[GRASS-dev] [bug #4598] (grass) g.region from raster wrong when crossin 180 deg

[Request Tracker]

this bug's URL: http://intevation.de/rt/webrt?serial_num=4598
-------------------------------------------------------------------------

Subject: g.region from raster wrong when crossin 180 deg

Platform: Mac OSX
grass obtained from: Trento Italy site
grass binary for platform: Compiled from Sources
GRASS Version: CVS 2006_05_27

an example is the best way I can describe this.  I have an SRTM tile at S18W180, that is it covers 
180W-179W/18S-17S.  Due to the cell-center alignment of SRTM, it extends 1.5" outside this area, 
so the real region of the tile is:

north:      16:59:58.5S
south:      18:00:01.5S
west:       179:59:58.5E
east:       178:59:58.5W
nsres:      0:00:03
ewres:      0:00:03
rows:       1201
cols:       1201

I want to merge this with the nex tile over, S18W179, so I set the region from the two rasters:

g.region rast=S18W180,S18W179

now the region is wrong:

north:      16:59:58.5S
south:      18:00:01.5S
west:       179:00:01.5W
east:       178:59:58.5W
nsres:      0:00:03
ewres:      0:00:03
rows:       1201
cols:       432001

The west edge should not change (179:59:58.5E), and the east edge should be another degree east 
(177:59:58.5W) and the column count should be calculated from that, not spanning the globe 
(2401).  Based on the above wrong E/W, the column count is still wrong, it should be 1 (but it 
displays correctly in a monitor).

g.region rast=S18W180

gives the correct region for the single raster.

g.region rast=S18W180,S18E179

for a regionwith the tile to the west instead, gives the correct region.

Manually setting the region to what it should be also gives the wrong region:

g.region w=179:59:58.5E e=177:59:58.5W -p
projection: 3 (Latitude-Longitude)
zone:       0
datum:      ** unknown (default: WGS84) **
ellipsoid:  wgs84
north:      16:59:58.5S
south:      18:00:01.5S
west:       179:59:58.5E
east:       177:59:58.5W
nsres:      0:00:03
ewres:      0:00:00.000007
rows:       1201
cols:       1037234401

The extents are right, but the ewres/cols is wrong.  Forcing the resolution finally fixes the problem:

g.region w=179:59:58.5E e=177:59:58.5W -p -a res=0:00:03
projection: 3 (Latitude-Longitude)
zone:       0
datum:      ** unknown (default: WGS84) **
ellipsoid:  wgs84
north:      16:59:57S
south:      18:00:03S
west:       179:59:57E
east:       177:59:57W
nsres:      0:00:03
ewres:      0:00:03
rows:       1202
cols:       2402

But only for manually setting e/w, with rast= it doesn't help.

-------------------------------------------- Managed by Request Tracker