I think the problem we have in our current automated processing using the FlowAccumulationFullWorkflow tool is related to a large DEM input (20527x9187 Pixel) and limited memory available on a docker host (Processing runs within a Python container).
I wonder if there were changes made to Whitebox improving this situation. I also think it would be good to not fail gracefully in Whitebox to give a t least some feedback to the user why the tool exists before finishing.
I would be especially interested what solutions others found to address the problem with too large DEM inputs.
I also wonder if using a chunked approach works with this tool as it does both, flow direction and accumulation at the same moment.
In case our issue is really the lack of memory, then maybe a best approach (like chunking) could be directly integrated into Whitebox?
Hi,
I think the problem we have in our current automated processing using the FlowAccumulationFullWorkflow tool is related to a large DEM input (20527x9187 Pixel) and limited memory available on a docker host (Processing runs within a Python container).
Our observation: The FlowAccumulationFullWorkflow step exists gracefully after line 631 (Whitebox Version 2.3.1) https://github.com/jblindsay/whitebox-tools/blob/59d345138f86a50602182a2dd5b0a410449f6511/whitebox-tools-app/src/tools/hydro_analysis/flow_accum_full_workflow.rs#L613 (It still writes the DEM file but then no further output)
Input to that tool is the
I found the following threads already touching the same topic of large DEMs
I wonder if there were changes made to Whitebox improving this situation. I also think it would be good to not fail gracefully in Whitebox to give a t least some feedback to the user why the tool exists before finishing.
I would be especially interested what solutions others found to address the problem with too large DEM inputs. I also wonder if using a chunked approach works with this tool as it does both, flow direction and accumulation at the same moment.
In case our issue is really the lack of memory, then maybe a best approach (like chunking) could be directly integrated into Whitebox?
Thanks for looking into this!