[Proj] How to transform between NAD83 HARN/HPGN and WGS84

Aaron Friesen aaron at cartopac.com
Fri Feb 20 11:30:52 PST 2009


Thank you all for your initial responses...

===== Short follow-up =====

A short follow-up question for those that don't want to read my
long-winded follow-up below:

Do either NADCON or CorpsCon provide datum transforms between a
given NAD83 and WGS84?  The websites seem to indicate that they
are only for transforms among various NAD datums and I do not
know of a NAD83 variant that is truly assumed to be equivalent
to WGS84/ITRF00.  Please correct me if I am wrong on that.

Thanks again,

Aaron

===== Long-winded follow-up that you may want to ignore =====

Hopefully, you do not mind continuing the thread.  I am responding
to multiple people in this message, hoping to keep the thread more
coherent.

Due to non-disclosure limitations, my original description of my needs
only indicated Colorado as a location under NAD83 HARN, thinking I might
be able to extrapolate responses to a more general solution.  I apologize
if I wasted your time with my original query.

Let me expand a bit and re-emphasize one point that seems to not have
been addressed (i.e., conversion to/from WGS84).

I will be working with data originating in various parts of the US.
The data represents various assets in the field where sub-foot accuracy 
is important (or claimed to be important).  The first assumption is that 
existing data has the accuracy it claims to have.  The second assumption 
is that existing data is in the datum it claims to be in.  Both of these 
assumptions may be incorrect, but need to be assumed to be true for this 
discussion.

For existing data, I need to be able to go to the field and locate that 
asset within the combined error of the original reading and the current 
reading from the GPS device used to locate the asset.  For new assets, I 
need to be able to take a reading and add that information to the data set.  
All GPS field readings for this discussion can be assumed to have sub-foot 
accuracy.  If they do not, that is an issue between me and the GPS 
manufacturer claiming such accuracy.  In any case, we are still talking
sub-foot mapping grade GPS, not sub-centimeter survey grade.

The data starts out in various projections (some in State Plane, some
in UTM, some living its entire life as lat/long).  I would also like to
assume that any projection does not add significant error.  Yes, another
assumption that may or may not be valid, but for now...

The problem is that these various data-sets are in various NAD83 incarnations,
and to take the data into the field or bring back new data from the field, one
or more datum transforms are necessary, getting it to WGS84 and back retaining
its original sub-foot accuracy.

The transform to WGS84 seems to be the point not directly addressed.  At least
I did not see it that way.  For sub-foot work, the old assumption of NAD83==WGS84
doesn't do it.

> From: C. Mugnier
> 
> Such questions regarding geodetic-accuracy transformations in Colorado
> are best directed to the U.S. National Geodetic Survey.  NADCON indeed
> does have NAD83 HARNHPGN files, and can provide the necessary
> operations either on-line or through a download.  The same goes for
> current versions of CorpsCon.  Both solutions are not commercial and
> both are free.
> 
> Go to www.NOAA.NGS.gov and you'll get your problem solved for free
> courtesy of the U.S. Government.  A Google search will get you to the
> current version of CorpsCon.

Both NADCON and CorpsCon descriptions talk of conversion between various
incarnations of NAD83, but neither seems to address the transform to/from
WGS84/ITRF00.

> From: Landon Blake
> 
> When you say NAD83 HARN are you using Colorado State Plane Coordinates?
> Or are you dealing with lat/long values?

All lat/long WGS84 in the field.

> I don't think there isn't a huge difference between NAD83 and WGS84,
> but
> I'd think Proj would have a conversion between the two.

Depending on the NAD83 incarnation, NAD83 to WGS84 can differ by several
meters.  Much of the Proj documentation implies that Proj does a null
transformation for NAD83 to WGS84 by default.  That may or may not be
true, but the documentation seems to imply it.

With the +towgs84 switch, I might be able to over-ride that, but I'm not
sure this is sufficient to meet my sub-foot requirements and cannot find
any information on what those parameters might be for a given transform.

> If you are dealing with state plane coordinates a local surveyor can
> probably point you to the conversion formulas that you need. Or you
> might check with the state land surveying association in Colorado.
> 
> If you get stuck, let me know. I can bounce a message off a couple of
> land surveying message boards with Colorado Surveyors listening.

I'm hoping to find a single conversion package (or at least a small
set) to address this problem in the general case.

> I'd be careful using something like Corpscon for data that is sub-foot
> accuracy and needs to stay that way.
> 
> We wouldn't use that program or any like it to convert data that needs
> to be that accurate. I don't know what the purpose of Aaron's data
> conversion is, but he might need to be careful.
> 
> Transforming the location of gopher holes using Corpscon is one thing.
> Transforming building corner positions, or the alignment of a new
> pipeline with Corpscon is another thing entirely.

My needs are closer to the pipeline example than the gopher hole one.

> From: Rich Greenwood
> 
> Corpscon (http://crunch.tec.army.mil/software/corpscon/corpscon.html)
> based on NADCON is a good, free, authoritative tool when working in
> the US.

(Sorry for the repeat) But neither Corpscon nor NADCON seem to be able
to get me to WGS84.

> From: Landon Blake
>
> I was recommending that one consider the purpose of the data
> transformation before deciding to use ANY transformation package. If I
> was building a nuclear reactor I would want to perform field surveys to
> tie into local survey control.
> 
> Aaron is probably fine using CopsCon. I just get nervous when I hear
> "sub-foot" accuracy. Usually when you are dealing with coordinates at
> that level of accuracy you are leaving the GIS arena and entering an
> arena that demands field work to reference local control on the ground.

Rest assured, no nuclear reactors are in the works.  I understand what
you are saying.  Many times, I deal with highly-precise but very inaccurate
data, with GPS readings recorded down to 7 or more decimal places from a
consumer grade device with only 30 meter accuracy.

Thanks again,

Aaron



More information about the Proj mailing list