Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

[Phys-L] Re: earthquake



Michael Edmiston wrote:
do we need any adjustment in the GPS system?

Yes, indeed. The GPS system is routinely and continually adjusted.

The GPS coordinates are effectively phase-locked to the solid earth.
Specifically, there are five permanent monitoring stations, and
corrections are applied to ensure that the coordinates of those
stations do not change ... or rather, change as little as possible,
considering that the stations move relative to one another due to
plate tectonics.

There are Air Force officers (yes, officers, with higher rank than you
might have guessed) whose sole job is to monitor the system and make
sure nothing very bad happens. The routine corrections are handled
by a bunch of computers, but humans make decisions in response to
relatively rate events such as the demise of a satellite.

For details see e.g.
http://www.af.mil/factsheets/factsheet.asp?fsID=119

Also note that users who want high accuracy set up their own fixed
reference stations and use Differential GPS. For instance, if you're
running a large farm, you set up a reference station at one corner of
the field, so you can use Differential GPS to locate anything anywhere
on the farm with cm-scale reproducibility. You don't actually care
whether the Air Force lets the absolute lat/long of your farmhouse
drift or not; that just looks like an arbitrary gauge, and your
Differential GPS measurements are gauge-invariant.

Ditto for aircraft making GPS-based precision approaches to an airfield.
A pseudo-satellite (pseudolite) sitting at mid-field will compensate
for a multitude of sins.

FWIW note that internally, GPS time differs from UTC, in that it never
intercalates leap seconds. The last I knew, GPS minus UTC was 13 seconds.
Your garden-variety GPS receiver corrects for this, though, so that it
normally displays UTC, which is what most users expect.