Summary
T003 lint check finds 293 reaches where facc decreases downstream on non-bifurcating (1:1) links after the v17c facc correction pipeline.
Root cause
These slipped below the detection threshold of the hybrid facc correction pipeline. They are not residual MERIT Hydro errors — they are the same class of problem the pipeline targets (facc corrupt on 1:1 links) but were not flagged by any of the detection rules (fwr_drop, entry_point, extreme_fwr, jump_entry).
Fix
Use T003 output as a supplemental detection list and feed those reach_ids directly into the facc correction pipeline. Cross-check against the no-facc sanity model to verify corrections are plausible.
Notes
- Release notes updated in 0.0.2: removed false claim that "monotonicity violations on non-bifurcating links no longer occur in any region"
- 293 is small relative to 248,673 total reaches (0.12%)
- Deferred from 0.0.2
Acceptance criteria
- T003 passes with 0 violations
- No regressions on T001, T005, F001–F003
Summary
T003 lint check finds 293 reaches where facc decreases downstream on non-bifurcating (1:1) links after the v17c facc correction pipeline.
Root cause
These slipped below the detection threshold of the hybrid facc correction pipeline. They are not residual MERIT Hydro errors — they are the same class of problem the pipeline targets (facc corrupt on 1:1 links) but were not flagged by any of the detection rules (fwr_drop, entry_point, extreme_fwr, jump_entry).
Fix
Use T003 output as a supplemental detection list and feed those reach_ids directly into the facc correction pipeline. Cross-check against the no-facc sanity model to verify corrections are plausible.
Notes
Acceptance criteria