Home on Emulated and Hybrid ML-Physics Earth System Models

Save the date: 12–14 May 2027 at NSF NCAR

Fri, 16 Oct 2026 00:00:00 +0000

The workshop on Emulated and Hybrid ML-Physics Earth System Models will take place on 12–14 May 2027 at the NSF NCAR Mesa Laboratory in Boulder, Colorado. The meeting will convene hybrid model developers, climate emulator developers, and researchers working at their interface to define a staged path toward trustworthy coupled hybrid models and emulators.

The call for contributions, registration, and the detailed program will be announced on this site.

About the workshop

Mon, 01 Jan 0001 00:00:00 +0000

Dates: 12–14 May 2027
Venue: NSF NCAR Mesa Laboratory, Boulder, Colorado
Size: approximately 50–80 participants
Format: plenary and contributed talks, a poster session and reception, walk-and-talks, and working groups

Scientific motivation

Machine learning now produces climate-relevant emulators that are fast, skillful, and increasingly stable in long rollouts; hybrid model components that augment process-based Earth system models; and the first fully differentiable physics-based climate and weather models. Coupled climate modeling, however, sets a stricter standard than matching observed metrics or forecast skill. A coupled system must exchange fluxes and state variables across components, preserve essential budgets, generate credible internal variability, and remain stable over many simulated years. Climate modeling also requires extrapolation beyond the observational record: the model must represent feedbacks and respond physically to forcing.

Agenda

Mon, 01 Jan 0001 00:00:00 +0000

The detailed program will be posted here as speakers and sessions are confirmed. All times will be given in Mountain Daylight Time (MDT, UTC−6).

Day 1 · Define scope, trust, and readiness

Wednesday, 12 May 2027

What do hybrid ML-physics models and climate emulators each contribute, where do their requirements overlap, and what is ready now under an explicit readiness scale? Focus on shared vocabulary, failure modes, budgets, drift, feedbacks, long-rollout credibility, and a candid readiness ladder.

Call for contributions

Mon, 01 Jan 0001 00:00:00 +0000

The call for contributions has not yet opened. Submission instructions, deadlines, and any travel support will be posted here and announced under News.

Scope

We welcome contributed talks and posters across three threads:

Hybrid ML-physics models — ML parameterizations, hybrid CAM/CESM components, learned closures, stochastic parameterizations, differentiable or corrected physics, and process-constrained ML.
Climate emulators — whole-atmosphere, ocean, and coupled emulators; response emulators; surrogate large ensembles; and fast forced-response or scenario tools.
Shared coupling and evaluation — coupler interfaces, conservation, low-order tests, idealized benchmarks, online diagnostics, uncertainty quantification, software, intercomparison, and governance.

Readiness self-assessment

Each submission should state the readiness level of the system it describes, from R0 (conceptual or offline prototype) to R6 (multi-decadal / centennial coupled response), using the workshop readiness scale. Early-stage work (R0–R2) is explicitly welcome; the scale exists to make claims comparable, not to filter contributions.

Organizers

Mon, 01 Jan 0001 00:00:00 +0000

Will Chapman Department of Atmospheric and Oceanic Sciences, University of Colorado Boulder
David John Gagne Computational and Information Systems Lab, NSF National Center for Atmospheric Research (NSF NCAR)
Aneesh Subramanian Department of Atmospheric and Oceanic Sciences, University of Colorado Boulder
Linnia Hawkins Department of Earth and Environmental Engineering, Columbia University; NSF NCAR
Andrew Gettelman Pacific Northwest National Laboratory; University of Colorado Boulder

Speakers

Mon, 01 Jan 0001 00:00:00 +0000

The program includes plenary and highlight talks alongside contributed talks selected from submitted abstracts.

To be announced.

Themes and readiness scale

Mon, 01 Jan 0001 00:00:00 +0000

Workshop themes

Hybrid ML-physics model development
Where should ML enter the Earth system model — parameterization, full component, correction term, stochastic closure, solver accelerator, data-assimilation layer, or ensemble generator?
Climate emulators and surrogate systems
What is being emulated — a process tendency, a component, a full-state trajectory, a coupled system, a response surface, or an ensemble distribution?
Coupling as the bridge
How do hybrid components and emulator components exchange fluxes and states without hiding physical inconsistencies?

Venue and travel

Mon, 01 Jan 0001 00:00:00 +0000

The workshop takes place at the NSF NCAR Mesa Laboratory, 1850 Table Mesa Drive, Boulder, Colorado (map).

The Mesa Laboratory sits at approximately 1,860 m (6,100 ft) above sea level at the base of the Flatirons. Visitors arriving from low elevation should stay well hydrated and allow for reduced exertion during the first day.

Getting there

Most participants will fly into Denver International Airport (DEN), about 70 km from Boulder. Ground transportation and parking details will be posted here.

Home on Emulated and Hybrid ML-Physics Earth System Models

Save the date: 12–14 May 2027 at NSF NCAR

About the workshop

Scientific motivation

Agenda

Day 1 · Define scope, trust, and readiness

Call for contributions

Scope

Readiness self-assessment

Organizers

Speakers

Themes and readiness scale

Workshop themes

Hybrid ML-physics model development

Climate emulators and surrogate systems

Coupling as the bridge

Venue and travel

Getting there