Depression Evaluation completes with Float Raster as input but not Integer Raster

[GarretDuffy]

I am trying to extract depressions from a DEM (1600 x 2400 pixels). The following workflow works, but only using the float raster as input:

1. Project to UTM; resample (cubic); specify vertical coordinate system;

2. Run Depression evaluation

3. Done.

For a Float Raster, the run-time is 4 hours.  To speed this up, I am trying to use integer raster as input.

I then rescale the pixel values to integer centimetres and define the vertical coordinate system accordingly with the appropriate scale factor but when using this as input, the procedure only completes for a small part of the raster, when I set the extent to the full raster, it hangs at the "Computing drainage area for depressions..." (RasterToPolygon)

step. Note that the raster is 32 bit signed, if that makes any difference.  I don't think 16 bit unsigned will have the range I need.

I've also nullified zero pixels that appear around the edges after projecting the data to UTM from Geographic but no use.  I am not sure what is going on and would be grateful for some advice.

Thanks!

[ChristineDartiguenave · Accepted solution]

The performance issue seems to be caused by the joins. I have migrated the tool to Python and added it to the Arc Hydro Tools Python toolbox in versions 10.5 and later as well as Pro. Give it a try. The process should now complete in a couple of minutes. For big datasets, you may need to run in background mode to complete successfully (in foreground you may see the dreaded 9999 error otherwise).

Christine Dartiguenave

Esri Water Resources Team

[GarretDuffy]

Thanks so much for migrating the tool to Python.  The tools works so much faster now for both float and integer rasters.  In fact, I don't need to integerise the DEM raster now it is so much faster.

Best,

Garret Duffy