Register entry: docs/studio/OPEN_QUESTIONS.md → SCIENCE-5. Studio Phase 3. Status: open in principle, answered in code.
The question. Does the entrained background reservoir evolve, or is it static?
What the model does. One box with a static background. The entrained background gas is the initial plume state with dilution_zero_species zeroed and dilution_background pptv values applied as overrides (coupled/driver.py:261-276); the background aerosol is the initial TomasState, held immutable (tstate_bg = tstate, :276). It never evolves photochemically. Entrainment itself is physically correct — background gases and background aerosol genuinely enter the box at the entrainment rate (coupled/dilution.py:81,91), not merely as a sink on plume species.
Standing decision already recorded in this repo (Ali, 2026-07-08, coupled/paper_ensemble/run_ensemble.py:49-55): all future production runs should initialise from a spun-up control run rather than the static background list, as run_60day.py does via runs_60day/frank_control_ic.json. The static list is retained only to reproduce the existing 810-run ensemble.
The genuinely open part. Whether one-box-with-spun-up-IC is sufficient, or whether a two-box model is required. Until that is decided, background_evolves is a schema field whose only accepted value is false, and any other value raises rather than being silently ignored (ADR-005).
Done when: the sufficiency question is answered, and either background_evolves stays a single-valued field with the rationale recorded, or a two-box design gets an ADR.
Register entry:
docs/studio/OPEN_QUESTIONS.md→ SCIENCE-5. Studio Phase 3. Status: open in principle, answered in code.The question. Does the entrained background reservoir evolve, or is it static?
What the model does. One box with a static background. The entrained background gas is the initial plume state with
dilution_zero_specieszeroed anddilution_backgroundpptv values applied as overrides (coupled/driver.py:261-276); the background aerosol is the initialTomasState, held immutable (tstate_bg = tstate,:276). It never evolves photochemically. Entrainment itself is physically correct — background gases and background aerosol genuinely enter the box at the entrainment rate (coupled/dilution.py:81,91), not merely as a sink on plume species.Standing decision already recorded in this repo (Ali, 2026-07-08,
coupled/paper_ensemble/run_ensemble.py:49-55): all future production runs should initialise from a spun-up control run rather than the static background list, asrun_60day.pydoes viaruns_60day/frank_control_ic.json. The static list is retained only to reproduce the existing 810-run ensemble.The genuinely open part. Whether one-box-with-spun-up-IC is sufficient, or whether a two-box model is required. Until that is decided,
background_evolvesis a schema field whose only accepted value isfalse, and any other value raises rather than being silently ignored (ADR-005).Done when: the sufficiency question is answered, and either
background_evolvesstays a single-valued field with the rationale recorded, or a two-box design gets an ADR.