[GRASSLIST:4418] Re: watershed delineation and stream map

[Jaro Hofierka]

Thomas Dewez wrote:

> Dear all,
> 
> I looked at the different functions concerning the hydrological analyses
> and I am baffled as to which is best suited for my purpose. I'd appreciate
> some expert insight on the question.
> 
> My aim is to extract several morphotectonic parameters from a 10x10m DEM.
> One of the features I am after is the presence of coastal terraces
> preserved on valley ridges. To get them out of the DEM I was thinking of
> extracting the ridges location (with perhaps a r.watershed command?) then
> process each individual ridge profile looking for flat surfaces (perhaps
> with a histogram of elevation of the ridge?). That's for the principle but
> I am not sure how to implement this efficiently.
> 
> The r.watershed interactive module requires the manual input of the pour
> point coordinates before it can run (is this correct?). Well, I'd rather
> not have to input those manually and have an algorithm do it for me.
> That's perhaps a little lazy but surely it exists. Does anybody have a
> suggestion about this question?
> 


Thomas,

r.watershed can be run non-interactively as well. Streams can be 
extracted from DEM using stream and threshold options. Using an 
"inversed" DEM you can get the ridges you are after.

Jaro