[GRASS-user] flow accumulation (GRASS and Esri ArcGIS)

GRASS GRASS user
1

Hi,

one of my colleague is dealing with flow accumulation calculation in
GRASS, and comparing results with Esri ArcGIS (builtin Flow
Accumulation [1]). Tool in ArcGIS to me knowledge uses D8, so we run
r.watershed with -s flag and also -a flag to avoid negative values.
Resultant values are very different.

* Esri ArcGIS (range: 0 - 20741)
* r.watershed -s (range: -34068 - 5697)
* r.watershed -sa (1 - 34068)

Do you have idea why results are so different? Thanks for any pointers
in advance! Martin

[1] http://desktop.arcgis.com/en/arcmap/10.5/tools/spatial-analyst-toolbox/flow-accumulation.htm

--
Martin Landa
http://geo.fsv.cvut.cz/gwiki/Landa
http://gismentors.cz/mentors/landa

2

On Wed, May 2, 2018 at 7:13 PM, Martin Landa <landa.martin@gmail.com> wrote:

Hi,

one of my colleague is dealing with flow accumulation calculation in
GRASS, and comparing results with Esri ArcGIS (builtin Flow
Accumulation [1]). Tool in ArcGIS to me knowledge uses D8, so we run
r.watershed with -s flag and also -a flag to avoid negative values.
Resultant values are very different.

* Esri ArcGIS (range: 0 - 20741)
* r.watershed -s (range: -34068 - 5697)
* r.watershed -sa (1 - 34068)

I'd suggest to look at the univariate statistics first (r.univar -e)
and compare that instead of min/max only.

markusN

Do you have idea why results are so different? Thanks for any pointers
in advance! Martin

[1] http://desktop.arcgis.com/en/arcmap/10.5/tools/spatial-analyst-toolbox/flow-accumulation.htm

--
Martin Landa
http://geo.fsv.cvut.cz/gwiki/Landa
http://gismentors.cz/mentors/landa
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

--
Markus Neteler, PhD
http://www.mundialis.de - free data with free software
http://grass.osgeo.org
http://courses.neteler.org/blog

3

On Wed, May 2, 2018 at 2:21 PM, Markus Neteler <neteler@osgeo.org> wrote:

On Wed, May 2, 2018 at 7:13 PM, Martin Landa <landa.martin@gmail.com> wrote:

Hi,

one of my colleague is dealing with flow accumulation calculation in
GRASS, and comparing results with Esri ArcGIS (builtin Flow
Accumulation [1]). Tool in ArcGIS to me knowledge uses D8, so we run
r.watershed with -s flag and also -a flag to avoid negative values.
Resultant values are very different.

* Esri ArcGIS (range: 0 - 20741)
* r.watershed -s (range: -34068 - 5697)
* r.watershed -sa (1 - 34068)

I'd suggest to look at the univariate statistics first (r.univar -e)
and compare that instead of min/max only.

Also, r.watershed uses least-cost path, unlike the arcmap tool. Did
you fill the depressions before running the arcmap tool?

Anna

markusN

Do you have idea why results are so different? Thanks for any pointers
in advance! Martin

[1] http://desktop.arcgis.com/en/arcmap/10.5/tools/spatial-analyst-toolbox/flow-accumulation.htm

--
Martin Landa
http://geo.fsv.cvut.cz/gwiki/Landa
http://gismentors.cz/mentors/landa
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

--
Markus Neteler, PhD
http://www.mundialis.de - free data with free software
http://grass.osgeo.org
http://courses.neteler.org/blog
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

4

On Wed, May 2, 2018 at 8:29 PM, Anna Petrášová <kratochanna@gmail.com> wrote:

On Wed, May 2, 2018 at 2:21 PM, Markus Neteler <neteler@osgeo.org> wrote:

On Wed, May 2, 2018 at 7:13 PM, Martin Landa <landa.martin@gmail.com> wrote:

Hi,

one of my colleague is dealing with flow accumulation calculation in
GRASS, and comparing results with Esri ArcGIS (builtin Flow
Accumulation [1]). Tool in ArcGIS to me knowledge uses D8, so we run
r.watershed with -s flag and also -a flag to avoid negative values.
Resultant values are very different.

  • Esri ArcGIS (range: 0 - 20741)
  • r.watershed -s (range: -34068 - 5697)
  • r.watershed -sa (1 - 34068)

I’d suggest to look at the univariate statistics first (r.univar -e)
and compare that instead of min/max only.

Also, r.watershed uses least-cost path, unlike the arcmap tool. Did
you fill the depressions before running the arcmap tool?

If you want to get nearly identical results, use r.hydrodem to remove sinks from the DEM, then use this conditioned DEM as input for both ArcGIS and r.watershed. Note that with D8 there is some randomness about where surface flow will go to in case of several neighbors with the same elevation. In any case, if surface accumulation of ArcGIS and r.watershed differ, literature cited in the corresponding manuals should allow to decide which method is more appropriate for the current task.

Markus M

Anna

markusN

Do you have idea why results are so different? Thanks for any pointers
in advance! Martin

[1] http://desktop.arcgis.com/en/arcmap/10.5/tools/spatial-analyst-toolbox/flow-accumulation.htm

–
Martin Landa
http://geo.fsv.cvut.cz/gwiki/Landa
http://gismentors.cz/mentors/landa

grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

–
Markus Neteler, PhD
http://www.mundialis.de - free data with free software
http://grass.osgeo.org
http://courses.neteler.org/blog

grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

5

Thanks all for your suggestions!

I created a comparison of methods [1] which were mentioned. But unfortunately, the results are still very different (at least from my point of view). What do you think about it? What approach should be chosen?

[1] https://docs.google.com/document/d/15MyV-l2BH5agjN1T3LtQBotH7RJsIL_5K7exSZypTnQ/edit?usp=sharing

···

2018-05-02 22:17 GMT+02:00 Markus Metz <markus.metz.giswork@gmail.com>:

On Wed, May 2, 2018 at 8:29 PM, Anna Petrášová <kratochanna@gmail.com> wrote:

On Wed, May 2, 2018 at 2:21 PM, Markus Neteler <neteler@osgeo.org> wrote:

On Wed, May 2, 2018 at 7:13 PM, Martin Landa <landa.martin@gmail.com> wrote:

Hi,

one of my colleague is dealing with flow accumulation calculation in
GRASS, and comparing results with Esri ArcGIS (builtin Flow
Accumulation [1]). Tool in ArcGIS to me knowledge uses D8, so we run
r.watershed with -s flag and also -a flag to avoid negative values.
Resultant values are very different.

  • Esri ArcGIS (range: 0 - 20741)
  • r.watershed -s (range: -34068 - 5697)
  • r.watershed -sa (1 - 34068)

I’d suggest to look at the univariate statistics first (r.univar -e)
and compare that instead of min/max only.

Also, r.watershed uses least-cost path, unlike the arcmap tool. Did
you fill the depressions before running the arcmap tool?

If you want to get nearly identical results, use r.hydrodem to remove sinks from the DEM, then use this conditioned DEM as input for both ArcGIS and r.watershed. Note that with D8 there is some randomness about where surface flow will go to in case of several neighbors with the same elevation. In any case, if surface accumulation of ArcGIS and r.watershed differ, literature cited in the corresponding manuals should allow to decide which method is more appropriate for the current task.

Markus M

Anna

markusN

Do you have idea why results are so different? Thanks for any pointers
in advance! Martin

[1] http://desktop.arcgis.com/en/arcmap/10.5/tools/spatial-analyst-toolbox/flow-accumulation.htm

–
Martin Landa
http://geo.fsv.cvut.cz/gwiki/Landa
http://gismentors.cz/mentors/landa

grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

–
Markus Neteler, PhD
http://www.mundialis.de - free data with free software
http://grass.osgeo.org
http://courses.neteler.org/blog

grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

grass-user mailing list
grass-user@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/grass-user

6

Hi Radek,

út 12. 6. 2018 v 11:06 odesílatel Radek Novotný
<radeknovotny94@gmail.com> napsal:

I created a comparison of methods [1] which were mentioned. But unfortunately, the results are still very different (at least from my point of view). What do you think about it? What approach should be chosen?

look at *mean* values which are very close (except of r.hydrodem +
flow_accumulation). Maximum values are probably just few pixels
located at region edges. Ma

--
Martin Landa
http://geo.fsv.cvut.cz/gwiki/Landa
http://gismentors.cz/mentors/landa