There is a number of modules suitable for different applications:
If you are getting lots of straight lines the resolution of your DEM is too high
for D8 compared to the surface complexity - lower the resolution to get more realistic
stream network pattern, although some detail will be lost. or you can use r.flow
if there are no depressions in the DEM.
The version r.flowmd distributes water to 2 cells rather than a single one
Compare different algorithms for stream extraction from IFSARE
see
GRASS manual for detailed description of commands
r.watershed ch20ftdem accum=ch50f.accum basin=ch50f.basin threshold=50000 stream=ch50f.stream
The module uses SFD-D8 algorithm and finds the route through the pits using a shortest path
algorithm.
To really see the flowaccumulation map you need to change the color table, for example like this:
(note that the negative values are streams that may originate outside the given area)
r.colors ch50f.accum co=rules
>-500000 black
>-100 black
>-10 grey
> 0 white
> 10 yellow
> 100 cyan
> 1000 blue
>300000 0 0 30
To get a nice stream network for a selected flow accumulation threshold use map algebra
r.mapcalc acc100="if(abs(ch50f.accum)>100,1,null())"
r.watershed ch20ftdem accum=ch20f.accum basin=ch20f.basin threshold=5000 stream=ch20f.stream
r.flow elrst20ft.t50sm skip=4 flout=flowrst lgout=lgdrst20ft dsout=dsdrst20ft
r.flow -u elrst20ft.t50sm skip=4 flout=flowrst.u lgout=lgdrst.u dsout=usdrst
/usr4/grass5cvs/grass5sourcecvs/grass50_exp_2003_02_22/src.contrib/DUKE/r.terraflow
r.terraflow -s elevin=elrst20ft.t50 filled=elrst20ft.t50.fill direction=elrst20ft.t50dir swatershed=terrawshd accumulation=terra.accum tci=terra.tci
MFD flow accumulation for Chagres, Panama
MFD floods the depressions
MFD floods the depressions - closer look, note the noise on IFSARE DEMj
Close look at the comparison between SFD terraflow result and r.watershed
/usr4/grass5cvs/grass5sourcecvs/grass50_exp_2003_02_22/src.contrib/GMSL/r.simwe
r.sim.water
r.sim.sediment