From:    "David W. Inouye" 
Subject: Summary of responses about differential GPS now that SA is turned off

Here's a useful web site (Interagency GPS Executive Board) and a few of the
many responses that I got to my query a few hours ago.  Thanks to those who
responded. I got some responses that said differential GPS would no longer
provide any benefit, but the majority report that it still does, both in
theory and in practice.

David Inouye


David,

This question is covered in the IGEB's FAQ page....see
http://www.igeb.gov/sa/faq.shtml


It would depend on the level of accuracy you are trying to obtain.  The
"descrambled signal" that is now being put out increases accuracy ten fold.
Which basically means for the handheld devices that you can purchase
relatively cheaply, the accuracy goes from 100 meters to 10 meters.  For
survey grade equipment you will still need to use DGPS, especially if you
want submeter or even subcentimeter reliability.  One of our survey guys
went out to a fixed coordinate position in the morning and afternoon to see
how "off" the new descrambled signal is.  In the morning he was four feet
off and in the afternoon he was 18 feet off.  He did find that it was much
easier to lock onto the satellites and obtain a position, and that once that
position was obtained the signal did not budge, but for the sake of accuracy
that doesn't do too much good.  Again, it appears that if you want submeter
accuracy or even subcentimeter accuracy you will still need to use DGPS or
post processing.


David--

Empirically, we still need differential GPS with post-processing for good
elevation data.  Using a pair of survey-grade 5600s, we can get horizontal
fixes in the range of 1-2cm per km of separation from the base station, and
our vertical fixes are ~10cm.  That may not matter for you out there in the
promised land of nearly 3 dimensions, but we needed that precision for
determining flooding patterns in the Congaree Swamp, where 10cm is a ridge.

Theoretically, post-processing can deal with the atmospheric variation.  The
survey-grade units get cm accuracy, which will always require high-quality
differential mode to make such corrections.  What we haven't tested yet is
whether the removal of selective availability allows post-processing of
_moderate_ units to get much better.  With selective availability,
moderate receivers gave about 1m accuracy, while the military units gave ~2m
real time non-differential.  Removing SA should make decent units the
equivalent of the military units in nondifferential mode.  My guess is
that moderate units (or at least the software compatible with them)
can't do the atmospheric corrections to get much better, so the middle
of the market ($1-5k) may be obsolete.

David,

We are currently using Trimble Geo II GPS units and still do differentially
correct.  We were told by our Trimble distributor that the Geo II will be
within 10 meters without correcting (now that selective availability is
turned off).  We still can correct to get < 5 m accuracy.  If you can live
with +/- 10 m, there is no need to correct.  However, we collect attribute
data (e.g. habitat measurements) with out GPS unit (which stores these data
in a file).  Because we need to download the stored file with the
attributes, we just correct the coordinates anyway.  I would say if you are
just recording the coordinates on a data sheet and not worried about storing
attribute data in the GPS unit, the 10 m accuracy would be fine for most
ecological work.

David,
Yes.  Removal of the selective availability will improve the accuracy of
uncorrected GPS readings to approximately 20 to 40 m.  Using a differential
correction, either a base station or subscription service, can improve the
accuracy of the position fix to the meter to sub-meter range.

David,

I had this discussion with our GIS manager and we concluded that
differential is
still the way to go.  With the SA turned off you accuracy will still be + or -
10 meters or so.  Differential can get you sub-meter accuracy.  If you think
about how differential works, the base station constantly collects data and
creates a cloud of points around the actual location.  By turning off SA all
that will happen (I think) is the cloud of points around your tower will have
less variance.  The average (your actual tower location) should not change.

I suppose if +/- 10 meters will suffice, then the extra expense (and work) of
differentially corecting your GPS data might not be worth it.  Hope some of
this
helps.

Jeff

I would say 'yes', because, although gps fixes will be much improved,
  there are still inaccuracies caused by atmospheric conditions, solar
storms, whatever that can be corrected for.  The best part is that
immediate readings in the field will be much more accurate.
    Our newly gathered non-SA data is much improved from
the SA days.  We are currently correcting some readings to see whether
it is worth the extra post-processing step.

Hi,
     According to the garmin website:

While the termination of SA will enhance GPS position data, better
accuracy is still attained through the use of differential GPS
(DGPS), which offers users accuracy within 1-5 meters.

GPS receivers will automatically adjust to the termination of SA, so
no software upgrade or hardware changes are needed to take advantage
of this development. Instead, consumers will enjoy the benefits of
better performance at the push of a button.

------------------------------