GPS Error

[chuck]

Hello,

I recently downloaded a shapefile of LA county roads from the census
web site. I am using ArcGIS 9.2. Anyway, I set the datum to NAD 83
just as specified in the Meta Data at the census web site. I set the
map units and display units to decimal degrees. I have an old Garmin
12XL and I set the datum on the unit to NAD 83 with decimal degrees
showing. I took a reading with the GPS and the error indicated 13
ft. I know the error is based on a 60% probability. So, I am
expecting it to be off a number of feet. The probability is "poor".
I mapped the point on ArcMap using the "road" layer and I was off
around 80 feet according to the measuring tool. Why am I so far off?
Please help.

CB

[Jean H]

... maybe it is not the GPS accuracy but the map accuracy... Census
data is usually not known to be the most accurate one!

Now, the "accuracy" that your read is the one of the GPS, not of your
point. You can see it as the accuracy if you have a perfect satellite
coverage.... which is not often the case.
Your point accuracy would be something close to: 1.96 (95% confidence
interval) * DOP value * 3.962m (=13ft) = 7.76 * DOP

DOP (dilution of precision) is a function of the satellite coverage. DOP
is "good" between 1 and 6... so, with a "good" coverage, your point
precision would be up to 7.76*6 = 46.56m ... you say your are around 24m
off... so there is nothing to worry about!

Jean

[dan]

"Jean H" <jgha...@DELTHIS.ucalgary.ANDTHIS.ca> wrote in
message news:fsj5e7$si1$1...@news.ucalgary.ca...

The link below shows the counties having the most accurate
data--7.6 m or better. Los Angeles, CA, doesn't make the
list so Jean's post about map accuracy seems correct.

http://www.census.gov/geo/www/tiger/tiger2006se/tl06semtaip.txt

If you overlay the Tiger data onto ortho-imagery, you'll
probably be able to see some of the road data problems.

Dan