4

Im trying to analyse where water is flowing which rains down to my project area. This is the DEM I have (my project area is bordered blue, the red marks are my pour points):

enter image description here

I learned from another thread (Could bridges prevent ArcMap from creating correct watersheds?) that culverts could disturb the creating of watersheds. So I used the "DEM Reconditioning" tool of ArcHydro to burn the streams:

enter image description here

Then I used this DEM (called it "agreedem") and the following workflow to generate my watersheds (snap distance set to 3 m):

enter image description here

This was my first result:

enter image description here

Since it is my main aim to analyze what I have to do to keep the water inside the center of my area I started building small dams using the "Build Walls" tool of ArcHydro. After that I rerun my watershed workflow (several times with different dams, this was the best result I could get):

enter image description here

I was pretty happy with the result. Then I asked myself: "Why should I use the "burn the streams" thing in this case?". There are no streams with culverts in this area (though I have the situation with the culverts in other areas I'm looking at) so there should be no reason to do so. I deceided to rerun the watershed workflow for my original DEM (called "dtm_orig"):

enter image description here

As you can see, the watersheds are different from the ones I generated with the "agreedem" DEM. So I asked myself how there could be such a great difference in the watersheds between the ones created basing on the "agreedem" and the ones based on "dtm_orig". I couldn't imagine that bigger and deeper ditches / streams (whats the only difference between "agreedem" and "dtm_orig") lead to this results and analyzed all the steps of my "create watershed" workflow.

I then realized that there are huge differences in the DEMs which are generated in the first step by using the "Fill" tool between the "agreedem"...

enter image description here

...and the "dtm_orig" DEM:

enter image description here

This are the values of one of the points where one could see the difference:

enter image description here

Why is there such a big difference between the both filled DEMs?

As suggested by Martin I ran the Sink Evaluation (on the dtm_orig DEM) today. To be honest I don't know how to interpret the results.

This is the "SinkPoly" Layer:

enter image description here

Same with added "SinkDA" Layer:

enter image description here

  • I'll leave the long (and always informative) answer to @whiteboxdev, but it is a fact that Fill sinks is changing your DEM. In this case you probably have a sink in your area of interest, which you need to handle properly. I'd recommend running Sink evaluation, and go on from there. Or use the method in Whitebox for fill sinks, namely to breach sinks instead which leaves less changes to your DEM. – Martin Dec 17 '15 at 7:06
  • 1
    Today I tried running the "Breach depressions" tool in Whitebox GAT and used the result as an input for my workflow (without another fill sink). Now I have a third opinion where my watersheds are but absolutely no clue which one to trust. – Simon Dec 18 '15 at 13:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.