Hello all~
I'm working with SRTM data and trying to produce watershed maps in Honduras.
For most of the country, the threshold seems to work great at 100000, but
this is not the case on the shore - those watersheds are neglected likely
due to their smaller scale. Does anyone have experience merging multiple
scales? Does this make sense? I'd rather not reduce my threshold or the
whole country because those watersheds seem too small.
As an aside, what do people think about using SRTM for watershed mapping in
the first place - given it is a surface model and not a DEM.
--
View this message in context: http://osgeo-org.1560.x6.nabble.com/SRTM-r-watershed-at-multiple-scales-tp5072884.html
Sent from the Grass - Users mailing list archive at Nabble.com.
On Fri, Aug 16, 2013 at 6:51 PM, jencor808 <jencor808@gmail.com> wrote:
Hello all~
I'm working with SRTM data and trying to produce watershed maps in Honduras.
For most of the country, the threshold seems to work great at 100000, but
this is not the case on the shore - those watersheds are neglected likely
due to their smaller scale. Does anyone have experience merging multiple
scales? Does this make sense? I'd rather not reduce my threshold or the
whole country because those watersheds seem too small.
As an aside, what do people think about using SRTM for watershed mapping in
the first place - given it is a surface model and not a DEM.
While I cannot help you too much with your first question, you may
take a look at
Metz, M., Mitasova, H., and Harmon, R. S., 2011: Efficient extraction
of drainage networks from massive, radar-based elevation models with
least cost path search, Hydrol. Earth Syst. Sci., 15, 667-678,
doi:10.5194/hess-15-667-2011
http://www.hydrol-earth-syst-sci.net/15/667/2011/hess-15-667-2011.html
Markus
Hi jencor808,
I can only recommend using SRTM for subbasin delineations. We are using it on a regular basis in our hydrological modelling working group. You'll have to play around with thresholds a little. I use different thresholds for different areas of my catchment (eg. in mountainous regions I'd want smaller subbasins). By masking those areas and run r.watershed several times, you can vary your thresholds and then patch the different subbasin rasters back together. As the r.watershed algorithm often leaves small areas at the edges uncovered, you'll have to fill those before or after patching, otherwise you'll have holes in your subbasin map. You can also get rid of small subbasins by converting your subbasin raster into a vector map and then run
v.clean remove small areas (check manual)
Hope it helps. I have got a python/GRASS script that is station-based that does the varying thresholds for you and all the cleaning etc. Let me know if you'd want to give that a try.
Regards,
Michel
On 08/16/2013 06:51 PM, jencor808 wrote:
Hello all~
I'm working with SRTM data and trying to produce watershed maps in Honduras.
For most of the country, the threshold seems to work great at 100000, but
this is not the case on the shore - those watersheds are neglected likely
due to their smaller scale. Does anyone have experience merging multiple
scales? Does this make sense? I'd rather not reduce my threshold or the
whole country because those watersheds seem too small.
As an aside, what do people think about using SRTM for watershed mapping in
the first place - given it is a surface model and not a DEM.
--
View this message in context: http://osgeo-org.1560.x6.nabble.com/SRTM-r-watershed-at-multiple-scales-tp5072884.html
Sent from the Grass - Users mailing list archive at Nabble.com.
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
Michel~
Thank you for your response, though I have taken a long time to say so. I am coming back to this now, and would like to take you up on your offer to look at your script. I am familiar with python, but I am inexperienced with implementing it within GRASS, so that will be a good challenge.
Please let me know what you think about it,
~Jennifer
···
On Wed, Aug 21, 2013 at 4:18 AM, Michel Wortmann <wortmann@pik-potsdam.de> wrote:
Hi jencor808,
I can only recommend using SRTM for subbasin delineations. We are using it on a regular basis in our hydrological modelling working group. You’ll have to play around with thresholds a little. I use different thresholds for different areas of my catchment (eg. in mountainous regions I’d want smaller subbasins). By masking those areas and run r.watershed several times, you can vary your thresholds and then patch the different subbasin rasters back together. As the r.watershed algorithm often leaves small areas at the edges uncovered, you’ll have to fill those before or after patching, otherwise you’ll have holes in your subbasin map. You can also get rid of small subbasins by converting your subbasin raster into a vector map and then run
v.clean remove small areas (check manual)
Hope it helps. I have got a python/GRASS script that is station-based that does the varying thresholds for you and all the cleaning etc. Let me know if you’d want to give that a try.
Regards,
Michel
On 08/16/2013 06:51 PM, jencor808 wrote:
Hello all~
I’m working with SRTM data and trying to produce watershed maps in Honduras.
For most of the country, the threshold seems to work great at 100000, but
this is not the case on the shore - those watersheds are neglected likely
due to their smaller scale. Does anyone have experience merging multiple
scales? Does this make sense? I’d rather not reduce my threshold or the
whole country because those watersheds seem too small.
As an aside, what do people think about using SRTM for watershed mapping in
the first place - given it is a surface model and not a DEM.
–
View this message in context: http://osgeo-org.1560.x6.nabble.com/SRTM-r-watershed-at-multiple-scales-tp5072884.html
Sent from the Grass - Users mailing list archive at Nabble.com.
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
Markus,
last year you asked me whether the r.watershed output is not aligned to the r.water.outlet output in GRASS7 (s.b.). Ie. it leaves small areas out. Back then, I wasnt using GRASS7, but I can confirm now that it still does it and it is still rather annoying. Have you ever heard of the reasons or attempted to fix this?
Thanks,
Michel
On 08/26/2013 03:25 PM, Markus Neteler wrote:
On 08/22/2013 08:52 AM, Markus Neteler wrote:
>>
>>Hi Michel,
>>
>>On Wed, Aug 21, 2013 at 1:18 PM, Michel Wortmann
>><wortmann@pik-potsdam.de> wrote:
>>...
>>>
>>>As the r.watershed algorithm often leaves small areas at the edges
>>>uncovered, you'll have to fill those before or after patching, otherwise
>>>you'll have holes in your subbasin map.
>>
>>... just curious: does this happen still in the latest GRASS 7 version?
>>
>>Markus
I’ve been using GRASS 7 for this and have not seen a problem – I hope I’m not overlooking the issue. Michel, can you provide an image that shows this?
Thanks,
Tom
On Wednesday, February 19, 2014, Michel Wortmann <wortmann@pik-potsdam.de> wrote:
Markus,
last year you asked me whether the r.watershed output is not aligned to the r.water.outlet output in GRASS7 (s.b.). Ie. it leaves small areas out. Back then, I wasnt using GRASS7, but I can confirm now that it still does it and it is still rather annoying. Have you ever heard of the reasons or attempted to fix this?
Thanks,
Michel
On 08/26/2013 03:25 PM, Markus Neteler wrote:
On 08/22/2013 08:52 AM, Markus Neteler wrote:
Hi Michel,
On Wed, Aug 21, 2013 at 1:18 PM, Michel Wortmann
wortmann@pik-potsdam.de wrote:
…
As the r.watershed algorithm often leaves small areas at the edges
uncovered, you’ll have to fill those before or after patching, otherwise
you’ll have holes in your subbasin map.
… just curious: does this happen still in the latest GRASS 7 version?
Markus
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
–
Thomas E Adams, III
718 McBurney Drive
Lebanon, OH 45036
1 (513) 739-9512 (cell)
I’ve been having the same problem. I have seen figures illustrating the issue in riparian areas, where the watersheds ‘end’, leaving holes where perhaps very small watersheds should be delineated along rivers/streams instead. Personally, I am having this issue along coastlines, but its the same as in the riparian case. I can’t even fill the holes with a tiny threshold and I’m hesitant to make up my own watersheds at the coast.
Is there a way to provide a bounding line (ie. coastline) and wall-to-wall watersheds at a given scale are delineated within that area?
On Wednesday, February 19, 2014, Thomas Adams-2 [via OSGeo.org] <[hidden email]> wrote:
I’ve been using GRASS 7 for this and have not seen a problem – I hope I’m not overlooking the issue. Michel, can you provide an image that shows this?
Thanks,
Tom
On Wednesday, February 19, 2014, Michel Wortmann <[hidden email]> wrote:
Markus,
last year you asked me whether the r.watershed output is not aligned to the r.water.outlet output in GRASS7 (s.b.). Ie. it leaves small areas out. Back then, I wasnt using GRASS7, but I can confirm now that it still does it and it is still rather annoying. Have you ever heard of the reasons or attempted to fix this?
Thanks,
Michel
On 08/26/2013 03:25 PM, Markus Neteler wrote:
On 08/22/2013 08:52 AM, Markus Neteler wrote:
Hi Michel,
On Wed, Aug 21, 2013 at 1:18 PM, Michel Wortmann
<[hidden email]> wrote:
…
As the r.watershed algorithm often leaves small areas at the edges
uncovered, you’ll have to fill those before or after patching, otherwise
you’ll have holes in your subbasin map.
… just curious: does this happen still in the latest GRASS 7 version?
Markus
grass-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-user
–
Thomas E Adams, III
718 McBurney Drive
Lebanon, OH 45036
1 (513) 739-9512 (cell)
grass-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-user
If you reply to this email, your message will be added to the discussion below:
http://osgeo-org.1560.x6.nabble.com/SRTM-r-watershed-at-multiple-scales-tp5072885p5104742.html
To unsubscribe from SRTM r.watershed at multiple scales, click here.
NAML
–
Jennifer M. Corcoran, PhD
NASA Postdoctoral Program Fellow
Jet Propulsion Laboratory
4800 Oak Grove Drive
La Cañada Flintridge, CA 91011
alt email: [hidden email]
website: www.jennifermcorcoran.com
View this message in context: Re: [GRASS-user] SRTM r.watershed at multiple scales
Sent from the Grass - Users mailing list archive at Nabble.com.
On Wed, Feb 19, 2014 at 2:56 PM, Michel Wortmann
<wortmann@pik-potsdam.de> wrote:
Markus,
last year you asked me whether the r.watershed output is not aligned to the
r.water.outlet output in GRASS7 (s.b.). Ie. it leaves small areas out. Back
then, I wasnt using GRASS7, but I can confirm now that it still does it and
it is still rather annoying. Have you ever heard of the reasons or attempted
to fix this?
The reason is that r.watershed determines basins from streams, the ID
of basins is set from the streams. If a basin on a shore does not have
a stream outlet with at least <threshold> surface flow accumulation,
it will not be delineated, and there will be no streams in that area.
With r.water.outlet, you define an outlet and the upstream catchment
area will be traced from that outlet.
Markus M
Thanks,
Michel
On 08/26/2013 03:25 PM, Markus Neteler wrote:
On 08/22/2013 08:52 AM, Markus Neteler wrote:
>>
>>Hi Michel,
>>
>>On Wed, Aug 21, 2013 at 1:18 PM, Michel Wortmann
>><wortmann@pik-potsdam.de> wrote:
>>...
>>>
>>>As the r.watershed algorithm often leaves small areas at the edges
>>>uncovered, you'll have to fill those before or after patching,
>>> otherwise
>>>you'll have holes in your subbasin map.
>>
>>... just curious: does this happen still in the latest GRASS 7 version?
>>
>>Markus
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user