I just projected a feature class originally in Web Mercator (WKID 3857) to WKID 6318 (ESRI calls this "GCS_NAD_1983_2011") using ESRI's projection tool, and the same feature class to WKID 6318 (OGR calls this "NAD83(2011)") using ogr2ogr. They are very different: if I put them side by side in ArcMap they are 38" off in some places. What's more, I added the GCS_NAD_1983_2011 shapefile (the ESRI-projected shapefile) to ArcMap first and when adding the NAD83(2011) shapefile (the ogr-projected shapefile) ESRI prompted to use a transformation since they are different datums.

This is unexpected: I expected WKID 6318 to be the same between the two (ESRI and ogr). Why would they be so different?

This is the full description of GCS_NAD_1983_2011/ESRI (looking at the 'XY Coordinate System' tab in the Properties window of ArcCatalog): GCS_NAD_1983_2011 WKID: 6318 Authority: EPSG

Angular Unit: Degree (0.0174532925199433) Prime Meridian: Greenwich (0.0) Datum: D_NAD_1983_2011 Spheroid: GRS_1980 Semimajor Axis: 6378137.0 Semiminor Axis: 6356752.314140356 Inverse Flattening: 298.257222101

And this is the same info for NAD83(2011)/ogr: NAD83(2011) Authority: Custom

Angular Unit: Degree (0.0174532925199433) Prime Meridian: Greenwich (0.0) Datum: D_NAD83_National_Spatial_Reference_System_2011 Spheroid: GRS_1980 Semimajor Axis: 6378137.0 Semiminor Axis: 6356752.314140356 Inverse Flattening: 298.257222101

Everything is the same except for 'Datum'.

Why would that be? Which one is the "authoritative" datum??

  • 1
    The important part is the SRID (WKID). Both are identical. Even though the datums have different names their parameters would be the same... and a rose by any other name... is still a rose! – Michael Stimson Jul 20 '17 at 0:04
  • Thanks for the response Michael. I get that the SRIDs are the same, which is why I would expect the projected layers to be the same. But they're not: when I add both layers to ArcMap I expect they would exactly overlay each other (or nearly so, within centimeters at most), but that's not what's happening: they're a few feet off in places. – crb Jul 20 '17 at 0:20
  • Yes that is odd, I suspect it has more to do with the transformation that is applied. I'm not an expert in northern hemisphere projections but from my own experience (AGD66 to GDA94 NTv2) the Esri transformation is more accurate compared to the OSR default or approximate transformation; I must admit though I'm not sure how to specify a datum transformation file for OSR. It would be good if you could compare both outputs with a known to be accurate dataset to prove in your situation which one is more correct. – Michael Stimson Jul 20 '17 at 0:28
  • My guess is that on the OGR-side, there's a proj.4 string that includes a +TOWGS84 parameter...or not. Was a transformation used when you converted from 3857 to 6318 in ArcGIS? If the same tfm wasn't used in OGR--different output. – mkennedy Jul 20 '17 at 19:50
  • @mkennedy: good point. The WGS_1984_(ITRF08)_To_NAD_1983_2011 transform was used for the Esri projection, and none for the ogr2ogr command: ogr2ogr -f "ESRI Shapefile" C:\temp\bridges.shp C:\temp\nad83.shp -s_srs EPSG:3857 -t_srs EPSG:6318 – crb Jul 21 '17 at 15:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.