A Hybrid Atmospheric Model Incorporating Machine Learning Can Capture Dynamical Processes Not Captured by Its Physics‐Based Component
Open Access
- 26 April 2023
- journal article
- research article
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 50 (8)
Abstract
It is shown that a recently developed hybrid modeling approach that combines machine learning (ML) with an atmospheric global circulation model (AGCM) can serve as a basis for capturing atmospheric processes not captured by the AGCM. This power of the approach is illustrated by three examples from a decades‐long climate simulation experiment. The first example demonstrates that the hybrid model can produce sudden stratospheric warming, a dynamical process of nature not resolved by the low resolution AGCM component of the hybrid model. The second and third example show that introducing 6‐hr cumulative precipitation and sea surface temperature (SST) as ML‐based prognostic variables improves the precipitation climatology and leads to a realistic ENSO signal in the SST and atmospheric surface pressure.Keywords
All Related Versions
Funding Information
- Defense Advanced Research Projects Agency (HR00112290035)
This publication has 27 references indexed in Scilit:
- Prognostic Validation of a Neural Network Unified Physics ParameterizationGeophysical Research Letters, 2018
- Could Machine Learning Break the Convection Parameterization Deadlock?Geophysical Research Letters, 2018
- Hybrid forecasting of chaotic processes: Using machine learning in conjunction with a knowledge-based modelChaos: An Interdisciplinary Journal of Nonlinear Science, 2018
- Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing ApproachPhysical Review Letters, 2018
- A Practical Guide to Applying Echo State Networks2012
- Global Energy and Water Budgets in MERRAJournal of Climate, 2011
- Reservoir computing approaches to recurrent neural network trainingComputer Science Review, 2009
- A New Look at Stratospheric Sudden Warmings. Part I: Climatology and Modeling BenchmarksJournal of Climate, 2007
- Decadal interactions between the western tropical Pacific and the North Atlantic OscillationClimate Dynamics, 2005
- Atmospheric simulations using a GCM with simplified physical parametrizations. I: model climatology and variability in multi-decadal experimentsClimate Dynamics, 2003