Category Archives: Engineering

NAD83 vs WGS84

It has been a while since the last time I spoke with a surveyor who had a clear idea about his supposed to be deliverables should be. Their central misunderstanding was about the reference frames. Collecting the data apparently is something easy to do. Which reference frame they have to use, its the problem. I don’t know why most of them seem to be thinking that they only available reference frame for any job in their portfolio is the NAD83.

I want to remind the most significant differences between NAD83 and WGS84. The NAD83 is a goddamned FIXED reference frame in time AND space where the WGS84 is NOT. And using a fixed reference frame designed for cadastre and taxation purposes can have huge implications if you also try to use it for inertial mapping (create a map based on Inertial Measurement Units). Let me tell you how this is working. First of all, your surveying equipment is getting the GPS coordinates in WGS84. You can’t change this. Now, for a reason that is a bit blurry, you feel the necessity to transform those coordinates to bring them into the NAD83 space even if you map with an inertial tool. For this reason, you have to apply a 14 parameters transformation. 3 for the XYZ, 3 for the rotations around XYZ axis, 3 for the velocities along XYZ, 3 for the angular velocities around XYZ, 1 for the XYZ scale and 1 for the angular scale. Total, 14 parameters.

You can find this table at the NGA.STND.0036_1.0.0_WGS84 file, page 7-5.

The result of this transformation will give you the NAD83 coordinate transformed from WGS84, with an absolute accuracy of about 1 meter (NAD83 2011). If you have to go ahead to “reverse” the data back to their original state of WGS84 you are about to hit a wall. The transformation is not linear. You are going to introduce an additional error of maybe 1 meter… again. But you promise “less than a centimeter accuracy” and at the best, you are now 1-meter off.

Let’s talk about accuracy a little bit. Every time you ask a surveyor about terms such as accuracy and precision, they will tell you the same thing “less than a centimeter”. This accuracy is achieved between your base and your rover. It is what we call RELATIVE accuracy. It has NOTHING to do with the absolute accuracy of “where you are over the planet”. The FAA positional specification for FATO operations (Final Approach and Take Off) for jets is about 1-meter. Even if you expose your base for 10 hours static observation (DGPS and WAAS included), you are about to go down 0.4m to 0.6m ABSOLUTE positional accuracy. Add here the 1-meter from the transformation process between NAD83 and WGS84 and you get 1.4m to 1.6m absolute accuracy.

Conclusion:  Just use the proper reference frame for the job and if you have the problem to find which one is proper, use the damned WGS84.