Decided 2026-08-18: the climatology product (task 1.1) is built from ERA5 via the Copernicus CDS
API, confirming BLOCKING-5. This issue tracks the deliberate follow-up: whether to add or switch to
another platform later.
Candidates and the reasons they might matter:
- MERRA-2 — an independent reanalysis; a cheap cross-check on ERA5's stratospheric temperatures,
and some groups prefer it above 10 hPa.
- Aura MLS — specifically for water vapour. This is the strongest candidate for an actual
change: reanalysis stratospheric H₂O is known to be biased dry (recorded in BLOCKING-5 and as the
site.h2o_ppmv caveat), so ERA5 is a poor source for the one field the sulfate microphysics is
most sensitive to. A plausible end state is ERA5 for p/T + MLS for H₂O, with the two dataset ids
recorded separately in provenance.
The zonal-mean monthly convention (SCIENCE-1) and the dataset-identifier-in-provenance mechanism
(ADR-006) were designed so a second product can be added without reinterpreting configs made against
the first — this issue is the intended use of that door.
Not blocking anything: task 1.1 proceeds on ERA5.
Decided 2026-08-18: the climatology product (task 1.1) is built from ERA5 via the Copernicus CDS
API, confirming BLOCKING-5. This issue tracks the deliberate follow-up: whether to add or switch to
another platform later.
Candidates and the reasons they might matter:
and some groups prefer it above 10 hPa.
change: reanalysis stratospheric H₂O is known to be biased dry (recorded in BLOCKING-5 and as the
site.h2o_ppmvcaveat), so ERA5 is a poor source for the one field the sulfate microphysics ismost sensitive to. A plausible end state is ERA5 for p/T + MLS for H₂O, with the two dataset ids
recorded separately in provenance.
The zonal-mean monthly convention (SCIENCE-1) and the dataset-identifier-in-provenance mechanism
(ADR-006) were designed so a second product can be added without reinterpreting configs made against
the first — this issue is the intended use of that door.
Not blocking anything: task 1.1 proceeds on ERA5.