I am deriving flow paths on a hu12-by-hu12 scale. Using the following workflow, I have successfully derived flow paths for a handful of subwatersheds, so I don't understand why my methods aren't working now.
Step 1: (LAS to DEM) I started by setting my filtered LAS files to ground-only points. I derived a bare earth DEM with the LAS Dataset To Raster tool, using a Triangulation Interpolation Type (Natural_Neighbor, No_Thinning), Output Data Type = FLOAT, and Sample/Cell Size = 1 ft (yeah, it took a while, haha). This produced a DEM that looked pretty good...I even derived contour lines, which were very comparable to pre-existing contour lines. (Attachment 1) (scale = 1:50,000)
Step 2: (DEM RECONDITIONING) I used my derived DEM, a polyline as my stream input (Derived from geomorphon analysis...I don't think the polyline is the problem, it looks great. If my derived streams looked more like my geomorphon-based streams, I would be ecstatic.), a stream buffer of 5, smooth drop/raise of 10, and a sharp drop/raise of 10. (AGREE results = Attachment 2)
Step 3: (Fill Sinks) Filled all sinks. (Attachment 3)
Step 4: (Flow Direction) (Attachment 4)
Step 5: (Flow Accumulation) This is where it looks like stuff starts getting weird. Those big horizontal cuts where the flow paths fork are not accurate. (Attachment 5a) (zoomed in 1:5,000) (same scale DEM = Attachment 5b)
Step 6: (Stream Definition) I just went with Arc Hydro's default, which for this data was 859555 cells (Attachment 6 shows the polyline derived from the Stream Definition raster)
So...what's going on? This method worked very well so far, and here I'm getting some terribly anomalous lines. Any ideas?
The resulting lines look offset. In Attachment 6, the yellow = my flow paths and the pink = NHD. My lines also obviously don’t match my generated DEM. Finally, where my flow paths fork, there are huge horizontal cuts.
What's going on???