Abstract
Abstract. Over the last couple of years, machine learning parameterizations have emerged as a potential way to improve the representation of subgrid processes in Earth system models (ESMs). So far, all studies were based on the same three-step approach: first a training dataset was created from a high-resolution simulation, then a machine learning algorithm was fitted to this dataset, before the trained algorithm was implemented in the ESM. The resulting online simulations were frequently plagued by instabilities and biases. Here, coupled online learning is proposed as a way to combat these issues. Coupled learning can be seen as a second training stage in which the pretrained machine learning parameterization, specifically a neural network, is run in parallel with a high-resolution simulation. The high-resolution simulation is kept in sync with the neural network-driven ESM through constant nudging. This enables the neural network to learn from the tendencies that the high-resolution simulation would produce if it experienced the states the neural network creates. The concept is illustrated using the Lorenz 96 model, where coupled learning is able to recover the “true” parameterizations. Further, detailed algorithms for the implementation of coupled learning in 3D cloud-resolving models and the super parameterization framework are presented. Finally, outstanding challenges and issues not resolved by this approach are discussed.
Highlights
The representation of subgrid processes, especially clouds, is the main cause of uncertainty in climate projections and a large error source in weather predictions (Schneider et al, 2017b)
The key difference is that the scale separation is not clearly defined as in Lorenz 96 (L96) or SP, but rather downscaling and upscaling are required to get the HR state on the LR model grid and, reversely, apply the forcing term, which is computed on the LR model grid, in the HR model
This is the case in SP where only the LR model prognostic variables are forced during cloud-resolving model (CRM) integration
Summary
The representation of subgrid processes, especially clouds, is the main cause of uncertainty in climate projections and a large error source in weather predictions (Schneider et al, 2017b). Machine learning (ML) has emerged as a potential shortcut which would allow using short-term, high-resolution simulations in order to improve climate and weather models. Simulations with neural networks turned out to be unstable at times. Even if stable, the resulting simulations had biases compared to the reference model. In pre-ML climate model development, biases were reduced by manual tuning of a handful of well-known parameters (Hourdin et al, 2017). With thousands of nonphysical parameters in a neural network, this is no longer possible. I propose coupled online learning as a potential mechanism to tackle these two issues and illustrate the principle using the two-level Lorenz 96 (L96) model, a common (but probably too simple) model of multiscale atmospheric flow (Lorenz, 1995)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
