[Proj] MGRS notation in extended UTM zones?

Charles Karney charles.karney at sri.com
Tue Jun 21 08:30:45 EST 2011

On 06/21/11 07:35, Mikael Rittri wrote:
>> Also I can unearth my most recent E-mails with the GeoTrans developers
>> to document what I perceive as the shortcomings of the current version
>> of GeoTrans.
> That could also be interesting.

Here are 3 "bugs" with MGRS conversions with GeoTrans 3.1

(1) Geotrans converts from UTM to MGRS via truncation *except* for a
pesky 0.5m offset

38N 444999   3688999   -> 38SMB4488 (OK)
38N 444999.5 3688999.5 -> 38SMB4589 (no, should be 38SMB4488)

(2) Geotrans converts from MGRS to UTM using the SW corner of the
square.  THis is inconsistent with the truncation rule for the reverse
transform (using the center of the square is the consistent choice).
Repeated forward and backward transformations lead to a drift.

38SMB => (1km accuracy) 32.53 43.94
       => (100km) 38SMA
       => (1km) 31.63 43.95
       => (100km) 38RMV

Maybe this is a misfeature rather than a bug.  The GeoTrans maintainers
say this is a feature.

(3) Geotrans is too lax about latitude bands, e.g.,

38RNL0099 is accepted and converts to 45E 41.54N

However band R covers [24N, 32N) so the north edge of the band is 1000km
away from 41.54N.


More information about the Proj mailing list