con statement

4836
3
05-04-2015 12:19 PM
MichaelBrown3
New Contributor II

Hello, I am delineating streams using the hydrology tools in the arc toolbox and I have an issue. It is not that I dont know how to do it or what tool to use, but it is a threshold decision. The "conditional statement" defines the minimum number of pixels that are required to "flow" into another pixel in order for that pixel to be delineated as a stream. I have a 2m, high resolution DEM spanning 164 km2.

My question is, if I can put in what ever value I want, how can I remove the subjectivity from that decision? I am currently using a conditional statement of 13,000 pixels (an area of about .1 km). How can I validate this or is there even a way to validate this?

Thank you for your time, Michael

0 Kudos
3 Replies
DanPatterson_Retired
MVP Emeritus

There is no magic number, if it is too low, you will soon find out since you will have way too many 'streams' whether or not they could support water flow or not.  A long time ago, I ran that same analysis on a study area I live in and soon discovered that I was way too generous and ended up with water flow in areas that never had or never seen it.  You might want to qualify your study by expanding your inclusion area incrementally, there will come a point where real streams will be evident and garbage removed.  There are probably methodologies out there but I can assure you they will be partially area dependent as well as terrain and surficial cover dependent..  Good luck and report back with some examples so that others can chirp in.

0 Kudos
SepheFox
Frequent Contributor

It also depends on what category of streams you are trying to model. For instance, do you just want to model perennial streams, or do you want intermittent streams also. The basin size for intermittent streams is obviously small than perennial.

I have recently created modeled streams for our state forests, and I did this by a combination of asking staff to supply points where known streams went from being perennial to being seasonal, and then using those points to calculate a basin size. I then used that basin size in my hydrology analysis as a starting point. I also created several stream layers with smaller and larger basin sizes, and sent all these back to staff asking them to pick the most suitable one.

This basin size does vary by such things as topography, geology, and rainfall, so It's not valid across huge areas. In my case I did it by districts, knowing that it was an average and approximation. You might be able to find information about basin sizes in your area.

0 Kudos
MichaelBrown3
New Contributor II

I am looking for more transient, low order streams. I am conducting this research on the Greenland ice sheet. I cant seem to find any validation for the conditional statement still... It appears you are correct in that there is no magic number. The closest I have found to citable work is here the paper titled:

DEM resolution and stream delineation threshold effects on the results of geomorphologic based rainf...

They use a percentage of the basin area for there conditional statement. The lowest percentage they use is .25% but I have calculated mine and it appears I have used ~ .06% threshold. I think mine is very low in order to capture the smallest streams. (My DEM is high resolution, 2m). So I think its ok?