Having re-read this whole thread
> Yes, Star*Net, for example, might report that the GPS-derived azimuth between the two control points from which you began your traverse has an uncertainty of, say, +/-0'10" at 95% confidence. Even if your traverse measurements were perfect (instead of including measurements made on Mondays and Fridays), that initial uncertainty will generate a closing error all by itself that is quantifiable along with the additional uncertainty contributed by the Monday and Friday angles in the traverse.
:good:
I have seen it, back in the old days running large traverses, not sure why it's East-West however. You need to check the Laplace corrections. I used to have to apply them in some environments.
Hopefully, there is a program these days to do it.
His stated raw closure of 1:16,000, if all due to azimuth, is 12.9 arc seconds. I sampled a few tri station data sheets around Atlanta and found Laplace for most of them in the range of only 0 to -2 seconds. Maybe in a larger area it might be somewhat more.
And Laplace is only significant if you have a significant tilt to the sights while measuring the horizontal angle, such as for Polaris sights.
So I don't think Laplace accounts for his problem.