Tier II: Volcanic response — Pinatubo (1991–1993)

[duncanwp]

Description

Evaluate whether models reproduce a realistic volcanic signature during the post-Pinatubo period.

Specification

• Assess from historical simulation (no separate volcanic experiment)
• Ensemble-mean anomalies of surface SW↓ and near-surface temperature for 1991–1993
• Require correct co-variation of radiative and thermal anomalies (not in isolation — guards against overfitting)
• Compare against BSRN-derived and observational estimates

Required variables

rsds, tas

Required simulations

historical

References

Protocol §Tier II — Volcanic response

[duncanwp]

Proposed solution: partially reuse ICONEval's timeseries infrastructure

ICONEval does not have a dedicated Pinatubo diagnostic, but recipe_basics_timeseries provides the preprocessor chain for global-mean tas and TOA flux timeseries with multi-obs references.

• Reuse the timeseries preprocessor infrastructure (regridding, area statistics, multi-obs references)
• Write a new diagnostic script that isolates the 1991-1993 window and computes:
  • Peak cooling magnitude vs observed ~0.5 K drop
  • TOA SW anomaly recovery timescale
  • Ensemble spread during the volcanic response window
• This is partially ICONEval-derived (infrastructure) but requires a custom diagnostic

Phase: 3 (Tier II observational benchmarks) in the integration plan
ICONEval reference: recipe_basics_timeseries.yml (preprocessor chain only)