I've been working in GIS for about a year and have a functional grasp of Projections and Transformations from within ArcGIS (10.0). Geosoft Oasis is another program we use. In working with our data we primarily use NAD83 Zone 13 and 14 North. Although WGS84, NAD27, and NAD83(CSRS) often join the mix. With each of these different Coordinate Systems, there are many transformation options to go from one to the other.

I would really like to have a better understanding of how much of an error is created in each transformation. To do that, could someone tell me the best way to test this out?

As an example, we often have data that spans Zones. What happens when we using Zone 13 projection in our arc project in an area that should be Zone 14, but we don't want to break the data up? How much of a penalty is there if I work in WGS84 and then transform the data into a NAD83 projection (or a NAD27 one)? Are the errors cumulative? If I switch back and forth 50 times will my data move? Or if I switch between 3 Projections using 3 Transformations?

I'm not sure of a good method for finding an answer to these questions. I would like to be able to work through it myself instead of taking someone's word for it, but that might be a more technical task than I'm able to do...

Any help is appreciated, thanks,


  • 1
    Here's a thread about the distortion of data that crosses zones: gis.stackexchange.com/questions/31701/…
    – kenbuja
    Commented May 31, 2013 at 19:36
  • 1
    You put up several scenarios and ask what will happen, but have you tried running them and comparing the results with your starting coordinates? I think trying to explain any discrepancies observed from what you expected could lead to one or several more focussed questions.
    – PolyGeo
    Commented May 31, 2013 at 23:16
  • I suppose that's what I'm getting at. I'll mess around and see if I can observe discrepancies and then come back. Right now I'm just guessing they are there, and not sure how to flush them out. I'll be back in a few days with what I come up with, thanks.
    – cndnflyr
    Commented Jun 3, 2013 at 16:10

2 Answers 2


NADCON errors (NAD27) to (NAD 83) are in the range of 12-18cm http://www.ngs.noaa.gov/TOOLS/Nadcon/Nadcon.shtml. I think this answera a large part of your question. So yes multiple shifts can start to add up for high (very) quality survey data but it will take many NADCON shifts to start to meaningfully cause problems in most datasets.

The published accuracy of WGS 84 to NAD 83 for continental North America is 2 meters (http://support.esri.com/cn/knowledgebase/techarticles/detail/24159) with values lower than 2 in southern states and towards the top end in northern states. In Hawaii (4m) it increases as do other non-continental states. So yes multiple transformations can cause aggregated error.

The cross zones part of your question is fully answered at gis.stackexchange.com/questions/31701.

I think between these 3 sections your answer is complete.

  • Yes, thank you for the links you provided. They have answered my questions well.
    – cndnflyr
    Commented Jul 31, 2013 at 21:47

The quickest way to estimate errors is to set up some sample points (say five, more is better) with exact coordinates, and perform round trip calculations.

Datum A -> transformation X -> Datum B -> transformation Y -> Datum A

Then you directly can compare coordinates before and after.

Alternatively: WGS84 -> transformation X -> Datum A ; WGS84 -> transformation Y -> Datum A

  • That's a great idea. Thank you. I'll employ this in the future.
    – cndnflyr
    Commented Jun 4, 2019 at 21:31

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.