[Proj] EPSG:28992 to EPSG:23031 conversion

Clifford J Mugnier cjmce at lsu.edu
Wed Oct 8 22:28:04 EDT 2008

I find these sorts of discussions disconcerting in that TR8350.2 is a military publication that offers crude relations that are sufficient for indirect artillery fire control.  That means it's good enough for shooting cannon shells over the horizon to hit military targets with an accuracy good enough for the blast radius of a 155mm high explosive Howitzer round!
GIS applications are not the same venue, and one should recognize the differences.
The Dutch government publishes software that will provide the MAXIMUM possible accuracy for a generalized datum shift within their Kingdom.  Nothing else will be superior, muchless its equal.
Even with such superb accuracy, one must realize that a "datum shift" will DEGRADE the accuracy of ANYTHING submitted for processing as a "datum shift" and it is a best guess.  The only "perfect" way to change datums is to re-compute the original field observations with new starting and ending survey point coordinates.
Short of that, it's a guess, the quality of the guess dependant on the quality of the algorithm - the official nationally-published version being the absolute best guess of them all.
Wanna invade a country and provide indirect cannon fire to support your infantry?  Then use TR8350.2.  However, don't consider it the first choice - it should be the last choice.  I quote it often in my columns when there's nothing better available.  It's better than nothing, but don't compare it to a national geodetic organization's offering!
Cliff Mugnier
P.S.  ED50 was computed by the U.S. Army Map Service in the 1950s to provide a good unified system for future indirect artillery fire control ... if needed.  It is NOT the "last" word in geodetic elegance - it was the best guess for the ARMY ... because WE won the war.


From: proj-bounces at lists.maptools.org on behalf of Paul Kelly
Sent: Wed 08-Oct-08 16:32
To: PROJ.4 and general Projections Discussions
Subject: Re: [Proj] EPSG:28992 to EPSG:23031 conversion

On Wed, 8 Oct 2008, Bart van den Eijnden (OSGIS) wrote:

> Hi Jean-Claude,
> I am using PROJ 4.4.9 so that should not be necessary or am I mistaken?

I think you are mistaken, but would be interested if you could explain
your reasoning behind that understanding. There seems to be a lot of
confusion about the differences in this area in PROJ 4.4.9 and it might
help to understand what is causing this confusion.

Jean-Claude is indeed correct about the second transformation lacking
datum transformation paramters. If I use +towgs84=-87,-98,-121 (a general
solution for ED 1950 across Europe) with your command-line I get the
530201.420293435        5708616.572244617 1.595183260
which is much closer to the official results. I suspect if you found a
datum tranformation set for ED 1950 specifically covering the Netherlands
rather than Europe as a whole you could get very close. (The NIMA TR8350.2
document might perhaps be useful here; I haven't looked.)

Best regards,


> Best regards,
> Bart
> Jean-Claude Repetto wrote:
>> bartvde at osgis.nl wrote :
>>> when I use PROJ.4 to do a conversion between EPSG:28992 (the Dutch
>>> national grid) and UTM zone 31 ED50 I get:
>>> cs2cs -f %.9f +proj=sterea +lat_0=52.15616055555555
>>> +lon_0=5.38763888888889 +k=0.999908 +x_0=155000 +y_0=463000 +ellps=bessel
>>> +units=m
>>> +towgs84=565.2369,50.0087,465.658,-0.406857330322398,0.350732676542563,-1.8703473836068,4.0812
>>> +no_defs +to +proj=utm +zone=31 +ellps=intl +units=m +no_defs
>>> 19465 394814
>>> 530108.861080788        5708613.329898705 -150.818087836
>>> Northing 5708615,754
>>> Easting 530199,479
>>> so especially the difference in easting is too large.
>>> Is there a logical explanation for this difference? TIA.
>> Hi Bart,
>> I think you have forgotten the +towgs84 parameters to do the WGS84 -> ED50
>> transformation.
>> Jean-Claude

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.maptools.org/pipermail/proj/attachments/20081008/70cacb7b/attachment.html

More information about the Proj mailing list