[GRASS-user] r.stream.extract Montgomery

[Jarek Jasiewicz]

stephen sefick pisze:
> What is good guidance for where to set the threshold?  I am in the
> coastal plain, but In the very upper part ~10 miles from the piedmont,
> so it is not terribly low gradient (I am going to quantify this "not
> terribly low gradient" soon).  I don't see the -b flag for r.watershed
> 6.4svn checkout (probably a week ago) in the man pages.  I said
> realistic because you can set the threshold to 1 for a 1m res dem.
> There are lines that don't even look like streams all over the place.
> Is it possible to extract streams with the landscape (Mongomery) as a
> guide for the threshold? 
the rule for Montgomery is described in literature cited in 
r.stream.extract help. Look into this.
In general, Montomery's assumptions take into account annual 
precipitation versus slope gradient * SCA (specific catchment area). It 
is not a problem if you have 1m dem, because in that one specific 
situation SCA = accumulation, so you can use a threshold = 200, as 
suggested by Montogmery and Dietrich for Appalachian. But it is not 
possible to say which value is best.
>  I want to be able to do this so all of the
> stream ordering is done on a consistent stream network among
> watersheds.  If this is not possible then I will use a threshold and
> use this for all stream network extraction on all of the other dems
> that I need to process so they are all comparable.  I appologize if
> this doesn't make sense, but I will explain in greater detail if I
> need to.
> kindest regards,
>
> Stephen Sefick
>
> On Thu, Aug 5, 2010 at 9:35 AM, Jarek Jasiewicz <jarekj at amu.edu.pl> wrote:
>   
>> stephen sefick pisze:
>>     
>>> I would like to use the Montgomery method to extract streams (exp~2
>>> from the paper).  What should the threshold value be?  And in general
>>> what should the threshold value be?  I am working in the southeastern
>>> coastal plain (USA), which is characterized by low gradient (if this
>>> helps).
>>>       
>> for flat areas Montgomery's method is not a good option. Generally that
>> method was created and tested on areas with gradient > 5%.
>>     
>>>  I am using 1m resolution LIDAR data.  I would like to extract
>>> the most realistic map of streams that I can.
>>>       
>> Chmmm... what you mean realistic? Maybe use existing stream network will be
>> the best solution?
>> For coastal plains where is no real vallyes the r.watershed's treeshold with
>> -b option seems to be best option
>>     
>>>  I am also trying to
>>> track down the inttermitance perminance threshold.  Any guidance would
>>> be greatly appreciated.
>>>       
SWAT (soil water assessment tool) but not for GRASSS.

>>> kindest regards,
>>>
>>>
>>>       
>>     
>
>
>
>