[GRASSLIST:1305] Converting sites to rasters

GRASS GRASS user
1

I'm trying to use grass5b10 with bathymetry data and am having some trouble getting started.

I can import my data into a sites file (approx 150,000 sites) but I haven't been able to convert it to a raster.

s.to.rast gives me:

     ERROR: error while writing to temp file

Has anyone got any ideas what be causing this?

s.surf.idw seems to take about 1 CPU hour per site on my 400MHz PII (I killed it after running all weekend).

s.surf.rst also seems to be taken an enormous amount of time.

My data is basic lat,lon,depth data.

I suspect I'm trying to use the wrong tools for the job, can anyone suggest what I should be using?

Regards
Gordon

2

Gordon,

I'm not particularly experienced at this, but I had a similar (but less
extreme) problem the first time I tried to generate a surface from sites
data. I think all of your problems here might be explained *if* you did
the same thing that I did.

The raster map that is created by all of these procedures is determined by
your current region and mask settings. Make sure that your current region
is no larger than you need it to be and that the grid resolution is no
smaller than you need it to be.

If your region settings aren't right then you may be asking grass to give
you far more than you need or want.

I think my first choice of methods would be to generate a raster map from
the sites file using s.to.rast (assuming you can get it to work) then
generate the interpolated surface using r.surf.idw or r.surf.idw2.

Roger Miller

On Wed, 20 Dec 2000, Gordon Keith wrote:

I'm trying to use grass5b10 with bathymetry data and am having some
trouble getting started.

I can import my data into a sites file (approx 150,000 sites) but I
haven't been able to convert it to a raster.

s.to.rast gives me:

     ERROR: error while writing to temp file

Has anyone got any ideas what be causing this?

s.surf.idw seems to take about 1 CPU hour per site on my 400MHz PII (I
killed it after running all weekend).

s.surf.rst also seems to be taken an enormous amount of time.

My data is basic lat,lon,depth data.

I suspect I'm trying to use the wrong tools for the job, can anyone
suggest what I should be using?

Regards
Gordon

3

On Wed, Dec 20, 2000 at 11:30:19AM +1100, Gordon Keith wrote:

I'm trying to use grass5b10 with bathymetry data and am having some
trouble getting started.

I can import my data into a sites file (approx 150,000 sites) but I
haven't been able to convert it to a raster.

s.to.rast gives me:

     ERROR: error while writing to temp file

Has anyone got any ideas what be causing this?

Yes, it's either a bad region setting, or the temp file creation/open
failed, but it's not caught until later. I just updated the CVS copy to
hopefully be a little better about this (and at least report the problem
where it occurred). I can't be sure why program was unable to create a
temp file (it would be located as ...
<location>/<mapset>/.tmp/$HOST/$PID.nn ).

That's a big chunk of data to process at one time.

s.surf.idw seems to take about 1 CPU hour per site on my 400MHz PII (I
killed it after running all weekend).

s.surf.rst also seems to be taken an enormous amount of time.

My data is basic lat,lon,depth data.

I suspect I'm trying to use the wrong tools for the job, can anyone
suggest what I should be using?

I've had good luck using s.to.rast, then r.surf.idw. For something big,
it should work better than the surface generation from sites (as the
textual nature of sites makes reading/parsing slower and requires a fair
amount of memory). It will take a while to generate a surface no matter
what with that many locations.

--
Eric G. Miller <egm2@jps.net>