Position Error Question

Does the position error resulting from the *best* possible satellite
reading, for a given location, change much over a given time period. I know
this sounds like a vague question so let me give an example.
Another poster wanted to know if the Nuvi could be used to mark the
position of his car in a parking lot and the answer given was yes, but only
within 5 meters. If the user saved the location (with a 5 meter error) in
the Nuvi, and tried to return 2 hours later to the same location, would the
return location be closer than 5 meters based on the saved location? Or
would another reading error be present which would add to the 5 meter error
in the saved location?

Don F

Don F wrote:
Understandably, the error on my 60CSX reduces noticeably after dark. I
assume that of the 5m error had reduced to 3m then you would be within
3m of the centre of the 5m possible location, logic says there is no way
it could reduce the original error.

Den Sun, 14 Jan 2007 15:36:41 +0000 skrev Broadback:

Logically, if you got the same error everytime you tried in the same
position, you could get a very precise same spot, even though there
was a large error.

However, this could only happen if the satellites were fixed position.
As they move, the error will change, and the errors will add up.

/Kent
--
"So there I was surrounded by all these scary creatures
They were even scarier than what Microsoft call features"
- C64Mafia: Forbidden Forest (Don't Go Walking Slow).

Don F wrote:

The second measurement may well be determined using a different set of
satellites than the first one, so it should be assumed that they are
two independent measurements. The uncertainty would then add in
quadrature; i.e. total uncertainty = sqrt(unc1^2 + unc2^2). In this
case, if both uncertainties are 5m, then the total uncertainty would be
about 7m.