Importing Elevation data from an ASCII format

GRASS GRASS user
1

Tom Nelson writes:>

with a sites file then do either s.surf.idw or s.surf.tps.
Both create a surface file but tps takes a lot longer and with possibly better results.
then r.contour which produces a vector contour map of interpolated contours from the
surface file.

------------------------------------------------

I will have to disagree with "do either s.surf.idw or s.surf.tps". These two
interpolation modules are NOT the same! If you are creating an elevation surface
then I believe s.surf.tps is the one to use, because all "sites" are equally
weighted, whereas ".idw" refers to inverse distance weighting. The ".idw" could
create unsatisfactory results. I suggest testing these two modules on a sample
dataset and comparing results.

A good reference is Helena Mitasova article in the Spring 1993 GRASSCLIPPINGS.

Hope this helps...

--
R. Brian Culpepper Research Assistant
---------------------------------------------------------------------
Center for Advanced Spatial Technologies (fax) 501 575-3846
Rm. 214 Ozark Hall (work) 501 575-6159
University of Arkansas brian@cast.uark.edu
Fayetteville, AR 72701

All opinions are my own and do not necessarily reflect those of the
Center for Advanced Spatial Technologies or the University of Arkansas.

2

Brian Culpepper (brian@cast.uark.edu) writes on 9 Dec 93:

Tom Nelson writes:>

with a sites file then do either s.surf.idw or s.surf.tps.
Both create a surface file but tps takes a lot longer and with possibly better results.

I will have to disagree with "do either s.surf.idw or s.surf.tps". These two
interpolation modules are NOT the same! If you are creating an elevation surface

The important thing here is the underlying model that the data should
follow, characteristics of the program (smoothing or exact), and the
quality of the output.

IDW will tend to give peaks at the point values, which may not be
desirable. Exact interpolators (e.g., s.bissf, an implmentation of
Akima's method, forthcoming) expects data with no noise
(bissf=Bivariate Interpolation and Smooth Surface Fitting). TPS can do
smoothing, which is good for most of the kind of data that I've seen
grass folks working with.

The original poster may also be looking for s.to.rast, a contrib
program, that does not do any interpolation but simply makes each
site into a cell (depending on your resolution, see 'g.region -p').

--Darrell