Climate models

Corrective machine learning to improve climate models

Editors’ Highlights are summaries of recent articles written by AGU’s journal editors.
Source: Journal of Advances in Earth System Modeling

A new generation of global atmospheric models with 1-5 kilometer grid spacing can accurately simulate extreme weather conditions, such as heavy local rainfall from thunderstorm systems and flow through complex mountain ranges. They could also help us better plan for the local-scale impacts of future climate change, but they are far too computationally heavy to be used for simulations of decades or centuries. Bretherton et al. [2022] present a machine learning approach to correcting coarse-grid climate models that we can afford to run using the results of short benchmark simulations with fine-grid climate models. The correction allows the coarse-grid model to more closely follow weather forecasts and time-averaged precipitation patterns from the baseline simulation.

Quote: Bretherton, CS, Henn, B., Kwa, A., Brenowitz, ND, Watt-Meyer, O., McGibbon, J., et al. (2022). Correcting coarse-grid weather and climate models by machine learning from global storm resolution simulations. Journal of Advances in Earth System Modeling14, e2021MS002794.

―Jiwen Fan, Editor, JAMES

Text © 2022. The authors. CC BY-NC-ND 3.0
Unless otherwise stated, images are subject to copyright. Any reuse without the express permission of the copyright holder is prohibited.