[PROJ] Explaining geodesic distance discrepancies

[Charles Karney]

I believe PROJ does have some way of computing the scale of a
projection; but it uses numerical differencing.  GeographicLib computes
the scale of TM from an analytic expression, so it's probably more
accurate.  The command I used to get the scale (using GeographicLib) is:

     echo 0.5 |
     GeodSolve -I 31.4617 -83.4849 31.4704 -83.4781 -F -p 10 |
     cut -f1,2 -d' ' | GeoConvert -c -p 10
=>
     -1.2959245331415 1.000285683174366

The GeodSolve expression gets the midpoint of a geodesic.  GeoConvert -c
returns the meridian convergence and scale for the UTM projection.

The command line utility geod uses the same code as pyproj.Geod.  So
unless you have extra versions of geod lying around they should return
the same results.


On 5/15/20 5:00 PM, Dima Kogan wrote:
> Thank you very much. That's VERY useful info. How did you get that scale
> factor? Is there a proj tool and/or function call?
> 
> I have just tried to run the "geod" cmd on another box with another proj
> install, and it matches up over there. Which is probably enough for me
> to go figure this out. We DO expect the "geod" cmdline and the
> pyproj.Geod results to match 100%, right?
> 
> Thanks again.
> _______________________________________________
> PROJ mailing list
> PROJ at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/proj
>