How to develop a dem map

[Hal Mueller]

At 5:50 PM +0200 9/17/99, Antonio Rodríguez Verdugo wrote:
>With s.surf.tps I only get a plain raster file with no appreciable data. I
>tried with s.surf.idw, and it seems to fit my needs. But when I run the
>d.3d module I get some "holes" or depression as if I put negative values,
>but it isn't the case. So I think I must tune up something.

You're simply running into the limitations of the interpolation 
schemes.  Whenever you try to fill in holes in some data sets, you 
need to understand the algorithm being used and judge its 
appropriateness to your purpose.  There are many many schemes for 
building surfaces from contours or points; none are correct for every 
problem.

It sounds to me like you're choosing among GRASS programs because 
they're readily available.  What you should be doing instead is 
choosing a particular mathematical algorithm because it's the right 
one for your problem.  Eventually you're going to need to write your 
own replacement for these GRASS modules, implementing the scheme 
you've chosen based on literature reviews or your own mathematical 
formulation.

I suggest you take a look at Noel Cressie's book "Statistics for 
Spatial Data", or perhaps spend some time hanging out on the Geostats 
mailing list (send subscribe/unsubscribe requests to 
ai-geostats-request at gis.psu.edu).

----
Hal Mueller
Seattle, Washington    hal at seanet.com