Methodology

Upstream property — not independently re-verified by SHFT v1.0

Evidence basis · Carbon conservation is a property of the pinned upstream SURFER v3.0 formulation and is evidenced only by the reproduced upstream carbon-budget benchmark (carbon_budget_2000_2009_and_cumulative). The earth-state-0.2 schema stores no carbon reservoir masses. Of the two catalogued flux-like quantities, net_primary_productivity is defined but not represented; weathering_rate is represented. SHFT therefore holds no complete reservoir/flux inventory and cannot construct an independent closure test. Completeness is not inferred.

Required to promote · Store the complete upstream carbon reservoir and flux diagnostics per stored time point in a new schema version, then add an engine-0.2+ criterion computing a residual against a sourced tolerance.

Independent SHFT verification · inherited upstream

Active — evaluated for engine-0.2

Evidence basis · Evaluated for every engine-0.2 run by evaluate_engine_gates within its stated scope.

Required to promote · None; already active. Any widening of scope requires a new criterion and tolerance set.

Independent SHFT verification · performed

Upstream property — not independently re-verified by SHFT v1.0

Evidence basis · Energy closure is a property of the pinned upstream formulation. The adapter/payload stores only global mean surface temperature and ocean heat content; no top-of-atmosphere or component energy fluxes are stored, so no auditable closure residual can be computed by SHFT v1.0.

Required to promote · Store the complete upstream energy-flux inventory (top-of-atmosphere imbalance and component fluxes) per stored time point, then add a criterion checking the imbalance against a sourced tolerance.

Independent SHFT verification · inherited upstream

Deferred — not independently validated in v1.0

Evidence basis · Deferred in v1.0. No reconstruction-based ice benchmark has been sourced or reproduced for the pinned backbone, and no engine-0.2 criterion evaluates ice volume against any target. No pass is inherited from tol-0.1.

Required to promote · A cited deglaciation ice-volume reconstruction with a declared tolerance, an executed adapter-path run over that interval, and an engine-0.2+ criterion comparing served ice_sheet_volume output against it.

Independent SHFT verification · not performed

Deferred — not independently validated in v1.0

Evidence basis · Deferred in v1.0. The upstream publication compares thermosteric sea level with IPCC evidence, but no sourced, reproducible adapter-path comparison exists in SHFT and no engine-0.2 criterion evaluates it. The seeded tol-0.1 demonstration result is hand-authored and is not evidence about the SURFER-backed engine.

Required to promote · A sourced tide-gauge or reconstruction target with a declared tolerance, an executed historical adapter-path run, and an engine-0.2+ criterion comparing served sea_level output against it.

Independent SHFT verification · not performed

Upstream property — not independently re-verified by SHFT v1.0

Evidence basis · Upstream reproduction evidence only. The reproduced pinned-model benchmarks atmospheric_co2_curve_1850_2015_rmse_ppm and gmst_anomaly_2014_k compare upstream output with published historical evaluation, and are recorded under upstream_reproduction. No engine-0.2 criterion compares served SHFT adapter-path output against a sourced historical temperature target, so there is no independent SHFT hindcast gate.

Required to promote · A stored historical-period adapter-path run with a sourced observational temperature target and declared tolerance, plus an engine-0.2+ criterion comparing served SHFT output against it.

Independent SHFT verification · inherited upstream

Active — evaluated for engine-0.2

Evidence basis · Evaluated for every engine-0.2 run by evaluate_engine_gates within its stated scope.

Required to promote · None; already active. Any widening of scope requires a new criterion and tolerance set.

Independent SHFT verification · performed

Active — evaluated for engine-0.2

Evidence basis · Evaluated for every engine-0.2 run by evaluate_engine_gates within its stated scope.

Required to promote · None; already active. Any widening of scope requires a new criterion and tolerance set.

Independent SHFT verification · performed

Active — evaluated for engine-0.2

Evidence basis · Evaluated for every engine-0.2 run by evaluate_engine_gates within its stated scope.

Required to promote · None; already active. Any widening of scope requires a new criterion and tolerance set.

Independent SHFT verification · performed

Deferred — not independently validated in v1.0

Evidence basis · Deferred in v1.0. No long-run SURFER-backed stability experiment has been executed and stored: the longest reproduced evidence is the LTMIP long-tail CO2 comparison, and the interface offers horizons far short of the 1 Myr integration the catalogued gate describes. The seeded tol-0.1 demonstration does not count.

Required to promote · A defined and executed long-run adapter-path experiment at the declared resolution with stored trajectories, plus an engine-0.2+ criterion checking for numerical blow-up over that integration.

Independent SHFT verification · not performed

Active — evaluated for engine-0.2

Evidence basis · Evaluated for every engine-0.2 run by evaluate_engine_gates within its stated scope.

Required to promote · None; already active. Any widening of scope requires a new criterion and tolerance set.

Independent SHFT verification · performed

Active — evaluated for engine-0.2

Evidence basis · Evaluated for every engine-0.2 run by evaluate_engine_gates within its stated scope.

Required to promote · None; already active. Any widening of scope requires a new criterion and tolerance set.

Independent SHFT verification · performed

Not applicable to the active production backbone

Evidence basis · SURFER v3.0, the sole active production backbone, models no hydrological cycle and SHFT Layer 1 stores no water inventory. The catalogued river_discharge quantity is defined but not represented. There is no quantity against which water mass closure could be tested, so the gate is outside the scope of the active backbone and must never be displayed as passed.

Required to promote · A production backbone that resolves a water inventory (reservoir masses and fluxes) plus a sourced closure tolerance. Not achievable by adapter work on SURFER v3.0.

Independent SHFT verification · not applicable