[GRASS-user] r.watershed identify inland watershed

[Micha Silver]

On 4/2/21 5:37 PM, ming han wrote:
> Maybe I am the only one who has this demand. Following is just a 
> recommendation to GRASS r.watershed function.
> Maybe it is worth having an option to avoid r.watershed overcome 
> depressions.
> The reasons are 1) there are many hydrologically pre condition DEM 
> data available globally, such as:HydroSHEDS, MERIT
>                             2) the depression in these DEM are real 
> depressions, overcome these depressions will make the entire drainage 
> system


Regarding Hydrosheds, the documentation[1] in section 3.4 explains how 
they overcame the problem of sinks. They performed a regular "fill 
sinks" operation on areas that were SRTM artifacts. True natural 
depressions were identified manually, then another manual procedure of 
carving rivers was done to force flow thru these depressions and produce 
hydrologically correct streams and basins. So pre-conditioning to 
overcome depressions is not a magic bullet...


In my opinion, the best results are obtained when true depressions 
(pits, salt playas or karst regions) are identified, and set to NULL in 
the elevation raster. That will allow r.watershed to stop routing at 
those locations, and produce correct stream and basin layers.


[1]https://hydrosheds.org/images/inpages/HydroSHEDS_TechDoc_v1_2.pdf


> incorrectly.
>
> I understand GRASS has other functions to solve this problem, but just 
> a user recommendation. I use GRASS a lot.
>
> Thanks
> Ming
>
> ming han <dustming at gmail.com <mailto:dustming at gmail.com>> 
> 于2021年3月30日周二 上午8:06写道：
>
>     Got it, thanks everyone~
>     Ming
>
>     Micha Silver <tsvibar at gmail.com <mailto:tsvibar at gmail.com>>
>     于2021年3月29日周一 下午2:40写道：
>
>         Hello:
>
>         You might try `r.param.scale`, or even better `r.geomorphons`
>         modules to
>         identify geomorphology features, then filter out all pixels
>         identified
>         as pits.
>
>
>         r.watershed is purposely designed to overcome depressions, and
>         find flow
>         routing thru these spots. So I don't think you can use that
>         module to
>         identify depressions.
>
>
>         On 3/27/21 8:49 PM, ming han wrote:
>         > Hi  Everyone
>         >
>         >      When I do watershed delineation using r.watershed for
>         great salt
>         > lake watershed. I found r.watershed always tried to assign
>         an outlet
>         > for a great salt lake, which does actually not exist because
>         it is an
>         > inland lake and the great salt lake has no watershed outlet
>         at all.
>         >
>         >       I noticed that there is a depression option. But is
>         there any
>         > way that  r.watershed can automatically identify depressions
>         while
>         > defining flow accumulation and stream network?
>         >
>         > Thanks
>         > Ming
>         >
>         > _______________________________________________
>         > grass-user mailing list
>         > grass-user at lists.osgeo.org <mailto:grass-user at lists.osgeo.org>
>         > https://lists.osgeo.org/mailman/listinfo/grass-user
>         <https://lists.osgeo.org/mailman/listinfo/grass-user>
>
>         -- 
>         Micha Silver
>         Ben Gurion Univ.
>         Sde Boker, Remote Sensing Lab
>         cell: +972-523-665918
>
-- 
Micha Silver
Ben Gurion Univ.
Sde Boker, Remote Sensing Lab
cell: +972-523-665918