Access model parameter experimentation

[smnorris]

1. measure gradient over 25m
2. add max elevation parameter (per spp)
3. experiment with gradient thresholds (per spp) and report on results

Proposed models for evaluation:

model	spp	max grade	max elev
1	CH	20%	1500
1	CM	30%	1000
1	CO	48%	1500
1	PK	30%	1000
1	SK	24%	1500
2	CH	26%	1500
2	CM	20%	1000
2	CO	38%	1500
2	PK	18%	1000
2	SK	40%	1500
3	CH	20%	NA
3	CM	20%	NA
3	CO	25%	NA
3	PK	18%	NA
3	SK	24%	NA

[smnorris]

Note that with salmon being split out into 5 unique models, processing time increases significantly. As the workflows/jobs are currently configured, they will not fit within GHA 6hr limit.
For faster turnaround, it might be worth experimenting with:

• processing natural barriers separately from anthropocentric barriers (so naturally accessible processing does not need to be re-run on typical changes - data fixes/PSCIS submission/etc)
• are there alternatives to bcfishpass.load_dnstr for finding the minimal barriers in the per-species barrier tables and deleting all the rest (ie, a self join) - and is this processing even required, perhaps it is faster to just retain all barriers?

Otherwise, the scripts will have to be further split into separate jobs that complete in <6hrs