[GRASS-user] Re: understanding r.watershed

[Georg Kaspar]

On Thu, 29 Jan 2009 10:20:36 -0500, M S wrote:

> In short, r.watershed, without depression input,  will route water in
> and *up* and out of depressions in the terrain to illustrate the
> complete downward path.  This is why no DEM filling is necessary.  By
> entering in known (substantial and impactful) depressions, the water
> does not route out of it.

But in my case, providing known depressions leads to areas of null()s 
around those depressions (lakes by the way).