[GRASSLIST:3802] Re: interpolate rainfall - r.runoff

[Markus Neteler]

On Thu, May 30, 2002 at 06:48:15PM -0600, Roger Miller wrote:
> On Thursday 30 May 2002 19:12, felipe wrote:
> 
> > Hello list,
> >
> > What is the best method to interpolate rainfall maps in GRASS and/or Linux?
> >
> > Something how kriging....
> 
> I have used gstat ( http://www.gstat.org ) to produce GRASS raster maps via 
> kriging.  Gstat offers a lot of capabilities, but the version I have is 
> fairly difficult to use and behaves badly with even moderate-sized datasets.  
> With luck that has been improved.

Another method is to use s.vol.rst if you interpolate in an area
with non-flat elevations (refs: [1], [2])

You need
 - 3D sites of climate stations (x,y,z,attribs)
 - 2D raster DEM

The interpolation is done with 3D RST (3D splines) as provided by 
's.vol.rst'.

's.vol.rst' expects you have defined 3D region (as every other 3D module).
Therefore you have to run 'g3.createwind' script before 's.vol.rst'.
It doesn't matter that you will produce just a 2D, intersection file.
's.vol.rst' asks for 3D region during the input (remember, you must have
3D sites file).
 
With 'g3.setregion' you can change your vertical resolution to 3000m
(i.e. define only one depth layer, s.vol.rst will be happy with that).
 
The command line may be something like this:
  s.vol.rst in=precip3d cellinp=dem500 cellout=precip.topo zmult=50
 
The resulting map 'precip.topo' is a 2D raster map containing the
precipitation. This precipitation was calculated internally with
3D Regularized Splines with Tension (RST). The 'cellinp' method
extracts the precipitation values from the precipitation volume at
the individual elevations as stored in the 'dem500' raster map.

[1] Hofierka J., J. Parajka, H.  Mitasova, and L. Mitas, 2002. Modeling
    impact of terrain on precipitation using 3-D spatial interpolation.
    Transactions in GIS, March 2002, vol. 6, no. 2, pp. 135-150(16)
[2] The upcoming GRASS book

> > I need to generate some rainfall maps to scs runoff calculation.
> >
> > Where can I get the r.runoff code?
> > What is G-language?
> 
> r.runoff was part of grass4.  Documentation for r.runoff is still included in 
> grass5, but I haven't been able to find the code.  You may be able to get it 
> from a grass4 distribution or from somewhere deep in the Grass CVS tree, but 
> I can't lead you to believe that it will work with Grass5.

Check also:
r.cn help
Description:
 Generates a SCS curve number map layer

> I have no idea what G-language is.
Where did you find this reference?
 
Markus Neteler