Hi Martin, hi Stefan,
Just to give an update. I used the original geographic coordinate system. Values are in a similar range so that it seems to work with longitude/latitude and I avoid reprojecting twice. In some reference I found that values of 55 might still be okay and in order to avoid having so many 0.03 I set the threshold extremely high (10^6), so basically the opposite as in a watershed delineation where I would like to have a rather small threshold to get more subbasins.
Thanks again,
Laura
-----Original Message-----
From: Scherer Laura
Sent: 25 July 2014 12:05
To: 'Stefan Lüdtke'; Martin Zbinden
Cc: grass-user@lists.osgeo.org
Subject: RE: [GRASS-user] LS factor for USLE
Hi Martin, hi Stefan,
Thanks for your feedback. I already tried with projected coordinate systems and different thresholds before asking here, but always got results that seemed strange to me. I don't have any feeling for typical values, but the r.watershed documentation says it's usually below 1 and that's true for most of the cells, but some are as high as 55 (after division by 100). But these extremes are very much dependent on the threshold. And most values are 0.03 (at a small threshold all were 0.03). That's why I thought that maybe only values at outlets are reliable, but according to the link from Stefan it has nothing to do with outlets and I should get results for all cells (except for streams).
I just tested the approach with r.slope.aspect, r.flow and r.mapcalc as described in the GRASS GIS book, but the output is even more strange. Everything is zero up to the 75% quantile and afterwards much higher (up to 4000).
Maybe interesting to know is that I want to do it for the world and as at 1 km the processing never seemed to end, I lowered the resolution to 10 km.
Best regards,
Laura
-----Original Message-----
From: Stefan Lüdtke [mailto:sluedtke@gfz-potsdam.de]
Sent: 25 July 2014 11:12
To: Martin Zbinden; Scherer Laura
Cc: grass-user@lists.osgeo.org
Subject: Re: [GRASS-user] LS factor for USLE
Hi Laura,
for question 2)
I had the same question some time ago, see here for an answer why a threshold is required.
http://lists.osgeo.org/pipermail/grass-user/2012-December/066491.html
HTH,
Stefan
On 07/25/2014 10:43 AM, Martin Zbinden wrote:
Hi Laura,
I've done some RUSLE3d calculations before, but always based on
projected coordinate systems.
1) With r.watershed you seem to be on the right path, as trying to use
r.flow on a geographic coordinate system gives me this error:
"lat/long projection not supported by r.flow. Please use 'r.watershed'
for calculating flow accumulation."
However, I just tried by reprojecting elevation from north carolina
sampledata (nc_basic_spm_grass7.zip) to geographic and the results are
not the same before and after reprojection. I followed instructions in
[1], page 46f. I don't know anything about reprojections so the error
might lay there. But if your results are also very implausible, then
maybe better try again with projected coordinate system.
2) I never tried slope length based USLE calculations and cannot
answer this question. Page 158 ff in [2] explains many different
approaches to RUSLE in GRASS GIS.
Best regards,
Martin
sources:
[1] http://grass.osgeo.org/sampledata/north_carolina/
[2] Markus Neteler and Helena Mitasova, 2008, Open Source GIS: A GRASS
GIS Approach. Third Edition.
--
Martin Zbinden
Riedacker 523
3154 Rüschegg Heubach
+41 78 628 28 82
--
Martin Zbinden
Riedacker 523
3154 Rüschegg Heubach
+41 78 628 28 82
2014-07-24 9:56 GMT+02:00 Scherer Laura <scherer@ifu.baug.ethz.ch>:
Dear all,
I would like to calculate the slope length and steepness (LS) factor
for USLE using r.watershed with the parameter length.slope. I have
two
questions:
1) Is it necessary to reproject my raster into a projected coordinate
system or can I leave it in latitude-longitude?
2) It is required to set a value for the threshold. Is it not
possible to get a value for each cell? If not, do I always have to
delineate the watersheds as well and the length.slope returned is the value at the outlet?
Thanks in advance for any help.
Best regards,
Laura
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
--
Stefan Lüdtke
Section 5.4- Hydrology
Tel.: +49 331 288 2821
Fax: +49 331 288 1570
Email: sluedtke@gfz-potsdam.de
Helmholtz-Zentrum Potsdam
Deutsches GeoForschungsZentrum GFZ
(GFZ German Research Centre for Geoscience) Stiftung des öff. Rechts Land Brandenburg Telegrafenberg, 14473 Potsdam
-------------------
PGP Public Key: http://bit.ly/13d9Sca