Watershed delination

[ErnestDankwah]

I am trying to delineate watershed for road culverts using the culverts as pour points. I have 3, 10 and 30 meter resolution DEMs
which I downloaded from national map viewer for my study area. I snapped the pour points to specified distances; 10m distance for
the 3m resolution DEM 20m distance for the 10m DEM, and 50m distance for the 30m resolution DEM.

Can someone please help me how to get a reasonable catchment for each culvert? I am interested mainly in the little catchments
that drain into a particular point(s). I have tried the usual ways of delineating watersheds; creating flow direction, flow accumulation and using the snapped points to generate to watershed using the watershed tool. My problem is some of the catchments are way toobig for just one culvert. Can someone please help me how I can do this? Thanks.

[MarkBoucher]

Do you have a contour map that you can compare you GIS results to to see if the GIS results are reasonable?

[ErnestDankwah]

My study are covers 7 counties. I do not have a contour map map to compare my results to. You have any ideas in mind? Thanks

[MarkBoucher]

One of the ESRI base maps might have enough detail for you. See image that shows the toolbar you can use to get them. They load over the internet and can be slow, so zoom to your area first.

[ErnestDankwah]

Can you please tell me how it will that help me delineating the water for the culverts?? Thanks

[MarkBoucher]

Can you please tell me how it will that help me delineating the water for the culverts?? Thanks

This will help you by checking your assumption that the watershed coming out of the processing is too big. Maybe its not.

In the old days, watershed were delineated by hand using contour maps. A contour map of the area, however old, should be somewhat consistent with a modern day DEM. If I have a watershed that looks "wonky" I will use some other data source to check it against. Contours are a good start. The rules are that the watershed boundary must cross the contours perpendicularly. Flow direction is also perpendicular to the contours. If you have a GIS generated watershed, it should overlay on the contours (polylines or image) and the watershed boundary should cross the contours perpendicularly. If not, something is wrong. I suspect this is elementary to you. There are docs online that explain this.

I've attached an image of a recent project I did with the watershed boundaries plotted over Topographic basemap layer from ESRI.

[SharonHeller]

I am trying to delineate subwatershed from a DEM. The results are quite wonky. I suspect that when it calculates the flow direction it is coming up with erroneous results. The flow direction grid looks weird.

I've tried using a variety of point data, even so much as turning my stream arcs into points and I am getting stripes, and areas that don't look anything like watersheds.

[MarkBoucher]

Are you following a process that includes filling in the sinks?

[SharonHeller]

no, I didn't even know there was such a process.

I was wondering if I need to clip my DEM to my watershed area. Can it be too large?

What is the sink fill process? I just used the Spatial Analyst watershed tool. I think when it processes the DEM to get the flow direction there is some kind of problem.

[MarceloOyuela1]

You should burn the DEM with your river network too.

Regards