Previous methods used to count boundary crossing rivers are not effective - they are counting some basins which they shouldn't and missing some they should.
This PR will document a new method to do it more accurately. Unfortunately this will require creating a patch version of LSDJunctionNetwork.
The plan:
[x] Figure out a way to track pruned basins which intersect NDV, but do not have nested basins within them (which may not intersect NDV)
[x] Tidy up and document the code which has been added
[x] Rewrite the script to analyse the cross counts
I was initially planning to configure this to run as a separate series of jobs, distinct from the main gdbm code, but I think it might be simpler to run it within the existing framework - these results are not time critical, so I can wait for the current jobs to finish and use the results from that for next stages of work, before re-running everything to get these new data.
Previous methods used to count boundary crossing rivers are not effective - they are counting some basins which they shouldn't and missing some they should.
This PR will document a new method to do it more accurately. Unfortunately this will require creating a patch version of LSDJunctionNetwork.
The plan:
I was initially planning to configure this to run as a separate series of jobs, distinct from the main gdbm code, but I think it might be simpler to run it within the existing framework - these results are not time critical, so I can wait for the current jobs to finish and use the results from that for next stages of work, before re-running everything to get these new data.