The Graft-Versus-Host Problem for Data-driven Gravity Wave Parameterizations in a 1D Quasi-Biennial Oscillation Model

Official Version

Two key challenges in the development of data-driven gravity wave parameterizations are generalization, how to ensure that a data-driven scheme trained on present-day climate will continue to work in a new climate regime, and calibration, how to account for biases in the “host” climate model. Both problems fundamentally depended on the response to out-of-sample inputs, compared to the training dataset, and are often conflicting. The ability to generalize to new climate regimes often goes hand in hand with sensitivity to model biases. To probe these challenges, we employ a 1D Quasi-Biennial Oscillation (QBO) model with a stochastic source term that represents convectively generated gravity waves in the Tropics with randomly varying strengths and spectra. We employ an array of machine learning models consisting of a fully connected feed-forward neural network, a dilated convolutional neural network, an encoder-decoder, a boosted forest, and a support vector regression model. Our results demonstrate that data-driven schemes trained on “observations” can be critically sensitive to model biases in the wave sources. While able to accurately emulate the stochastic source term on which they were trained, all of our schemes fail to fully simulate the expected QBO period or amplitude, even with the slightest perturbation to the wave sources. The main takeaway is that some measures will always be required to ensure the proper response to climate change and to account for model biases. We examine one approach based on the ideas of optimal transport, where the wave sources in the model are first re-mapped to the observed one before applying the data-driven scheme. This approach is agnostic to the data-driven method and guarantees that the model adheres to the observational constraints, making sure the model yields the right results for the right reasons.