[GRASS-user] r.watershed negative values in flow accumulation

[Jarek Jasiewicz]

Margherita Di Leo pisze:
> Hi all,
>
> I found a strange result runnig r.watershed . There are _many_ 
> negative values in flow accumulation obtained with Multiple Flow 
> Direction algorithm (MFD, f flag). I compared with the accumulation 
> map obtained with r.terraflow. It is surprising to see that with 
> r.terraflow there are very few negative values. How is it possible? Is 
> the algorithm (MFD) the same? How is the true result?
>
> I'm runnig Grass 6.5 svn on Ubuntu 8.04 - 64 bit.
negative values are ascribed to that cells where real cumulated flow 
cannot be calculated. In most cases it is for streams which drain from 
outside area. See help for r.watershed, it is well described there.
If you want all values positve run: echo 'new_accum=abs(accum)'|r.mapcalc
BTW it is very good idea to use negative values for that cells, As far 
as I know only grass had that feature
Jarek
>
> Command lines:
>
> r.watershed -f elevation=elev1 at ex_20 accumulation=accumulation_MFD 
> convergence=5 memory=300
>
> r.terraflow elevation=elev1 at ex_20 filled=filled_terraflow 
> direction=dir_terraflow swatershed=swa_terraflow 
> accumulation=acc_terraflow tci=tci d8cut=infinity memory=300 
> STREAM_DIR=/var/tmp stats=/home/margherita/terraflow_stats.out
>
> Thank you in advance.
>
> margherita
>