Hybrid physics-ML & parameterization
Machine Learning Climate Model Dynamics: Offline versus Online Performance
In one sentence. Compares random forest and neural network moist-physics parameterizations trained on a 3 km global simulation; NN wins offline but crashes within 7 days when coupled.
Abstract
Climate models are complicated software systems that approximate atmospheric and oceanic fluid mechanics at a coarse spatial resolution. Typical climate forecasts only explicitly resolve processes larger than 100 km and approximate any process occurring below this scale (e.g. thunderstorms) using so-called parametrizations. Machine learning could improve upon the accuracy of some traditional physical parametrizations by learning from so-called global cloud-resolving models. We compare the performance of two machine learning models, random forests (RF) and neural networks (NNs), at parametrizing the aggregate effect of moist physics in a 3 km resolution global simulation with an atmospheric model. The NN outperforms the RF when evaluated offline on a testing dataset. However, when the ML models are coupled to an atmospheric model run at 200 km resolution, the NN-assisted simulation crashes with 7 days, while the RF-assisted simulations remain stable. Both runs produce more accurate weather forecasts than a baseline configuration, but globally averaged climate variables drift over longer timescales.
Labels
- Topic
- Hybrid physics-ML & parameterization, Climate emulation
- Type
- Evaluation or intercomparison of existing models
- Method
- Classical ML (trees, GP, linear, kernels)
- Time scale
- 3–15 days, Climate projections / centennial
- Domain
- Global
- Data
- km-scale or high-res simulation
- Status
- Research
- Identifiers
- arXiv:2011.03081 OpenAlex:W3101098605
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.