Abstract:Modern climate projections often suffer from inadequate spatial and temporal resolution due to computational limitations, resulting in inaccurate representations of sub-grid processes. A promising technique to address this is the Multiscale Modeling Framework (MMF), which embeds a kilometer-resolution cloud-resolving model within each atmospheric column of a host climate model to replace traditional convection and cloud parameterizations. Machine learning (ML) offers a unique opportunity to make MMF more accessible by emulating the embedded cloud-resolving model and reducing its substantial computational cost. Although many studies have demonstrated proof-of-concept success of achieving stable hybrid simulations, it remains a challenge to achieve near operational-level success with real geography and comprehensive variable emulation that includes, for example, explicit cloud condensate coupling. In this study, we present a stable hybrid model capable of integrating for at least 5 years with near operational-level complexity, including real geography, seasonality, explicit cloud condensate predictions, and land coupling. Our model demonstrates skillful online performance in metrics such as 5-year zonal mean biases compared to previous MMF emulation studies. The monthly error against reference MMF simulations with the same initial condition approaches the fundamental predictability limit. Key factors contributing to our online performance include an expressive U-Net architecture, additional input features that include large-scale forcings and convection memory, and physical thermodynamic constraints for microphysics. With microphysical constraints mitigating unrealistic cloud formation, our work is the first to demonstrate realistic multi-year cloud condensate climatology under the MMF framework. Our work showcases ML parameterization's potential for operational-level climate simulations.

A Deep Learning Earth System Model for Stable and Efficient Simulation of the Current Climate

ACE: A fast, skillful learned global atmospheric model for climate prediction

Probabilistic Emulation of a Global Climate Model with Spherical DYffusion

Earth System Modeling 2.0: A Blueprint for Models That Learn From Observations and Targeted High-Resolution Simulations

Predicting global patterns of long-term climate change from short-term simulations using machine learning

DiffESM: Conditional Emulation of Temperature and Precipitation in Earth System Models with 3D Diffusion Models

A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics

Neural General Circulation Models for Weather and Climate

A Hybrid Deep-Learning Model for El Niño Southern Oscillation in the Low-Data Regime

Data-Driven Weather Forecasting and Climate Modeling from the Perspective of Development

Causal deep learning models for studying the Earth system

A non-intrusive machine learning framework for debiasing long-time coarse resolution climate simulations and quantifying rare events statistics

Diffusion-Based Joint Temperature and Precipitation Emulation of Earth System Models

DiffESM: Conditional Emulation of Earth System Models with Diffusion Models

Iterative Integration of Deep Learning in Hybrid Earth Surface System Modelling

Stable Machine-Learning Parameterization of Subgrid Processes with Real Geography and Full-physics Emulation

DUNE: A Machine Learning Deep UNet++ based Ensemble Approach to Monthly, Seasonal and Annual Climate Forecasting

Deep learning and process understanding for data-driven Earth system science

Learning and Dynamical Models for Sub-seasonal Climate Forecasting: Comparison and Collaboration

CAS-Canglong: A skillful 3D Transformer model for sub-seasonal to seasonal global sea surface temperature prediction