Which version of NAD83 is assumed in geographic transformation?

FrancesBiles · ‎01-07-2016

I have some points in NAD83 (1986). In fact, A LOT of our GIS data is in NAD83 (1986). I am using Desktop ArcMap v 10.2.2.3552. I want to transform the points to ITRF00.

If I choose: WGS_1984_(ITRF00)_To_NAD_1983 + ITRF_2000_To_WGS_1984 for the geographic transformation method, which flavor of NAD83 is being assumed? Everything I found suggests that NAD83 (CORS96) is being assumed. If that is the case, how does one transform between NAD83 (1986) to ITRF00? To ITRF08? (Or the most current version of WGS84?). The GIS data covers southeast Alaska.

From what I have read, it looks like depending on what version of ArcMap one is using, the name "NAD83" could refer to completely different versions of NAD83. For example, in earlier versions of ArcMap, the transformation NAD_1983_To_WGS_1984_5 assumed that coordinates were in NAD83 (1986), but in current versions of ArcMap (10.x) NAD_1983_To_WGS_1984_5 assumes that coordinates are in NAD83 (CORS96). Is that correct? If so it would be VERY HELPFUL if datum names/labels always had their "version" attached. Does WGS84 in that transformation refer to the original WGS84 or WGS84(G1674), or?

Is there a comprehensive "cheat sheet" that crosswalks ArcMap versions with the datum versions assumed for all datum and geographic transformation names that involve NAD83, WGS84, ITRF reference frames?

DanPatterson_Retired · ‎01-07-2016

Melita Kennedy .... would be the person to address this to. I have flagged this post for her attention.

DanPatterson_Retired · ‎01-08-2016

some interim links

a big list
- Spatial Reference List -- Spatial Reference
particularly as they apply to Canada
- Confluence Mobile - QPS
- The Canadian Spatial Reference System (CSRS) | Natural Resources Canada

NeilAyres · ‎01-08-2016

The factoryCode for NAD83 (1986) is 4269.

And WGS_1984 is 4326.

So to list the available transformations I did this

sr1 = arcpy.SpatialReference(4269)
>>> sr1.name
u'GCS_North_American_1983'
>>> sr2 = arcpy.SpatialReference(4326)
>>> sr2.name
u'GCS_WGS_1984'
>>> trans = arcpy.ListTransformations(sr1, sr2)
>>> len(trans)
298

But a great deal of those 298 are composite transformations and I treat those with deep suspicion.

So listing out only those without "+" in the name gives me :

>>> for tr in trans:
...     if "+" not in tr:
...         print tr
...         
WGS_1984_(ITRF00)_To_NAD_1983
NAD_1983_To_WGS_1984_5
NAD_1983_To_WGS_1984_1
NAD_1983_To_WGS_1984_4
NAD_1983_To_WGS_1984_3
NAD_1983_To_WGS_1984_2

So that first one might do want you need to do.

NeilAyres · ‎01-08-2016

But looking into the geographic_transformations pdf I see this.

WGS_1984_(ITRF00)_To_NAD_1983 108190 CF 0.9956 -1.9013 -0.5215 0.025915 0.009426 0.011599 0.00062

WGS_1984_(ITRF00)_To_NAD_1983_2011 108354 CF 0.9956 -1.9013 -0.5215 0.025915 0.009426 0.011599 0.00062

WGS_1984_(ITRF00)_To_NAD_1983_CORS96 108151 CF 0.9956 -1.9013 -0.5215 0.025915 0.009426 0.011599 0.00062

WGS_1984_(ITRF00)_To_NAD_1983_HAR 108282 CF 0.9956 -1.9013 -0.5215 0.025915 0.009426 0.011599 0.00062

These are all the same thing.

So now I am not so sure

DanPatterson_Retired · ‎01-08-2016

They did say ... using a 7 parameter transformation... they failed to mention whether it was the same one ... hmmm ... what's a meter or so, province-wide

MelitaKennedy · ‎01-08-2016

We do have some "named" realizations for NAD83 in the US - HARN, CORS96, NSRS2007, 2011. "NAD 1983" should only be used for the 1986 version, but some people use it for everything. When I can find an ITRF/NAD83 transformation from the NGS (National Geodetic Survey) CORS website, I put it, but also put "equivalents" where I include it for HARN and CORS96 or NSRS2007 and 2011, and possibly "1986."

WGS84 is a generic definition. We've been debating adding the specifically named versions, but there's very little public information about how to convert between them. Also, what do we do with the existing "to WGS84" transformations? Do they apply also to any of the specific WGS84 realizations?

Specifically for Alaska, I'm not sure I have a full solution for you. There was no HARN grid published to convert from NAD83 (1986) to NAD 1983 HARN that I'm aware of. GEOCON v2.0 will convert from the Alaska FBN re-adjustment of 1992 to NSRS2007 and then from NSRS2007 to 2011. That still leave a gap between 1986 and 1992.

You might try contacting the NGS state advisor for Alaska to see what she recommends.

Melita

FrancesBiles · ‎01-11-2016

Thank you everyone for the useful information. First an FYI, and then a question.

Regarding the NAD83 version used for the WGS_1984_(ITRF00)_To_NAD_1983 transformation, I was able to find the answer from clues given by Neil and Melita. As Neil pointed out, the 7 parameter values are the same for all four ITRF00 to NAD83(1986), CORS96, 2011, and HARN transformations. I wondered which version of NAD83 these parameters are associated with. According to this NGS CORS page, the 7 parameters are specifically for ITRF00-to-NAD83 (CORS96).

Q1) I'm not sure what the + ITRF_2000_To_WGS_1984 is indicating in the WGS_1984_(ITRF00)_To_NAD_1983 + ITRF_2000_To_WGS_1984 composite transformation. Why is this needed/what is it doing? Could someone please explain?

-Frances

DanPatterson_Retired · ‎01-11-2016

Some minor shifts... the wgs84 associations are listed here, but I sent this before

Confluence Mobile - QPS

MelitaKennedy · ‎01-12-2016

It's converting from NAD83 to WGS84, then WGS84 to ITRF2000. The 2nd transformation's parameters are zeroes. I call it a "bookkeeping" transformation. Even though the first transformation says WGS_1984_(ITRF00), it's between WGS84 and NAD83. The (ITRF00) is to clue that the WGS84 version is the approximately ITRF00 realization.

Melita