[GRASS5] GRASS 5.3/5.7, GDAL and southern hemisphere

[Markus Neteler]

On Tue, Oct 07, 2003 at 11:48:28AM +0100, Paul Kelly wrote:
> 
> On Thu, 25 Sep 2003, Paul Kelly wrote:
> 
> >
> >
> > On Thu, 25 Sep 2003, Markus Neteler wrote:
> >
> > > WARNING: [tuc6] in mapset [tucuman] - in different zone [-4] than current
> > >          region [4]
> >
> > Looking into it further, it appears r.in.gdal will set the zone in the
> > raster header to a negative value if the region is in the southern
> > hemisphere. But g.setproj will never create a negative zone in the WIND
> > file (which has the same format as a raster/CELL header), and indicates
> > that it is in the southern hemisphere by putting south: defined in the
> > PROJ_INFO file.
> 
> And looking even further, it appears that some old coorcnv library
> functions add a large false northing if the zone in the WIND file is
> negative, so it appears this was the original intention of the GRASS
> authors and the authors of g.setproj (which was not developed at CERL if I
> remember correctly) just got this bit wrong.
> 
> So I have changed g.setproj to put a negative zone number in the
> DEFAULT_WIND if the answer is yes to the southern hemisphere question. But
> probably in the future this issue is bound to cause a few more problems to
> people because of the inconsistencies we have observed between different
> GRASS modules.
 
Paul,

thanks for the update. It seems to work OK now (Andrey Kinselev was
so kind to add Transv. Mercator support to LANDSAT/FAST driver in
GDAL which is needed for Argentina data).

So it seems to be complete now.

Markus