Problem when converting Oregon coordinates

1575
3
Jump to solution
02-10-2017 03:17 PM
JasonDick
New Contributor III

I'm writing unit tests for the Oregon Coordinate Reference System. The ArcGIS conversion does not match sample data provided by Oregon. Is it me or is it ArcGIS? I compared with an open source library based on proj4 and its conversion gives the expected results.

Baker Zone

Input:

44 49 57.80936, 117 48 54.56244

converted to:

44.832724822222225, -117.81515623333333

Expected:

Easting: 41437.60083
Northing: 36980.20833

Got:

Easting: 41438.67098

Northing: 36979.55245

I tried Authority IDs of both 6784 and 6786. I used 4326 for the input coordinates.

Tags (1)
0 Kudos
1 Solution

Accepted Solutions
MelitaKennedy
Esri Notable Contributor

Hi Jason,

I see exactly what’s happening. There’s a geographic/datum transformation occurring because the input is 4326 and the output is based on NAD 1983 (CORS96), 6783. The transformation WKID is 108151,  "WGS_1984_(ITRF00)_To_NAD_1983_CORS96". Here's the definition:

 

   PE_GCS_WGS_1984 \

   PE_GCS_NAD_1983_CORS96 \

   PE_MTH_COORDINATE_FRAME \

   PE_PAR_X_AXIS_TRANSLATION:+0.9956 \

   PE_PAR_Y_AXIS_TRANSLATION:-1.9013 \

   PE_PAR_Z_AXIS_TRANSLATION:-0.5215 \

   PE_PAR_X_AXIS_ROTATION:+0.025915 \

   PE_PAR_Y_AXIS_ROTATION:+0.009426 \

   PE_PAR_Z_AXIS_ROTATION:+0.011599 \

   PE_PAR_SCALE_DIFFERENCE:+0.00062

 

This default transformation is being pulled from either the gtdefaults.json file or its internal equivalent. If you instead use 6783 as the input GCS, you'll get the answers you expect.

Melita

View solution in original post

3 Replies
MelitaKennedy
Esri Notable Contributor

Hi Jason,

I see exactly what’s happening. There’s a geographic/datum transformation occurring because the input is 4326 and the output is based on NAD 1983 (CORS96), 6783. The transformation WKID is 108151,  "WGS_1984_(ITRF00)_To_NAD_1983_CORS96". Here's the definition:

 

   PE_GCS_WGS_1984 \

   PE_GCS_NAD_1983_CORS96 \

   PE_MTH_COORDINATE_FRAME \

   PE_PAR_X_AXIS_TRANSLATION:+0.9956 \

   PE_PAR_Y_AXIS_TRANSLATION:-1.9013 \

   PE_PAR_Z_AXIS_TRANSLATION:-0.5215 \

   PE_PAR_X_AXIS_ROTATION:+0.025915 \

   PE_PAR_Y_AXIS_ROTATION:+0.009426 \

   PE_PAR_Z_AXIS_ROTATION:+0.011599 \

   PE_PAR_SCALE_DIFFERENCE:+0.00062

 

This default transformation is being pulled from either the gtdefaults.json file or its internal equivalent. If you instead use 6783 as the input GCS, you'll get the answers you expect.

Melita

JasonDick
New Contributor III

Thanks very much. I buried myself in the documentation and the web, so I realize why my input GPS was wrong. I have no geo-background; I just use the transformations, so I apologize for the newbie questions.

I understand the different realizations for NAD83, but am still a little confused about how they are implemented in software.I think all the realizations use the same spheroid parameters, so does that mean just the actual transformation parameters between datums are different?

Related (?), OCRS has CORS96 and 2011, but MN County uses HARN. Are these just the ones that happened to be added? Is it possible to have 2011 for MN and HARN for OCRS?

Thank you for your time. Or feel free to point me somewhere else.

0 Kudos
MelitaKennedy
Esri Notable Contributor

Hi Jason, 

No worries. This area of GIS has been getting more, rather than less, complicated over the years! 

A realization is sometimes called a re-adjustment. All the control points in the network are run through a least-squares adjustment (which shakes everything up a little) and then the control point coordinates are updated based up where they are at the time. It takes into account subsidence, tectonic motion, seismic events, etc. Sometimes there are transformations (which are just best fit) between realizations, sometimes not, or the transformations are released years after the realization is put into practice. Pretty much every 'new' geographic coordinate system (datum) that was based on ITRF/WGS84 uses the same ellipsoid, GRS80, but how that surface fits the earth will vary a bit. 

ITRFxx and WGS84 (Gxxxx) have more mathematical definitions and are tectonic plate-agnostic. They vary very little over time. NAD83 and its realizations/re-adjustments are bound to the plates (North American, Pacific for Hawaii and American Samoa, and Mariana for the Guam and the Marianas) to try to minimize the change in coordinates over time. That works well for a lot of places, but the west coast, Louisiana, Alaska, Hawaii all have issues. That also means that the NAD83 realizations are moving away from the ITRFxx/WGS84 (Gxxxx) frames. NAD83 <> / != WGS84 any more!

The county systems like MN and OR got particular realizations versions depending on when they were added. The WisCRS system was added when the HARN re-adjustments will still going on in some parts of the US, but Wisconsin had had their done for a while. We added a NAD83 (2011) set at 10.4.1. The Oregon sets were added in discussion and cooperation of the Oregon DOT. Basically, it's very difficult to add and maintain several versions of the projected coordinate systems based on different realizations so both we (ESRI) and the EPSG (an online registry of coordinate reference systems and transformations) try to limit them. 

Melita