[GRASS-user] about r.threhold

[Margherita Di Leo]

Dear Rengifo Ortega,

On Thu, Jun 11, 2020 at 1:08 PM Rengifo Ortega <rengifoo at yahoo.de> wrote:

> Dear Grass community
>
> First of all thanks for this great piece of  software such  as GRASS GIS!
> Recently I have been using  the  r.threshold module to extract a  river
> networks at different  DTM resolutions and different areal extent. I
> noticed that   r.threshold values varies widely as a function DEM
> resolution and extent of watershed (areal extent). So , I have some
> questions, and hope someone in the community can shed some light on them.
>
> Looking at the description of the module  it says : "This approach
> provides a *best guess *about what makes sense when looking only at the
> DEM"
>
>  Question 1: means   that the results of the r.threshold module will
> depend on the resolution and  the extent of the watershed?
>
> Looking  at the source code, I  realised that r.threshold used r.stats to
> generate a flowacc  text.file ordered in ascendent order and put into a
> matrix called  mappatella, with 3 columns defined as:
>
> *mappatella is a matrix, in the first column the value of upslope area is
> stored, *
>
>
> *in the second the number of cells, in the third the distance from origin
> is calculated*
>
> further in the script the distance is defined as :
>
>
> *calculating distance from origin of each point; origin of the plot is in
> low left point *
>
>
> Question 2 : which is the origin?  is the center  of  a cell  of the
> corner of a cell?
> Question 3: which plot? and why  the low left point of the plot?
>
>  Although I understand the general aspect  of the module I still struggle
> to understand its details. I would appreciate some explanation to crarify
> it further, since I  am using it as part of workflow within  GRASS GIS to
> produce inputs to a Distance Distribution Dynamics( DDD) model for urban
> enviroments.
> https://onlinelibrary.wiley.com/doi/full/10.1002/hyp.10315?casa_token=ETYV_ojgMb4AAAAA%3AMWgGgX_jqbylpF6pHLoafciVZrlmCBiLCmxb7gRpsCV1F3SY7QQmOkBOIrhLzhj5BFCDSLobabm7Xw
>
>
Thank you for your interest in r.threshold. It's been written some time ago
but the principle is the following: in order to find the point "where the
stream begins" you can look at where you have an abrupt change in the slope
and flow accumulation increases. In order to do that, you can plot slope
and flow accumulation in a graph and pick the point that is the closest to
the origin. This is a very naive and preliminary guess but somehow allows
to start from a plausible value in order to find a better threshold.

Hope this helps

Kind regards,




Any help would be appreciated.
>  Thanks in advance !
>
> Best regards
> Rengifo Ortega
>
>
>
>
>
>
>
>

-- 
Margherita Di Leo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20200611/5c7b538e/attachment-0001.html>