ITRF epoch in compound NAD transformation

5121
5
03-30-2016 02:43 PM
Labels (1)
JoeBryant2
Occasional Contributor II

Hello,

Another GIS specialist up to my neck in Coordinate System information here.

I have a number of questions related to the current realization of WGS1984(G1674).

When transforming between NAD83 (2011) and the original NAD83 (1986 or "CONUS") in ArcMap, compound transformations are listed. I understand this (transform to WGS/ITRF first and then back into the other NAD83 Epoch given the defined NGS conversion values). But in 10.4, the most suitable listed transformation is "WGS_1984_(ITRF00)_To_NAD_1983 + WGS_1984_(ITRF08)_To_NAD_1983_2011". Shouldn't the epoch of each WGS reference frame match? I don't see a choice for ITRF08_To_NAD83 + ITRF08_To_NAD83_2011. There is a choice for "WGS_1984_(ITRF00)_To_NAD_1983 + WGS_1984_(ITRF00)_To_NAD_1983_2011", but it is down the list in suitability, below the HARN transformations.

I consulted the NAD_1983_To_WGS_1984 knowledge base article here: 24159 - Determine which NAD_1983_To_WGS_1984 transformation to use, and I can see that the "WGS_1984_(ITRF08)_To_NAD_1983_2011" transformation is the most accurate (0.030 m versus 0.100 m for ITRF00). But it is not listed for use in the USA, only for the Pacific Ocean. I'm in California (State Plane Zn 2), so maybe that's fine? There is not a "WGS_1984_(ITRF08)_To_NAD_1983" (epoch 1986) transformation listed to be used in the second half of the compound transformation I need to perform. Is that because it is not currently defined by NGS?

Thanks in advance.

0 Kudos
5 Replies
MelitaKennedy
Esri Notable Contributor

The function that puts together the list of suggested geographic/datum transformations is easy to fool. It's looking at extents and accuracies and adding a world base map can throw off its suggestions quite a bit.

When I include ITRFxx in parentheses in a geographic transformation name, it's specifying which realization of WGS84 it should be used with. It's not actually a part of the transformation definition so the projection engine library is not taking (ITRFxx) into account.

I also add "duplicates" of an original transformation but for other, related realizations or re-adjustments like NAD83 (CORS96) and HARN which assume the related realizations can be considered equivalent based on the data accuracy. I also sometimes added "NAD 1983" equivalents for some of the older, ITRF00 transformations assuming that people had data defined as "NAD 1983" but that it was more likely a later re-adjustment. I haven't done that for ITRF08. Now there's also GEOCON and GEOCON11 transformations to move between HARN/NSRS2007/2011/others. We have GEOCON/GEOCON11 v1 in 10.4.0, but not the beta v2 which adds yet more re-adjustments on a state-by-state basis.

Thank you for pointing that there are errors in the knowledge base article, 24159, the extent for WGS_1984_(ITRF08)_To_NAD_1983_2011 should be North America, with the next two, Pacific Ocean (Pacific plate), and Marianas Plate. I ended up talking to the article owner while I was writing this reply and she's opened a draft of it so I can fix it (the current version will remain on the website until it's fixed).

Melita

JoeBryant2
Occasional Contributor II

So if I'm hearing you correctly, "WGS_1984_(ITRF08)_To_NAD_1983_2011" and "WGS_1984_(ITRF00)_To_NAD_1983_2011" are the same underlying calculations, and should give me the same results? Why do they have different listed accuracies in the knowledge base article? Is this just based on the assumption that ITRF00 referenced data is older and thus has probably shifted more? Or perhaps ITRF00 velocities had not been as well established as ITRF08 velocities?

0 Kudos
MelitaKennedy
Esri Notable Contributor

No, they are different transformations. Here are the values:

PE_GT_WGS_1984_ITRF08_TO_NAD_1983_2011 \

   "WGS_1984_(ITRF08)_To_NAD_1983_2011" \

   PE_GCS_WGS_1984 \

   PE_GCS_NAD_1983_2011 \

   PE_MTH_COORDINATE_FRAME \

   PE_PAR_X_AXIS_TRANSLATION:+0.99343 \

   PE_PAR_Y_AXIS_TRANSLATION:-1.90331 \

   PE_PAR_Z_AXIS_TRANSLATION:-0.52655 \

   PE_PAR_X_AXIS_ROTATION:+0.02591467 \

   PE_PAR_Y_AXIS_ROTATION:+0.00942645 \

   PE_PAR_Z_AXIS_ROTATION:+0.01159935 \

   PE_PAR_SCALE_DIFFERENCE:+0.00171504

PE_GT_WGS_1984_ITRF00_TO_NAD_1983_2011 \

   "WGS_1984_(ITRF00)_To_NAD_1983_2011" \

   PE_GCS_WGS_1984 \

   PE_GCS_NAD_1983_2011 \

   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

The transformations occur in 3D Cartesian space, and you can see the differences between the two are pretty small.

CliffHoeffner1
New Contributor II

Hi Melita,

Can you also put in a request to change this across the product platform? This also comes up in Collector with the user profile and is misleading to customers. In other words, Collector says this transformation ITRF08 TO 2011 is for the Pacific Ocean while ITRF00 is for US Conus.

Thanks!

MelitaKennedy
Esri Notable Contributor

Hi Cliff, 

I fixed the extent issue in 10.6 / Pro 2.1, so it should get picked up in Collector soon.

Melita

0 Kudos