Expand GeoMIP

The Geoengineering Model Intercomparison Project (GeoMIP) is the primary mechanism to coordinate climate model simulations of stratospheric aerosol injection (SAI), and its standardized output underpins most SAI impact assessments. Yet only seven Earth system models (ESMs) currently participate — a fraction of the 100+ models and 35+ modeling groups involved in CMIP, the standard international climate model comparison project. This gap creates two compounding problems: researchers cannot reliably distinguish genuine physical uncertainty about how SAI affects climate from artifacts of model-specific biases, and the small ensemble makes SAI science appear less rigorous than mainstream climate research. Expanding GeoMIP participation directly addresses both.

Three categories of existing ESMs could run GeoMIP experiments but currently do not:

1. Chemistry climate models with interactive stratospheric aerosols and a coupled ocean — models that already have every necessary component but may lack personnel, knowledge, motivation, or computing resources.
2. Models where a CMIP-participating low-top version (one whose simulated atmosphere doesn't fully represent the stratosphere) has a coupled ocean, and a separate non-coupled version represents stratospheric aerosols. All the pieces exist, but combining them for SAI experiments requires additional development work.
3. CMIP-participating models that lack interactive aerosols or a resolved stratosphere entirely; these could run more limited GeoMIP experiments using prescribed aerosols (aerosol effects specified from other models' output rather than simulated directly), but would require substantial model development before doing so.

Reducing barriers in the first two categories through targeted funding, computing support, and technical guidance offers the most tractable near-term path to a larger, more diverse GeoMIP ensemble.

This research activity adds up to ten new Earth system models to the GeoMIP suite by having each simulate the most recent subtropical (G7-1.5K-SAI) and high-latitude (G7-1.5K-HiLLA) injection experiments and their associated preparatory experiments (Visioni et al., 2026). Participating models will come from categories one and two above. A near-term category-one candidate is EMAC (and its successor ICON/MESSy); category-two candidates include models from the MRI, IPSL, CNRM, GFDL, and CanESM families.