It's up now.
http://online.wr.usgs.gov/calendar/2012/apr12.html

For a good time sink, my favorite talk was March 2011:
Unraveling the Mystery of Avian Navigation
http://online.wr.usgs.gov/calendar/2011/mar11.html
A little more text back at:
http://online.wr.usgs.gov/calendar/2011.html

For anybody in the Silicon Valley area, their open house is May 19-20.  It's kid friendly, both big kids and little kids.  You get to talk to the people who are actually doing the work.  They are happy with geeky questions.

Back to somewhat time-nutty stuff...

Does anybody understand how they are using GPS and/or have performance numbers?

They don't need the actual position (DC), just the changes in position.  They need it now.  They can't wait for post processing.  I'm not sure how much accuracy they need.  I'd guess in the cm range.

(Maybe I can learn more at the open house.)

--
These are my opinions, not necessarily my employer's.  I hate spam.

On 04/28/2012 02:39 AM, Hal Murray wrote:

Extract the ECEF position, build a long term average position. Subtract
each given position with the known ECEF average, square the differences,
sum and square root to get an RMS value... set a threshold for
trigger... whenever the difference RMS is above threashold start
reporting. Keeping a memory of previous diffs will get a pre-trigger memory.

Not too hard, given good antenna and carrier phase, preferably double
frequency reciever.

Cheers,
Magnus

On 4/27/12 5:39 PM, Hal Murray wrote:

I'm going to guess that GPS doesn't have much to do with the detection,
per se, an accelerometer works quite nicely.  What you want GPS for is
precision timing, so you can figure out where the shock wave started and
is going, but combining data from multiple stations.

zipping along at a few thousand meters per second, timing to
milliseconds is probably good enough, but pretty darn tough to do in a
field site without something easy like GPS to give you a nice time hack.

the physical displacement during an earthquake isn't all that much (a
few mm or cm, unless you're right on the fault and/or it's a big one)

here's a M4 a couple months ago
http://earthquake.usgs.gov/earthquakes/eventpage/ci15141521#summary

shows max accels of around 0.2 %g  (I assume that means about 0.002g, or
0.02 m/s^2)

that works out to a displacement on the order of mm, over a time of tens
of milliseconds.

Hi Hal:

In the talk there was a slide showing a comparison between ground position calculated from an accelerometer and a real
time precision GPS.
The Accelerometer is AC coupled and so misses the DC coupled GPS answer that shows the permanent ground movement.
I'm guessing it takes a GPS receiver that has 100 Hz or faster outputs that can be reduced to cm or better position to
do this.

Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/Clarke4Congress.html

Hal Murray wrote:

Brooke,

In the papers they were getting some results with even 1 Hz sampling,
but, as expected, 10 Hz was better. That's probably sufficient for seismic
waves; 100 Hz is overkill. See Figure 5 of the Larson paper I for a nice
example of the AC vs. DC coupling that you mentioned.

I also agree with Jim's earlier comment; for earthquake detection it
seems an cheap accelerometer is more than adequate. One doesn't
need the expense of dual frequency carrier phase gps receivers just
to detect a local shake.

These days, there are always many cell sites where there are many
people; and each site already has GPS timing, battery backup, and
a fast connection to a central office; so it's the perfect place to add
a sensitive accelerometer. You could just call it an security intrusion
monitor and use it for earthquake detection as a free side effect.

/tvb

Hi Tom:

The USGS talk was the first time I'd heard about the need to look at an earthquake as happening along some length of
fault line.  For the big quake in Japan the forecast software assumed a point source for the quake and that cause them
to under estimate the magnitude and get other things wrong.  GPS is part of the solution to get better results.

In the S. CA example he showed a 180 mile long rupture of the San Andreas fault.  At 2 miles a second the quake would
last about 90 seconds.
Accelerometers that are not right on top of the fault will be overloaded with signals coming from each location where
there's a fracture and so the data will be nearly impossible to untangle in a short time frame.  But a GPS receiver will
show a DC displacement that unambiguous.

Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/Clarke4Congress.html

Tom Van Baak wrote:

Brooke,

Right, an overloaded accelerometer is a problem -- if you have
only one or a few of them.

But the beauty of using cellular sites is that you have hundreds
or thousands of them across populated areas; so it's no problem
if the a bunch of sensors near the epicenter overload. A clipped
signal is not worthless; at least you know something big happened
there; you can rely on slightly more distant cell tower sensors to
get readings a few seconds later that are less clipped or not clipped
at all. (There's another solution I heard about -- using smartphones
as a tiered network of synchronized accelerometers).

A high rate GPS solution sounds really cool to me but I bet its also
far more expensive.

Related to that, are there any seismometer experts on the list? I've
always wondered why they don't augment the extremely sensitive
detectors with less sensitive detectors? Of course a really good
detector will overload; so just co-locate cheap detectors that are 40
and 80 dB less sensitive. That way you get a clean signal no matter
how close or far the epicenter is from the detector.

/tvb

----- Original Message -----
From: "Brooke Clarke" brooke@pacific.net
To: "Tom Van Baak" tvb@LeapSecond.com; "Discussion of precise time and frequency measurement" time-nuts@febo.com
Sent: Saturday, April 28, 2012 2:49 PM
Subject: Re: [time-nuts] GPS, USGS Early Earthquake Warning

Hi Tom:

They do use two different seismometers at each location, a large movement and a sensitive.
http://www.prc68.com/I/Seismometer.shtml

Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/Clarke4Congress.html

Tom Van Baak wrote:

WIWAUG (when I was an under grad), it was explained to me that the seismometers were log scale and basically don't clip. When you get into a big event, the last few decimal places just don't matter.

We also used differential GPS across known fault lines to measure slip both over time and during events, which are often 'step' slips of a bit at a time over the course of the event, which may last seconds or hours.

Bob

On Apr 28, 2012, at 18:32, "Tom Van Baak" tvb@LeapSecond.com wrote: