Deriving Depth Information from Polygon and DEM

[DarrochKaye]

Hi folks,

I have a LiDAR dataset (ESRI grid / .asc format) and a polygon showing a flood extent. What I want to do is extract the depth information, i.e. space between polygon and DEM, to calculate the flood depth. Unfortunately the polygon has no elevation information, so I have nowhere to start from.

Is there a way to assume the level where the polygon intersects with the LiDAR? My theory is that at the extreme edges of the polygon, where the flood extent finishes (the watermark), the corresponding LiDAR data will be the level of the polygon at that location. If this can be repeated at set intervals around the extent of the polygon, then I could at least start from there.

Any help you could provide would be greatly appreciated.

Thanks,
DK

[MattMead]

If you want the exact elevation where the polygon intersects with the topography, you can convert the polygon to points(ET geowizards or similar) to get points everywhere there is a vertex or use a profiler(EZprofiler is free and works well for me. works in 10 http://arcscripts.esri.com/details.asp?dbid=13688) to set points every X map units.  Once you have these points you can use spatial analyst->Extract Values to Points to get the elevations at those points.

The problem then is how do you set the elevations through the middle of the flood plain.  I'm not sure if there is an automated way of doing this or not, but personally I would try and draw in cross sections where you think the water surface is likely to be about the same across your floodplain.  Estimate the elevation since it wont be exactly the same on both sides of the floodplain at your cross section, then create a TIN using these cross sections and elevations.  Either convert your topo to a TIN or your water surface TIN to a grid(probably the easier option) and use raster calc/spatial analyst to get the difference/depth grid.

Quick note if you use grids: make sure to use a snap grid(check the help files on setting it up) or else you may have incorrect depths.  When you create a grid/raster it will create it using the extents of the base data, so if you have a 5ft cell size starting at 0,0 and another 5ft cell starting at 1,1 there is a 1ft offset in your data in both the x+y.  Snap grid will set the start of your grid to the same location so even if you data starts at 1,1 it will begin your 5ft cells at 0,0. 

Might be a minor difference and the accuracy level you are likely to get by approximating makes it a wash most likely, but its a good practice to have.

[kbowe94]

I know this post is from 2012 but did anyone try this suggested method with success? I have the same problem where I have DEMs and the 100yr floodplain polygon, but no accurate depth information or cross sections. I basically want to subtract the 100yr flood water depth from the DEM and the normal stream height, for an area covering an entire county (in the united states). There's supposed to be a way to set this up with a NFHL dataset for Hazus analysis, but the county I am working in doesn't have cross sections for the entire floodplain in the NFHL dataset, so the resulting "depth grid" from the Hazus method is much smaller than the actual floodplain and won't accurately predict damage from a simulated flood.

[VrutikaGamit]

Did anyone tried the method? I have the same situation.