AnsweredAssumed Answered

Watershed Analysis Issues

Question asked by geonetadmin on Dec 20, 2010
Latest reply on Jan 26, 2011 by geonetadmin
Original User: Jvhowube

Dear GIS Community,

I am having some trouble with watershed analysis.  I know I'm not the first one, but I've searched the internet/forums, and I can't seem to find any exact workthroughs for the issues I'm having.

The eventual goal of this analysis is to create a layer that shows the result of installing hydroelectric in a certain area.  If this is not what watershed analysis is used for, please let me know.  I'm going by a user-guide that I was given here in a developing country, so I'm working with very little.  The before and after pictures show the kind of output that is desired - i.e. if you choke off a part of the watershed with a dam, what is the resulting reservoir?  At least that's what it looks like to me.  Again, I'm new to this watershed stuff so bear with me.

Basically, I understand the toolset, and how it's a step-by-step process to eventually delineate watersheds from a DEM.  My trouble starts when I go to set a pour point (what the heck is a pour point anyway), and then use the Watershed tool.  My understanding of a pour point is as follows: the point in a watershed or catchment where all the water flows out of said watershed or catchment.  When I set this manually by just creating a point file, it's never the "right" point, and any watershed analysis thereafter results in a very small 20-30 raster-cell-sized area around my pour point.

I've tried to remedy this by using the Snap Pour Point tool, and searching the entire area for a "better" pour point - if this isn't what that tool does, then I'm using it wrong.  I can't figure out what it does, if not that.  Anyway, this usually results in an area slightly larger, but still quite small, and it fails to delineate any reasonable watershed-esque area.

Is my issue possibly with DEM resolution?  I've been using Fill to correct the DEM from any Sinks (although I am using a Z of 100, I'm not sure this is correct).  The Flow accumulation raster never actually looks like a river/stream network (i.e. the one that's shown in ArcHelp) - more like a bunch of stringy confetti thrown on the map - there's minimal connectivity, which makes me think there's a problem with the original DEM data.

The guide also has me using a Flow_Accumulation layer that has been reclassed in the watershed analysis (instead of the flow direction raster that the watershed tool asks for).  This results in basically nothing.  But maybe I do have to use an accumulation layer of some sort instead?

I've tried analyzing both a large area (1800 sqkm) (too large?), and then also a catchment section of that larger area.  Both turn out the same broken watershed analysis.

This has served as a source of frustration for the past few weeks.  If you can lend any help, it'd be much appreciated.  Thank you very much.


Before and After .jpgs show the guides depiction of what I should get from this analysis.  I also attached flow_direction layer and flow_acc_with_watershed.