I have data in two projected coordinate systems (PCS), 1) NAD 1983 HARN StatePlane California III FIPS 0403 (US Feet), and 2) NAD 1983 StatePlane California III FIPS 0403 (US Feet). Using HARN, nonHARN going forward for these two.

The ArcGIS Pro Map viewer coordinate system is set to HARN. For nonHARN to HARN, it is using the datum transformation WGS 1984 (ITRF00) to NAD 1983 + NAD 1983 HARN to WGS 1984 (2).

I am converting all nonHARN data to HARN for my project. When I use the Project tool, the transformation recommended is the NAD_1983_To_HARN_CA_N. I've used this. When all of my data is converted to HARN, there is about a 2.5 foot offset between all the HARN and nonHARN datasets.

When I change the Map viewer datum transformation for nonHARN > HARN to the NAD_1983_To_HARN_CA_N, which matches the Project tool recommended transformation, all the datasets align with no material offset (fractional inch).

I am trying to understand what is going on here.

First, is there a more correct transformation in this case?

Second, could there really be that large of a variation between the datasets, or is this a case of misused transformations?

This variation may not really change the outcome for my project, but I'd rather like to figure out why this is occurring.

I started with a fresh Map and just added a few datasets in nonHARN and HARN and the system defaulted to the NAD_1983_To_HARN_CA_N transformation for the map. Apparently, in originally working with a lot of WGS based data as well, it set the transformation at the start (to the NAD83 > WGS > NAD83 HARN) and didn't revert even when all my data in the Map was either NAD83 or NAD83 HARN based. This isn't a true answer to the issue, but I suppose it could boil down to an error due to incorrect transformation, as mentioned above. I still will take any inputs as to what may have been occurring here.

  • 2.5' sounds more like a US Survey Foot/International Foot problem.
    – Pointdump
    Dec 20, 2022 at 18:32

1 Answer 1


I bet you have a world basemap layer. In a map, the union of all data layers is used for the "data extent". That information is sent to our gtlist (geographic transformation list function) or hvtlist (horizontal/vertical transformation list) which returns a sorted list of transformations based on how well extent of the 1 to 2 transformations for gtlist or up to 4 transformations for hvtlist match the data extent. If there are more than one transformation path that have the same percentage overlap, the accuracies are used as a tiebreaker. If they also match, I believe we list them by well-known ID order.

Anyway, a "world" extent is getting sent to the function. In Project, the layer extent is being used, thus a much more accurate transformation is applied.

Note: I work for Esri on the coordinate systems and transformations team. We've been discussing dropping any basemaps from the data extent; just haven't done it yet.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.