[GRASS-user] r.watershed: whats wrong?

[achim]

That helped!

Thank you very much.
I thought that in depression-input zero means depressions (like in
r.terraflow, which does not produce that good results).

Thanks,
achim

Btw: I suggest oceans as real sinks. Think thats ok... ;-)




the Problem was as you s

Markus Metz schrieb:
> 
> 
> achim wrote:
>> Hi all,
>>
>> I used r.watershed on a small test-region an all I produce
>> for accumulation and flow-direction was -1 on the whole area
>> except on the border-cells. Whats wrong?
>>
>> Curiously is that the first time I tried on my equal-area projection
>> everything seemed normal. Before on ll it was wrong and after on that
>> laea-projection it was like mentioned.
>>
>> I tried on grass63 and on grass64 (and different thresholds, new
>> output-names...). Both the same. (On opensuse 11.1 64)
>>   
> May be you need to run
> g.region rast=etopo1_fil
> first?
> 
> You can not use the same raster as elevation and as depression, this
> will give the you results you observed, nonsense results.
> The name etopo_fil suggests that you filled the elevation map first
> before using it as input for r.watershed. This can heavily distort flow
> accumulation and flow direction output, the input elevation map *must
> not* be filled for r.watershed. Input depression is only needed if there
> are real depressions in the area, depressions without outflow. The Dead
> Sea is such a real depression, ordinary lakes are not, they usually have
> an outflow.
> More details are in the help page:
> http://grass.osgeo.org/grass64/manuals/html64_user/r.watershed.html
> 
> Good luck,
> 
> Markus M
> 
>> r.watershed elevation=etopo1_fil at achim depression=etopo1_fil at achim
>> threshold=100000 accumulation=r_watershed_etopo1_accu
>> drainage=r_watershed_etopo1_dir basin=r_watershed_etopo1_basin
>> stream=r_watershed_etopo1_stream half.basin=r_watershed_etopo1_halfbasin
>> visual=r_watershed_etopo1_visual
>> length.slope=r_watershed_etopo1_lenghtslope
>> slope.steepness=r_watershed_etopo1_slopesteepness --overwrite
>>
>> Can anyone help?
>>
>> Best regards
>>
>> Achim
>> _______________________________________________
>> grass-user mailing list
>> grass-user at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/grass-user
>>
>>