[gdal-dev] gdal/ogr/proj and NAD83(CSRS)

Even Rouault even.rouault at mines-paris.org
Thu Jul 23 14:13:25 EDT 2009


Matt,

testepsg EPSG:3579 shows : +proj=aea +lat_1=61.66666666666666 +lat_2=68 
+lat_0=59 +lon_0=-132.5 +x_0=500000 +y_0=500000 +ellps=GRS80 +units=m 
+no_defs 

So it looks like it doesn't have the datum information. I think proj won't 
make any datum transformation then. If your input data is already in NAD83 
CSRS, the end result should be OK though.

Is there a difference between NAD83 and NAD83 CSRS ? (the difference between 
NAD83/GRS80 and WGS84/WGS84 is usually neglectable). If none or not very 
significant, you could try using -t_srs "+proj=aea +lat_1=61.66666666666666 
+lat_2=68 +lat_0=59 +lon_0=-132.5 +x_0=500000 +y_0=500000 +ellps=GRS80 
+datum=NAD83 +units=m +no_defs" (I've added +datum=NAD83 to the testepsg 
output)

Another thing to know, the WKT in a .prj file is in ESRI WKT form, not OGC 
WKT. There are differences in the citation strings and projection parameter 
names. Don't ask me more than that! There has been recently changes commited 
in trunk from ESRI to improve OGC WKT <--> ESRI WKT translation (the output 
in trunk and GDAL 1.6.1 are not the same for testeps EPSG:3579). But that 
doesn't really matter here for your use case, as the .prj file that is 
written by ogr2ogr has no influence in the reprojection process.

Best regards,
Even

Le Thursday 23 July 2009 19:13:42 Matt Wilkie, vous avez écrit :
> Hello All,
>
> I've recently noticed that projecting with ogr2ogr (and I presume
> gdalwarp since they share so much) and using epsg codes adds information
> to .prj files that apparently proj doesn't know about. For example:
>
>    # ogr2ogr -f "esri shapefile" -t_srs epsg:3579 (...)
>
> will write this .prj (note "D_North_American_1983_CSRS98"):
>
>    PROJCS["NAD83(CSRS) / Yukon
> Albers",GEOGCS["NAD83(CSRS)",DATUM["D_North_American_1983_CSRS98",SPHEROID[
>"GRS_1980",6378137,298.257222101]],PRIMEM["Greenwich",0],UNIT["Degree",0.017
>453292519943295]],PROJECTION["Albers"],PARAMETER["standard_parallel_1",61.66
>666666666666],PARAMETER["standard_parallel_2",68],PARAMETER["latitude_of_ori
>gin",59],PARAMETER["central_meridian",-132.5],PARAMETER["false_easting",5000
>00],PARAMETER["false_northing",500000],UNIT["Meter",1]]
>
> However asking proj to list the datums it knows about, there is nothing
> about NAD83CSRS:
>    # proj -ld|grep -i --after-context=1 nad
>    NAD83 GRS80   towgs84=0,0,0
>                  North_American_Datum_1983
>    NAD27 clrk66  nadgrids=@conus, at alaska, at ntv2_0.gsb, at ntv1_can.dat
>                  North_American_Datum_1927
>
> Environment:
>     osgeo4w on WinXPx64
>     Proj version: Rel. 4.6.1, 21 August 2008
>     ogr2ogr version: unknown, but gdal utilities report "GDAL 1.5.4,
> released 2009/01/07"
>
>
> I came across this curiosity while trying to figure out what the
> equivalent proj4 parameters were to epsg:3579. Using
> "+datum=nad83(csrs)" and combinations thereof result in "datum_unknown"
> being written to the .prj file.
> ((if there are no equivalent proj4 strings for the NAD83(CSRS) datum,
> then the proj4 reference files at spatialreference.org are in error and
> should probably be removed, or something.))
>
> So now I'm wondering if transforming a file with gdal/ogr actually
> results in NAD83(CSRS) data, or if it really puts out NAD83 and blindly
> writes out the epsg code?
>
> cheers,




More information about the gdal-dev mailing list