Jarek:
> > Interested question, but I think the easiest answer is carving or
> > stream burining (r.stream.carve). Next reproduce r.watershed or
> > r.stream.extract you shall receive (almost) ideally confidential
> > network with supplied
Achim Kisseler:
> Stream burning: Thats what I thought about.
Nikos:
Thanks guys, I'll be testing this quickly (or just pass the information to
the friend (uses Arc-something-9.3).
Steps tested (more or less successfully):
# comment: contours were digitised from topomaps with a scale of 1:50000
# scanned at 300dpi
# estimated cell size = 4.23 m
# same scale applies to digitised streams
# set region and rasterise
g.region vect=contours_20m res=4.23m -pa
v.to.rast in=contours_20m out=contours_4.23m
# comment: this is harder than I thought. Quick checking reveals
disconnected lines
# set resolution to 1m and (re-)rasterize
g.region res=1 -pa
v.to.rast in=contours_20m out=contours_1m
# checking... looks ok.
# I think this particular step is a "crux" for a successful processing
# any ideas to avoid disconnected lines at a resolution >4.23m?
# some special downscaling method? Neatest neighbor seems to be a bad choice
here.
# change resolution as the "1m" will be running for days?
g.region vect=contours_20m res=50m -pa
# contours to elevation
r.surf.contour in=contours_1m out=elevation_50m
# carving...
r.carve rast=elevation_50m vect=streams out=elevation_50m_carved
# extracting streams and direction map (not sure about the threshold)
r.stream.extract elevation=elevation_50m_carved stream_rast=streams_extracted
direction=stream_directions_extracted threshold=100
# strahler-ordering streams
r.stream.order stream=streams_extrected dir=stream_directions_extracted
strahler=strahler
# resulted streams look a bit squared due to the 50m res I guess. "Strahler"
looks correct to me (afaict from my studies 6-7 years ago).
Nikos
