Arc Hydro-Depression Evaluation Trouble

[angeldybas]

Hi,

I trying to delineate subwatersheds using Arc Hydro. I'm using a DEM derived from LIDAR. So far, I've converted the DEM to an integer DEM  and run the sink evaluation. The size of the prefilled DEM that I'm trying to run the depression evaluation on is 8.86 GB. My problem is when I run the depression evaluation, it runs for a while (over 8 hrs) then gives me the message "Build VAT might have failed make sure raster target location is not GDB and path does not contain space." Does this simply just mean I have to change the path of where it stores the depression data? Would I do this under "Set target locations for HydroConfig, Layers?" I looked back at some other depression evaluations and their all located in the geodatabase in layers so I'm not sure what the issue is. Is the DEM too big?

Any input would be appreciated.

Thanks.

[CatherineTulley]

I can't place the source right now, but I've read that ArcHydro has trouble processing high-resolution rasters. I had success resampling our 3.2ft-lidar-derived DEM (~3+ gb) to 10-ft resolution (300+ MB). I burned in a high-resolution stream network after running the dem evaluation & correction steps, and have been able to successfully set up a drainage area tool for a county-level study area.

If you don't necessarily need the higher resolution, that might be something to try?

[JeffLee]

Angel;

     If I am not mistaken, the maximum file size that the grid module can handle is 2GB.  However, on the practical side is probably half that size.

[MiguelGuerrero]

I am having the same problem, but I know for sure that it is not the size of the DEM because mine is only 46 MB.


Miguel

[JoshuaWolf]

Try converting your raster to 32 bit floating point using the copy raster tool and then rerun.  When depression evaluation encounters an integer grid it has to build VAT files and unless you change the VAT limits in the "secret esri settings" it will choke; floating point doesn't require the VATs.  I encountered the same problem using a 5m LIDAR derived DEM for a HUC8 watershed and this was my work around.