5811 Berkeley, CA 94720-4760
Remote: 530-752-5615 (Wed,Fri)