[GRASSLIST:6021] Re: How deep is a region?

[Roger Bivand]

On Sun, 6 Mar 2005, David Adam wrote:

> Hamish wrote:
> 
> >>I'm trying to set up a region, starting with a Landsat TIFF
> >>image from California.  I have imported the region with r.in.gdal,
> >>which produces a large file. I then close grass60 and reopen in the
> >>new location, at which point I get the errors
> >>
> >>
> >>GRASS_INFO_ERROR(5579,1): default region is invalid
> >>GRASS_INFO_ERROR(5579,1): line 15: <depths:     -1074916636>
> >>
> >>The relevant file for the region is
> >>
> >>DEFAULT_WIND --
> >>
> >>proj:       99
> >>zone:       0
> >>north:      213720.89780907
> >>south:      -6149.10219093
> >>east:       -168613.96484194
> >>west:       -404623.96484194
> >>cols:       7867
> >>rows:       7329
> >>e-w resol:  30
> >>n-s resol:  30
> >>top:        1
> >>bottom:     0
> >>cols3:      7867
> >>rows3:      7329
> >>depths:     -1074916636
> >>e-w resol3: 30
> >>n-s resol3: 30
> >>t-b resol:  1
> >>
> >>--------------
> >>
> >>Most of the fields in the file are explained in the REGION section
> >>of the manual, but I can't seem to find anyplace that defines or
> >>explains what the "depths" field is about.  But until I can get
> >>grass to accept the region, I can't do much of anything.
> >>
> >>All help is much appreciated...
> >>    
> >>
> >
> >
> >rows,cols, and res=30 all look right. I don't know what is causing the
> >depths to go nuts. Try editing the "WIND" file in your mapset to make
> >depths: 1?
> >
> >
> >maybe this works as a temporary workaround--
> >
> >try viewing the GeoTIFF meta-data with "gdalinfo".
> >
> >If the meta-data is bogus, wipe it clean with:
> >
> >tifftopnm map.tiff | pnmtotiff > map_clean.tiff
> >
> >
> >This will wipe all geo-coding, but you should be able to load into a X,Y
> >location ok and at least view the data. May rectify from there.
> >
> >
> >
> >Hamish
> >
> >
> >  
> >
> Thanks, I'll give that a try.
> In response to another reply, this is happening under Grass6.0.0beta2.
> 
> But one of my questions still stands: What does depth represent?  The 
> other terms in the
> DEFAULT_WIND file are defined in the g.region documentation, but not 
> depth.  If I knew
> what it stands for, perhaps I could guess the right value.

This seems to be another side-effect of the precipitate introduction of 3D 
windows in 6.0.0beta*. The code seems to be in lib/gis/adj_cellhd.c, about 
line 249 - it may not be, because working out what is going on is far from 
easy. r.in.gdal seems not to write a depth (about raster/r.in.gdal/main.c 
line 227 and following). So the next operation to access the data tries to 
make one, and:

if (!depth_flag)
{
    cellhd->depths = (cellhd->top - cellhd->bottom + cellhd->tb_res/2.0) / 
        cellhd->tb_res;
    if (cellhd->depths < 0)
    cellhd->depths = 1;
}

says:

> (1 - 0 + (1/2)) / 1
[1] 1.5

for this data. But while top, bottom and tb_res are double, depths is
integer, and the printed output looks very much like an integer that has
fallen over the edge. But maybe the functions called don't visit
G_adjust_Cell_head3(), just something else?

It would be really nice if we had a fixed set of operations that a release 
candidate had to pass before changes were introduced and made available to 
non-developers.

Roger

> 
> Thanks for the help!
> 
> david
> 
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93
e-mail: Roger.Bivand at nhh.no