[Proj] Dataset mismatch?

Ed McNierney ed at topozone.com
Fri Jan 27 14:49:34 EST 2006

Jarrett -

Just to give you a feel for the situation, take a look at the NGS NADCON
page at:


The top of the page shows two rough graphics charting the NAD27/NAD83
datum shift in both latitude and longitude.  The heavy blue isolines are
in 0.5-second increments, so you need to move from one line to the next
for the size of the datum shift to change by 0.5 seconds.  Very roughly
speaking, 0.5 seconds of arc in the US is about 12 - 15 meters, so these
charts show how far you have to travel for the size of the datum shift
to get to be 12 - 15 meters.

I realize you can't do precise calculations from these charts, but the
longitude isolines are generally closer together and they look to be
about 5 degrees of longitude apart.  That would mean the longitude datum
shift might be - on average) about 0.5 seconds per 5 degrees or 0.1
seconds per degree.  A 7.5-minute quad is one-eighth of a degree on a
side, so one would expect a datum shift (in each of latitude and
longitude) on the order of one-eigth of 0.1 or 0.0125 seconds.  That's
about a third of a meter in each direction, or perhaps as much as half a
meter altogether.

These are VERY rough "in the ballpark" calculations, but they're
intended to get a feel for the order of magnitude of the shift.  That's
why I would be surprised to find a datum shift that varies by more than
one pixel across a 7.5-minute quad.

	- Ed

Ed McNierney
President and Chief Mapmaker
TopoZone.com / Maps a la carte, Inc.
73 Princeton Street, Suite 305
North Chelmsford, MA  01863
ed at topozone.com
(978) 251-4242 

More information about the Proj mailing list