Cover all SIM_MET RUNs against MET model (#99 Bucket B)

[simnaut]

Summary

Closes the SIM_MET item of #99 Bucket B (was "1 of 9 RUNs").

SIM_MET ships eight model-exercising RUNs whose names suggest three atmosphere models (MET_VER / GRAM_MET / JAC_COMP), but its S_define instantiates exactly one model: jeod::METAtmosphere. JEOD v5.4 has no GRAM or Jacchia model anywhere — models/environment/atmosphere/ contains only MET and base_atmos, and a tree-wide grep for gram|jacchia returns nothing. The GRAM_MET / JAC_COMP suffixes name the altitude/latitude sample schedule used by external comparison studies, not a distinct model under test. Every RUN logs the same MET density/temperature via vehicle.atmos_state.

All RUNs are therefore portable Tier-2 cross-checks against our MetAtmosphere. This adds the four previously-uncovered sweeps:

RUN	Sweep	Notes
RUN_T01_GRAM_MET	100–1500 km altitude	extends ceiling 500 km past T02_MET_VER, into deep-hydrogen branch
RUN_T02_GRAM_MET	latitude -90°..+90° @ ~650 km	includes the poles
RUN_T01_JAC_COMP	100–1000 km altitude	Jacchia-comparison grid
RUN_T02_JAC_COMP	100–500 km altitude	denser low-altitude grid

Reference CSVs regenerated incrementally via the Docker xtask regenerate-tier3 flow (config from JEOD source only) and committed. CSV headers verified to carry all six expected columns (no DRAscii drop).

Tolerances

Observed max errors sit at the f64 noise floor (density rel ~3e-15, temperature ~1.1e-12 K). The latitude sweep reaches ~1440 K exospheric temps where one f64 ULP is ~2.3e-13 K, so its temperature tolerance is 1.2e-12 K (1.05× observed) vs the 1e-12 K floor elsewhere; density stays at 1e-12.

Not ported

RUN_data_compare re-runs RUN_T01_MET_VER's inputs only to confirm JEOD's several internal MET implementations (preferred vs deprecated AtmosphereState/METAtmosphereState) agree with each other — it adds no new sample point for us to validate. Documented in the module header.

Disposition

7 of the 8 model-exercising RUNs are now genuinely cross-validated against our MET model; the 8th (RUN_data_compare) is a redundant intra-JEOD self-consistency check. Nothing is out-of-scope-non-portable — there is no GRAM/Jacchia physics to port.

Verification

• cargo nextest run -p astrodyn_atmosphere — 39 passed
• cargo fmt --check — clean
• cargo clippy -p astrodyn_atmosphere --tests -- -D warnings — clean

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Expands Tier-2 JEOD cross-validation coverage for the MetAtmosphere implementation by adding the previously unvalidated SIM_MET RUN sweeps (GRAM/JAC schedules) and committing regenerated JEOD reference CSVs for those RUNs.

Changes:

• Add four new Tier-2 tests to validate additional SIM_MET altitude/latitude sweeps against JEOD reference outputs.
• Commit new JEOD reference CSVs for the added RUNs (GRAM altitude/latitude and JAC comparison grids).
• Update the Trick reference-generation script to include the additional SIM_MET RUNs (but see comment re: JAC required output filenames).

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
trick/generate_references.sh	Adds new SIM_MET RUNs to the reference-generation group (skip-check filenames for JAC runs need correction).
crates/astrodyn_atmosphere/tests/tier2_atmosphere.rs	Adds Tier-2 test cases for the additional SIM_MET RUN sweeps and documents the GRAM/JAC naming.
crates/astrodyn_verif_jeod/test_data/met_t01_gram_met.csv	New JEOD reference output for RUN_T01_GRAM_MET (100–1500 km sweep).
crates/astrodyn_verif_jeod/test_data/met_t02_gram_met.csv	New JEOD reference output for RUN_T02_GRAM_MET (latitude sweep).
crates/astrodyn_verif_jeod/test_data/met_t01_jac_met.csv	New JEOD reference output for RUN_T01_JAC_COMP (100–1000 km sweep).
crates/astrodyn_verif_jeod/test_data/met_t02_jac_met.csv	New JEOD reference output for RUN_T02_JAC_COMP (100–500 km sweep).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.