shrug-l: WGS84 vs NAD83 and data transformation

Jordan,Jonathan D jjordan at
Wed Jun 7 09:15:10 EDT 2006

Holli Brandt:

In this vein, just some twobits regarding metadata in such situations--regardless of what technical decision you make for the coordinate system, make sure that a record of the production history (including details about the computation precision as mentioned by J. Sykes in the previous post below) follows the dataset wherever it goes.  This can take the form of a simple readme.txt containing the info, if there is some geodetic/data-collection detail that you cant find a GIS-software readymade metadata-tool to handle.  

In general, when it is necessary to do a quick-n-dirty workaround to get something like a poster graphic made in a timely manner, I wouldnt worry too much about the coord-system submeter fuzziness other than to record the history of what-was-done in the graphic's metadata for my own future reference, but when preparing a submeter real-deal dataset for archive/sharing, the coord-system of the original dataset (and its downstream processed versions and products) needs to be correctly specified regardless of whether this has to mention any original data-collection irregularities such as an awkward GPS setting.  Future users will appreciate the info (in several cases it would have saved me time spent sleuthing the production history of an out-of-house-supplied dataset), and after a few months of water-under-bridge, even the data-producer may well need to consult the detailed record they themselves created.  


Jack Jordan
jjordan at

-----Original Message-----
From: shrug-l-bounces at [mailto:shrug-l-bounces at] On Behalf Of Sykes, John
Sent: Tuesday, June 06, 2006 4:08 PM
To: Brandt, Holli; shrug-l at
Subject: RE: shrug-l: WGS84 vs NAD83 and data transformation

Depends on what you are trying to do.  Here's the issues as I understand it
(I'm still trying to get my mind around it):

1) WGS84 varies from NAD83(86) by about 0.1 mm, however, NAD83 has been
readjusted several times since 1986.

2) Even non-tectonic areas like Florida are moving at about 1 cm per year, so
something that's at X, Y this year, isn't next year, California and Alaska
are zooming all over the globe!

3) The projection of data from State Plane Coordinates to the projection your
map is using (UTM, Albers, etc.) is probably done in double precision, at
best (some software only uses single precision), introducing more millimeter
size errors into your conversion (want to do it right -- get a FORTRAN
compiler which is capable of quad-precision - then you can hold this error to
about 1 micrometer when projecting to Florida DEP Albers by my estimation).

4) Also, the difference between NAD83(86) and NAD83(HARN) is a few
centimeters, and NAD83(HARN) is in the  process of being revised by NOAA/NGS
as we speak.  Actually, this might be good (eventually), since the purpose of
the readjustment is to fit everything to the network of Continuously
Operating Reference Stations (CORS) throughout the USA with about 2 cm
accuracy (or about 1 part in 10,000,000).
5) Finally, State Plane Coordinate systems are a planar system (hint, look at
the name).  Other coordinate systems are usually spherical or, better yet,
ellipsoidal, so the conversion from SPCs to Geographical Coordinates, for
example, are done by a "Best Fit" data method.  The errors from this
conversion can be in the 10 - 20 cm range at worst (in a state like Florida,
which has many fit points, it is much better).

So, what was your question again?

-- John

-----Original Message-----
From: shrug-l-bounces at
[mailto:shrug-l-bounces at] On Behalf Of Brandt, Holli
Sent: Tuesday, June 06, 2006 2:59 PM
To: shrug-l at
Subject: shrug-l: WGS84 vs NAD83 and data transformation

Good Afternoon fellow SHRUGGIES:

Ok we recently received sub-centimeter GPS data in State Plane WGS84 (it was
set up on the unit before they collected the data).  After I make it a
shapefile and go to Define the shapefile, it seems that my only choice in
ArcCatalog is State Plane NAD83.  I have read several posts on the web that
state WGS84 and NAD83 are 'really close' and I can simply lie to ArcMap by
choosing the State Plane NAD83 definition.  My concern is that I will be
losing some resolution of my sub-centimeter data by lying to ArcMap.  

So here are all my questions:
Are WGS84 and NAD83 close enough?
Is there a way to define State Plane WGS84 in ArcCatalog?
Should I lie?
Any other suggestions?

Thanks - Holli

Holli M. Brandt
GIS Project Manager
EnVetCo, Inc.
520 S. Independence Blvd.
Suite 200
Virginia Beach, VA 23462
Office:  (757) 456-0028
Fax:  (757) 456-0399
Cell:  (757) 438-0058
hbrandt at 
SHRUG-L mailing list

SHRUG-L mailing list

More information about the SHRUG-L mailing list