Raster Issues in MapServer 4.4.1

[Frank Warmerdam]

On 5/17/05, Chris Helm <helmster_99 at yahoo.com> wrote:
> Frank,
> I've placed a simple mapfile and the tif in a
> directory at:
> http://glims.colorado.edu/~chelm/glacierdata/example/

Chris, 

OK, the problem is that PROJ.4 maps some location on the 
other side of the globe into the UTM 8 zone you are using. 
This does seem to be some sort of wrap around instability 
issue. 

warmerda at gdal2200[305]% proj  +proj=utm +zone=8 +ellps=WGS84
5 60
-339686.29      7396243.98
10 60
-1313079.75     5684270.49
0 60
512501.80       8653308.24

warmerda at gdal2200[306]% proj -I +proj=utm +zone=8 +ellps=WGS84
512501.80       8653308.24
134d27'48.133"W 77d57'14.884"N
-1313079.75     5684270.49
159d42'17.209"W 48d35'37.222"N

What we can see from the above is that some locations such as 10E,60N
map to locations fairly near the UTM 8 zone, even though when you 
convert that location back to UTM you get a completely different longitude

10E,60N -> -1313079.75E 5684270.49N -> 159d42'17.209"W 48d35'37.222"N

I'm not exactly sure what to do about this.  One possible approach is to 
make the UTM (really transverse mercator) fail in cases where the foward
and reverse projections are so divergent.   

Would you mind filing a formal bug against PROJ.4 for the above "proj"
level issue?  Then I can take it up on the proj.4 mailing list for possible
resolution.

In general the problem is that the math for some projections becomes
very unstable or at least weird when you get far from the region of validity
and you can get the odd world-wrapping-around sorts of issues. 

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent