[Dhsvm-users] Re: stream network and DEM [Demon conversation]

[Keith Norman Adams]

Dear DHSVM Users,

Here are a few points that may be of interest:

1.
For the catchment I am modelling I have a stream network coverage from a 
the New Zealand 1 : 50,000 series topographic maps (in ArcInfo coverage 
form).
Before running 'createstreamnetwork.aml', I converted the coverage to a 
grid with each cell on the stream network having a value of 1.0
Then subtracted this grid from my DEM.
This helped to put the stream network where it should be.

I also drained a large (but shallow) depression on a low lying part of 
the catchment, by cutting a drain in the DEM a few metres deep (there is 
actually a similar drain in actual catchment).
This sped up the sink removal process in ArcInfo GRID.

I also fiddled the elevations of a few cells in some areas where 
'createstreamnetwork.aml' thought the stream should be flow backwards 
(i.e. upstream).
Then had to go back to almost the beginning of 'createstreamnetwork.aml'

That all took a few days.


2.
The GRID streamline function (as in 'createstreamnetwork.aml') reverses 
the direction of the last segment (outlet segment) of the stream network 
if a tributary joins the stream network at the outlet cell (the mouth cell).


3.
I find it is better to step through 'createstreamnetwork.aml' manually, 
and call each of the other AML programs manually as needed.


Regards,
Keith Adams.




Surfleet, Christopher wrote:

> One thing I must emphasize in my last message, is if you edit your 
> stream network, in my case shorten a few arcs.  You HAVE TOO run it back 
> through createstreamnetwork to create the updated files for DHSVM (when 
> you add your own stream network do not specify a source area tolerance).
> 
> ------------------------------------------------------------------------
> *From:* dhsvm-users-bounces at mailman1.u.washington.edu 
> [mailto:dhsvm-users-bounces at mailman1.u.washington.edu] *On Behalf Of 
> *Surfleet, Christopher
> *Sent:* Wednesday, September 14, 2005 1:48 PM
> *To:* User Support list for DHSVM Model
> *Subject:* RE: [Dhsvm-users] Demon
> 
> Peter,
> I appreciate your feedback.  Many of your points I did not know.  I 
> actually did a little bit of editing on both the DEM and the stream 
> network.  However, what I meant to suggest was that if your stream arc 
> extends too far upslope, in other words the contributing area is set too 
> low for a particular arc, then I would delete some of the arc.  
> My understanding is that the channel would be removed in the DHSVM 
> application.  The low spot in the DEM or topography is still there, but 
> it no longer forces overland flow into a stream channel. 
>  
> An example is createstreamnetwork made a few arcs that just crossed a 
> road in my watershed.  I knew that the road did not have a stream 
> crossing at that point nor overland flow from the upslope swale.  I 
> merely edited the stream arc a but shorter so it no longer crossed the 
> road.  My topography is unchanged, my DEM is unchanged, but the stream 
> channel is merely started a few pixels lower on the slope. 
>  
> Chris
>  
>  
> 
> ------------------------------------------------------------------------
> *From:* dhsvm-users-bounces at mailman1.u.washington.edu 
> [mailto:dhsvm-users-bounces at mailman1.u.washington.edu] *On Behalf Of 
> *Piotr Kuras
> *Sent:* Wednesday, September 14, 2005 1:38 PM
> *To:* 'User Support list for DHSVM Model'
> *Subject:* RE: [Dhsvm-users] Demon
> 
> Hi Chris,
> 
>  
> 
> I have edited the stream network (arcs) for my watershed before, but 
> have been advised by other longtime DHSVM users that I would need to 
> edit the digital elevation model as well – which I found is a lengthy 
> and often unsuccessful task.  If you only edit the stream network the 
> actual flow routing in the DEM is not altered.  Once the stream network 
> is edited it also needs to be “burned” into the DEM and this requires 
> editing each affected elevation point manually.  This is possible, but 
> once you run the watershed tool in GIS the resultant flow routing often 
> does not end up where you want it to.  The biggest problem I had was 
> altering a DEM with elevation points that are not integers, as GIS tools 
> will round to whole numbers, which creates more problems.  A colleague 
> of mine was successful in altering the DEM of a 10 sq.km watershed, but 
> it took him two months.  Have you encountered such problems yourself?
> 
>  
> 
> Peter
> 
>  
> 
> --------------------------------------------------------------
> Piotr Kuraś, FIT
> Department of Forest Resources Management
> University of British Columbia
> --------------------------------------------------------------
> 
> ------------------------------------------------------------------------
> 
> *From:* dhsvm-users-bounces at mailman1.u.washington.edu 
> [mailto:dhsvm-users-bounces at mailman1.u.washington.edu] *On Behalf Of 
> *Surfleet, Christopher
> *Sent:* September 14, 2005 1:03 PM
> *To:* User Support list for DHSVM Model
> *Subject:* RE: [Dhsvm-users] Demon
> 
>  
> 
> The source area for the createstreamnetwork is the area at which you 
> want a stream channel or overland flow to begin.  Thus if you specify 
> 10000 as your source area you are saying that a stream channel will 
> initiate at an area of 1 hectare (10000 m^2).  So the determination of 
> what source area to use should be a decision based on the climate and 
> conditions for the watershed you are working.
> 
>  
> 
> One thing to keep in mind, I did this, is once you create your stream 
> network in createstreamnetwork you can then take the resulting GIS layer 
> and edit the stream arcs within GIS software to more accurately reflect 
> your basin's stream network.  Then run your edited stream network 
> through createstreamnetwork to produce the necessary files for DHSVM use 
> (stream.map.dat, stream.network.dat).
> 
>  
> 
> Chris Surfleet
> 
>  
> 
> ------------------------------------------------------------------------
> 
> *From:* dhsvm-users-bounces at mailman1.u.washington.edu 
> [mailto:dhsvm-users-bounces at mailman1.u.washington.edu] *On Behalf Of 
> *fahad saeed
> *Sent:* Wednesday, September 14, 2005 5:13 AM
> *To:* dhsvm-users at u.washington.edu
> *Subject:* [Dhsvm-users] Demon
> 
> In order to create stream network, in DHSVM FAQ's refer to DEMON in 
> order to estimate source area. Can anybody help me in getting and 
> understanding this?
> 
> ------------------------------------------------------------------------
> 
> Yahoo! for Good
> Click here to donate <http://store.yahoo.com/redcross-donate3/> to the 
> Hurricane Katrina relief effort.
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Dhsvm-users mailing list
> Dhsvm-users at u.washington.edu
> http://mailman1.u.washington.edu/mailman/listinfo/dhsvm-users

-- 
_______________________________________________ 

  Keith Adams,  Ph.D. Candidate
  School of Geography and Environmental Science,
  The University of Auckland		
  Private Bag  92 019, Auckland, New Zealand
  location:  rm. 438  10 Symonds St.
  email:  k.adams at auckland.ac.nz
  phone:  +64-9-373 7599  ext. 84607
  dept. fax:  +64-9-373 7434