Thread: RE: Another example of filtering

Started: 2007-05-03 16:25:27
Last activity: 2007-05-03 22:33:57
Topics: IRIS EPO
John or Jan Lahr
2007-05-03 16:25:27
Hi Chris,

Good questions.

For the helicorder view AmaSeis only allows one filter to filter out
high frequencies and one to filter out low frequencies.

For example, if you want to enhance the high frequency signals from
local events on your helicorder record, you might filter out
frequencies below 1.5 Hz, which leaves frequencies between 1.5 Hz and
3 Hz. (Note that because we are only sampling at 6 Hz, the highest
frequency that can be resolved is 3 Hz.) The bottom of the
helicorder record would indicate, in this case, "High pass cutoff
period: 0.67 s"

To enhance the surface wave signals from distant earthquakes, one
could filter out frequencies above 1/10 Hz (shorter than 10 s
period). The AS-1 system doesn't record signals with frequencies
lower that 1/25 Hz (periods longer than 25 seconds), so I would also
filter out frequencies lower than 1/25 Hz. The bottom of the
helicorder screen would indicate, in this case, "Band pass cutoff
periods: 10 s 25 s"

When a highlighted period of time is first viewed in the event viewer
window, no filtering is applied, even if the helicorder record is
being viewed with a filter. Data in the event viewer window can be
filtered with the Control/Filter/Low pass.. and Control/Filter/High
pass... options. Each time one of those filters is selected the data
is filtered again. For example, to achieve a band pass from 10 s to
20 s, one would first apply a high pass filter with a corner at 20 s
and then apply a low pass filter with a corner a 10 s.

For my local event, I applied a high pass filter with the corner set
to 1 Hz. Then I applied the same filter twice more for a total of 3
applications. Each time the filter was applied it further reduced
the low frequency waves and enhanced the local event.

For a given event, I often try out various filter settings. If they
don't workout, I use use the Control/Unfilter option to return to the
original seismogram and then try again.

Note that in the helicorder view, along with filtering to reduce the
unwanted frequencies, the gain will need to be increased so that the
signal of interest is large enough to be clearly visible. I usually
increase the gain by about a factor of 10 when applying a filter. In
the event viewer window, the amplitudes are automatically adjusted to
fill the screen. This can make it a bit more difficult to adjust
the filter settings; the Control/Scale ... option can be used to
turn off the automatic scaling feature.

Hope this helps.
John

At 06:37 AM 5/3/2007, you wrote:
What would the low pass and high pass filter parameters be to highlight the
higher frequency for the local events?

Living near the New Madrid fault I was hoping to be able to see the micro
earthquakes that occur there regularly but so far I have not successfully
been able to identify them.

I saw on your web site that the low pass would be 1.5mhz. Should I enable
the high pass filter too? Or does the high pass filter not have an effect?

Then do you also filter again after the event is extracted? How is the
filter applied 3 times?


Thanks!

Chris Midden

-----Original Message-----
From: irised-bounces<at>iris.washington.edu
[irised-bounces<at>iris.washington.edu] On Behalf Of John or Jan Lahr
Sent: Thursday, May 03, 2007 1:30 AM
To: irised<at>iris.washington.edu
Subject: [irised] Another example of filtering

A while back I showed how to set up a helicorder filter in AmaSeis so
that the surface waves from distant earthquakes would be enhanced.
(See: http://jclahr.com/science/psn/as1/filtering/index.html)

The other day I received an earthquake notification message* from the
USGS about a 3.1 earthquake not too far from my station in Corvallis,
Oregon. Although I couldn't see anything on my helicorder record,
especially with the filter I was using to enhance low frequencies, by
switching to a filter that would bring out the high frequencies the
earthquake was clearly visible above the background noise. This
example clearly demonstrates how the frequency content of seismic
waves depends on the distance to the event. Local and regional
events will be rich in frequencies above 1 Hz, while distant events
are often best identified by surface waves with periods of 18 to 20 seconds.

See: http://jclahr.com/science/psn/as1/filtering/high_pass.html

*To subscribe to receive earthquake messages, go to this page:
http://earthquake.usgs.gov/eqcenter/ens/

Cheers,
John



#################################/ John C. Lahr
################################/ Emeritus Seismologist
###############################/ U.S. Geological Survey
==========================/ Central Region Geologic Hazards Team
#############################//#################################
############################//##################################
PO Box 548
/###################################
Corvallis, Oregon 97339 /===============================
Phone: (541) 758-2699 /####################################
Cell: (541) 740-4844 /#####################################
Fax: (413) 658-2699 /######################################
johnjan<at>lahr.org /#######################################
http://jclahr.com/science/


_______________________________________________
irised mailing list
irised<at>iris.washington.edu
http://www.iris.washington.edu/mailman/listinfo/irised

No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.467 / Virus Database: 269.6.2/784 - Release Date: 5/1/2007 2:57
PM


No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 5/2/2007 2:16
PM



_______________________________________________
irised mailing list
irised<at>iris.washington.edu
http://www.iris.washington.edu/mailman/listinfo/irised



  • Chris Midden
    2007-05-03 22:33:57
    Thanks John,

    This makes sense and will help me view my seismograms in a whole new way.
    It is exactly what I needed.


    Chris

    -----Original Message-----
    From: irised-bounces<at>iris.washington.edu
    [irised-bounces<at>iris.washington.edu] On Behalf Of John or Jan Lahr
    Sent: Thursday, May 03, 2007 11:25 AM
    To: irised<at>iris.washington.edu
    Subject: RE: [irised] Another example of filtering

    Hi Chris,

    Good questions.

    For the helicorder view AmaSeis only allows one filter to filter out
    high frequencies and one to filter out low frequencies.

    For example, if you want to enhance the high frequency signals from
    local events on your helicorder record, you might filter out
    frequencies below 1.5 Hz, which leaves frequencies between 1.5 Hz and
    3 Hz. (Note that because we are only sampling at 6 Hz, the highest
    frequency that can be resolved is 3 Hz.) The bottom of the
    helicorder record would indicate, in this case, "High pass cutoff
    period: 0.67 s"

    To enhance the surface wave signals from distant earthquakes, one
    could filter out frequencies above 1/10 Hz (shorter than 10 s
    period). The AS-1 system doesn't record signals with frequencies
    lower that 1/25 Hz (periods longer than 25 seconds), so I would also
    filter out frequencies lower than 1/25 Hz. The bottom of the
    helicorder screen would indicate, in this case, "Band pass cutoff
    periods: 10 s 25 s"

    When a highlighted period of time is first viewed in the event viewer
    window, no filtering is applied, even if the helicorder record is
    being viewed with a filter. Data in the event viewer window can be
    filtered with the Control/Filter/Low pass.. and Control/Filter/High
    pass... options. Each time one of those filters is selected the data
    is filtered again. For example, to achieve a band pass from 10 s to
    20 s, one would first apply a high pass filter with a corner at 20 s
    and then apply a low pass filter with a corner a 10 s.

    For my local event, I applied a high pass filter with the corner set
    to 1 Hz. Then I applied the same filter twice more for a total of 3
    applications. Each time the filter was applied it further reduced
    the low frequency waves and enhanced the local event.

    For a given event, I often try out various filter settings. If they
    don't workout, I use use the Control/Unfilter option to return to the
    original seismogram and then try again.

    Note that in the helicorder view, along with filtering to reduce the
    unwanted frequencies, the gain will need to be increased so that the
    signal of interest is large enough to be clearly visible. I usually
    increase the gain by about a factor of 10 when applying a filter. In
    the event viewer window, the amplitudes are automatically adjusted to
    fill the screen. This can make it a bit more difficult to adjust
    the filter settings; the Control/Scale ... option can be used to
    turn off the automatic scaling feature.

    Hope this helps.
    John

    At 06:37 AM 5/3/2007, you wrote:
    What would the low pass and high pass filter parameters be to highlight the
    higher frequency for the local events?

    Living near the New Madrid fault I was hoping to be able to see the micro
    earthquakes that occur there regularly but so far I have not successfully
    been able to identify them.

    I saw on your web site that the low pass would be 1.5mhz. Should I enable
    the high pass filter too? Or does the high pass filter not have an effect?

    Then do you also filter again after the event is extracted? How is the
    filter applied 3 times?


    Thanks!

    Chris Midden

    -----Original Message-----
    From: irised-bounces<at>iris.washington.edu
    [irised-bounces<at>iris.washington.edu] On Behalf Of John or Jan Lahr
    Sent: Thursday, May 03, 2007 1:30 AM
    To: irised<at>iris.washington.edu
    Subject: [irised] Another example of filtering

    A while back I showed how to set up a helicorder filter in AmaSeis so
    that the surface waves from distant earthquakes would be enhanced.
    (See: http://jclahr.com/science/psn/as1/filtering/index.html)

    The other day I received an earthquake notification message* from the
    USGS about a 3.1 earthquake not too far from my station in Corvallis,
    Oregon. Although I couldn't see anything on my helicorder record,
    especially with the filter I was using to enhance low frequencies, by
    switching to a filter that would bring out the high frequencies the
    earthquake was clearly visible above the background noise. This
    example clearly demonstrates how the frequency content of seismic
    waves depends on the distance to the event. Local and regional
    events will be rich in frequencies above 1 Hz, while distant events
    are often best identified by surface waves with periods of 18 to 20
    seconds.

    See: http://jclahr.com/science/psn/as1/filtering/high_pass.html

    *To subscribe to receive earthquake messages, go to this page:
    http://earthquake.usgs.gov/eqcenter/ens/

    Cheers,
    John



    #################################/ John C. Lahr
    ################################/ Emeritus Seismologist
    ###############################/ U.S. Geological Survey
    ==========================/ Central Region Geologic Hazards Team
    #############################//#################################
    ############################//##################################
    PO Box 548
    /###################################
    Corvallis, Oregon 97339 /===============================
    Phone: (541) 758-2699 /####################################
    Cell: (541) 740-4844
    /#####################################
    Fax: (413) 658-2699 /######################################
    johnjan<at>lahr.org
    /#######################################
    http://jclahr.com/science/


    _______________________________________________
    irised mailing list
    irised<at>iris.washington.edu
    http://www.iris.washington.edu/mailman/listinfo/irised

    No virus found in this incoming message.
    Checked by AVG Free Edition.
    Version: 7.5.467 / Virus Database: 269.6.2/784 - Release Date: 5/1/2007
    2:57
    PM


    No virus found in this outgoing message.
    Checked by AVG Free Edition.
    Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 5/2/2007
    2:16
    PM



    _______________________________________________
    irised mailing list
    irised<at>iris.washington.edu
    http://www.iris.washington.edu/mailman/listinfo/irised


    _______________________________________________
    irised mailing list
    irised<at>iris.washington.edu
    http://www.iris.washington.edu/mailman/listinfo/irised

    No virus found in this incoming message.
    Checked by AVG Free Edition.
    Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 5/2/2007 2:16
    PM


    No virus found in this outgoing message.
    Checked by AVG Free Edition.
    Version: 7.5.467 / Virus Database: 269.6.2/785 - Release Date: 5/2/2007 2:16
    PM




22:56:56 v.22510d55