GPS alert: Selective Availability

[cbadmin]

From: Fun Patrol (Original Message) Sent: 2/28/2003 11:55 AM
Greetings all:

For those that may not have been involved with GPS in 1991, here's some important information. During the first Gulf War in 91 the US Government activated what they call "Selective Availability" on the GPS system. What this does is dilute the precision of your onboard readings. This "S.A." was left on until only a couple of years ago, when they shut it off and we were all treated to the spectacular accuracy we use now.

For most uses it won't cause too much trouble, but if you rely on very close readings, you should be aware of the possibility of degraded performance if the govenment activates S.A. due to the military operation in Iraq.

I was cruising Mexico in 91, with my then new $2,500 single waypoint GPS (!!!), when the system was altered. The immediate effect was the anchor watch feature was useless, the indicated position would wander too much. In addition the SOG readings jumped around a lot, making comparative speeds due to sail changes, etc. invalid.

I am flying out Monday to the Marshall Islands for 6 weeks on a 45 ft catamaran where we will be attempting some narrow reef passes that have been previously entered in the plotter. These are sometimes only 50 ft wide and the boat has a beam of 25 !! So, we will be paying very close attention to the S.A. status.

I don't know how that might affect units with WAAS or DGPS, I'm sure we'll know after the fact.

Be safe out there.....Roy

From: Chuck S Sent: 2/28/2003 3:07 PM
Even with SA ON (degraded civilian positioning) you'll get 90 meter positioning 95% of the time and under 40 meter positioning over half the time. This is well under the width of a sharp pencil line on a near costal chart.

Won't tell you which end of your boat you're standing on, or if you're in the right slip at the yacht club, but SA has no meaningful effect on navigation and never did.

-- Chuck


From: Chuck S Sent: 3/1/2003 5:59 AM
I don't worry much about SA-ON or OFF.

Here's a link to researchg on GPS errors. Lots of other GPS techie stuff on the site, so look so look around: http://users.erols.com/dlwilson/gpscmpsa.htm This site reports under 55 meters 95% of the time, and 22.2 meter positioning 50% of the time.

With SA-ON even my early sequential channel GPS units were getting 10 meter positioning almost all the time and rarely over 40 meters. This was confirmed by standing on top of surveyed points or sitting on top of the cab of MLRS (multiple launch rocket system) vehicles in the field. The MLRS crews were getting encrypted sub meter positioning. Ad hoc testing to be sure, but after a couple dozen position checks with the MLRS crews I didn't worry about error.

GPS vertical positioning error is typically 3x that of horizontal (due to satellite geometry of something) so trying to use GPS for an aircraft landing system with SA ON could have 300 meter errors, about 1000 feet!

Datum is the critical setting on GPS receivers. Most marine use is Lat/Long and there's little difference between NAD27 and WGS84, but using a grid system like UTM or MGRS there can be differences of 350 meters. At Fort Knox it was usually pretty easy to see which troops were using personal GPS receivers on the land navigation courses, they'd end up 300 meters north of their intended position 'cuz they left them in default WGS84 datum, but the old maps we used for training were NAD27. We finally had to admit to progress and held classes on how to use these bootleg devices.

-- Chuck

From: B.Anderson Sent: 3/17/2003 1:10 PM
There IS a significant difference between NAD 27 and WGS 84 in Alaska (and I'm assuming other regions, as well). It can be up to 180 feet. Only significant if that rock is 180 feet from you!

Also keep in mind that if the military elects to turn on SA, the old workarounds are still available. DGPS using the Coast Guard differential system will still give you 3-7 meter accuracy.

WAAS will probably be left on, though it is unavailable to us in Alaska (too low on the horizon).