r.cost etc..

GRASS 4.0 grass at holstein.age.uiuc.edu
Wed Sep 28 14:31:47 EDT 1994


I have a multi-faceted question:

I am specifically trying to get legngth downslope
to a watershed outlet.

I ran my DEM through a program called r.fill.dir,
which told me that It could not fill all areas.
The result is that r.drain, of course, stops
at all depressions.

Next I made a cost map layer based on the DEM
and the watershed outlet as a starting point.

The result looked like a target...

Finally, I masked the watershed channels and copied
my dem to another file and used this layer as input
for r.cost.  The results, using this r.cost output for 
input for r.drain produces paths that do not match 
well with the topography map.  There seems to be a 
tendancy for the procedure to create straight
flow lines that are not necessarily normal to contours.

I am using a DEM generated by s.surf.tps in centimeters.
The watershed has approximately 30K acres and the relief
range is 30 meters.

1. does a script or program exist to display a flow line as
well as calculate it's length to the outlet that is more 
accurate? (based on the DEM)

2. how does r.cost do its calculations?  I am mystified as
to what the procedure is and the manual page does not cover 
it.  Does anyone have an address to the author or a source for
some documentation on the algorithm?
	mainly I want to know what is done with diagonals and
	the increasing area between the diagonals as a function of
	distance from starting point.  What does a "knights move" do?

Keith M. Mitchell
Dept. Of Ag Engineering
University Of Illinois.




More information about the grass-user mailing list