r.stream.distance elevation above streams produces unexpected null values

[madi]

As written in the summary. The commands i used:

r.watershed -a -b elevation=dem threshold=1500 accumulation=accum drainage=drainage stream=stream1500

r.stream.distance stream_rast=stream1500 direction=drainage elevation=dem method=downstream difference=elev2stream

elev2stream map has several null values. I tried with 2 different dems, srtm and aster2, on senegal area.

Linux Fedora 20, GRASS 7.1.svn (r63559)

[hellik]

Replying to madi:

elev2stream map has several null values.

where are the null values?

I know only the null values at the border of the DEM (red in the screenshot) where no stream information is available.

[madi]

Replying to hellik:

Replying to madi:

elev2stream map has several null values.

where are the null values?

I know only the null values at the border of the DEM (red in the screenshot) where no stream information is available.

Hi, in my case null values are scattered all over the map, see attached screenshot (the large spots however are water). I realised that the command line i actually gave is slightly different from the one i posted earlier (if matters) and is:

r.stream.distance --verbose stream_rast=stream1500 direction=drainage elevation=dem method=downstream distance=dist2stream_2 difference=elev2stream_2 memory=2000

Thanks madi

[madi]

Nulls

[hellik]

Replying to madi:

Replying to hellik:

Replying to madi:

elev2stream map has several null values.

where are the null values?

I know only the null values at the border of the DEM (red in the screenshot) where no stream information is available.

Hi, in my case null values are scattered all over the map, see attached screenshot (the large spots however are water). I realised that the command line i actually gave is slightly different from the one i posted earlier (if matters) and is:

r.stream.distance --verbose stream_rast=stream1500 direction=drainage elevation=dem method=downstream distance=dist2stream_2 difference=elev2stream_2 memory=2000

Thanks madi

some tests with a DEM subset:

- maybe differences in the r.stream.*-modules between all-in-memory and swapped memory mode; in my tests I can't see any difference

- I've done a r.geomorphon run on the DEM; it seems the DEM is very heterogenous. may be these null values are DEM artefacts?

- I've done a r.neighbor size=3 and size=5 on the DEM; there seems to be a less number of null values. maybe this indicates DEM artefacts?

- I've done a r.stream.distance run with r.stream.extract-outputs (streams and flowdir) instead; there are some differences regarding null values between using r.watershed or r.stream.extract-output as r.stream.distance input

[neteler]

Milestone renamed

[neteler]

Ticket retargeted after milestone closed

[neteler]

Ticket retargeted after milestone closed

[martinl]

Ticket retargeted after milestone closed