Transformations between various NAD83 and WGS84 realizations

[GennadyMogilevich]

I realize there are already many questions and answers out there related to this issue.  However, I continue having a difficult time pinpointing exactly which transformations I should use. 

Essentially, I have GPS data that was post processed against the local CORS stations in Pathfinder Office using the reference position from base provider option, which depending on the date and station would be ITRF00 or ITRF08 and exported as NAD83 CORS96 and in later years as NAD83 (2011).  Now, I have a Collector app that is using SBAS corrections, so the data is in ITRF 2008 (essentially equal to WGS84 within 10cm/4"), if I'm not mistaken.  This means that I have data in 3 different datums or realizations in regards to the NAD83 GPS points. 

I'm using ArcGIS Desktop 10.3.1, and by looking at the transformations list published with each release of desktop, it appears that the transformations between NAD_1983_HARN, NAD_1983_CORS96, and NAD_1986 (original/1986) are equivalent or simply "bookkeeping".  If this is the case, then there's really no method or transformation in the Project geoprocessing tool that would enable me to reproject my NAD_1983_CORS96 GPS data into NAD_1983_2011, correct?

As far as Collector goes, if the data is coming in as ITRF 2008 based on the SBAS correction source, if I overlay it with NAD_1983_CORS96 or NAD_1983_HARN, etc., what's the best transformation to use? Or does it really not make a difference if they're all treated as equal as far as ArcGIS is concerned? 

Finally, maybe I just haven't come across the details on this yet, but when ESRI references GCS_WGS_1984, which WGS 1984 are they referring to?  Is the reference to G1762, which is the latest realization, or is this in reference to G1674 or G1150? I realize that these are all based on ITRF and should all be within 10cm/4", but I'd like to #1 understand what ESRI is referring to and #2 be able to select the most accurate transformation settings for the data.

Thank you in advance,

Henry M.

[MelitaKennedy]

Hi Henry, 

So far we are not differentiating between the WGS84 realizations. That will change in an upcoming release, as EPSG plans to add them. 

I sometimes link a ITRFxx version with WGS84 in the transformation name, as a hint to which realization the transformatoin applies to. 

At 10.4, we added support for GEOCON v1.1 which converts between the "last" HARN re-adjustment and NSRS2007 or NSRS2007 to 2011. GEOCON is file-based, so I don't think Collector supports them. The files are also in a separate install, 'ArcGIS Coordinate System Data' or 'ArcGIS Pro Coordinate System Data' as the data files were getting too large to include with the core software. We don't have equivalent equation-based transformations because NGS hasn't released any.

These are probably the best we have:

"WGS_1984_(ITRF00)_To_NAD_1983_2011"
"WGS_1984_(ITRF08)_To_NAD_1983_2011"

You could take HARN as approximately equivalent to CORS96 and chain together the 1st transformation above, with 

"WGS_1984_(ITRF00)_To_NAD_1983_HARN"

Or just use them to get the HARN data up-to-date. Another possibility is to run the GEOCON transformations (there's a beta v2.0 as well) using the NGS converters to see what the offsets look like in your area of interest. 

Melita

[GennadyMogilevich]

Melita,

Thank you for providing the information regarding the 10.3.1 WGS84 realizations and transformations between NAD_1983_2011.  In some additional testing conducted using data points post-processed through Pathfinder Office (ver 5.20) using local CORS stations and exporting the data in both PCS Virginia_StatePlane_North_NAD_1983_2011_FT and in WGS 1984, it became apparent that the correct transformation in this case would be the "WGS_1984_(ITRF08)_To_NAD_1983_2011".  The "WGS_1984_(ITRF00)_To_NAD_1983_2011" introduced a slight shift of 1/4 inch (not that significant for sub-meter data collection). 

I assume this means that Esri does differentiate between ITRF00 and ITRF08, correct? 

If so, does Esri also differentiate between ITRF08 and WGS 1984?  Since we use Collector app with SBAS corrections for our new mobile applications, I want to make sure that I can use the transformation between ITRF 2008 and WGS 1984 at 10.3.1.

Thanks again!  I really appreciate your help.

Henry

[DanPatterson_Retired]

Melita, is there any documentation which easily gives an 'error' estimate if you don't use the transformation.  I know there are tech docs, but I am looking more for something more geared to intro students who ask 'how many meters will we be off?'

[MelitaKennedy]

I think the only doc I've written on it is on the forums. Any recent NAD83 (NSRS2007, 2011, probably CORS96 because it was used later than 1996 and possibly HARNs) versus WGS84/ITRF is now up to 1 m or more. 

Within NAD83, biggest jump was from original to the first HARN within a state. Some states have a 2nd or 3rd HARN / FBN adjustment which will be smaller. NAD83 original to first HARN, I'm guessing was centimeters. I checked Redlands, Californi which was about 6.6 cm. North Bay above San Francisco around 4 cm. Central Ohio, 0.263 m, New Orleans, 0.5 m. Subsequent re-adjustments should be less, normally.

WGS84/ITRFxx changes should be centimeters or less. 

Melita

[JoelCusick]

Henry,

May I chime in on the first layer of the onion that I see in this original question.  I'm experienced in Trimble workflows including Pathfinder Office and have fought my way out of the datum tidal flats more than i can mention.  So, if its not too late, or too far back, I have two comments on the precise terms you use when you wrote what your doing in PFO. If PFO output (the shapefiles you mention are considered "truth"), then I have a suggestion to do the right thing when it comes to "holding" a CORS solution. Then, when you get your PFO shapefile, you know you did everything you could prior to the loading into arcGIS.  

True accuracy is assessed only when you occupy truth better than your system.  That is a modern OPUS observed benchmark.  Try to find one nearby using NGS explorer or OPUS shared solutions.  you'll see the NAD83 2011 coordinate (assume you are trying to "get there" is posted.  Thats truth.

second level of truth is your Correction settings.  At PFO v. 5.2, pick the closest CORS station.  Next, when stepping thru the wizard your two radio button choices in the Diff Wizard will allow you to either "hold what Trimble has produced - thats the "Use position from base provider" or "Use position from base files".  The former is NOT a published CORS coordinate, and I never use.  The second option "base files" is the actual NAD83 (2011) coordinate for the Rinex file embedded by NGS.  Use that option at all times. Then, that COR file at the conclusion of Differential correction is in NAD83 (2011).  On export, you do not shift. You use a Trimble "bookkeeping" shifter (NAD83 (CONUS) which does nothing.  Define your shapefile with the correct PRJ (provided by ESRI with the datum tag 2011 and your exported shape is in the highest quality form.  I never use ESRI to shift data coming from PFO. Why?  Because you can harvest data directly from CORS published position  - thats the best approach in my view.  

So, this is going to have different results (likely about 90cm) depending on where you are in your world, relative to the WGS84 output you were exporting before, but your ready to load with no shifting in GIS.

Good luck. Hope this helped, or thoroughly confused you. 

Just know, Trimble has a mistake in the CBS list which is often overlooked and one path down PFO will lead you to holding NOT the IGS08 position for a CORS, but in reality, ITRF00 (1997) position.