Representing Moist Convection with a Collection of Linear Response Functions

How to represent the effect of moist convection in large-scale models is a longstanding problem. A number of recent studies have shown the promise of machine learning approaches such as random forest and deep neural net (Brenowitz and Bretherton 2018; O’Gorman and Dwyer, 2018; Rasp et al., 2018). I will present an alternative trajectory piecewise linear approach based on a collection of linear response functions and discuss the pros and cons of both approaches.

References:

Brenowitz, N. D., & Bretherton, C. S. ( 2018). Prognostic validation of a neural network unified physics parameterization. Geophysical Research Letters, 45, 6289– 6298.

O'Gorman, P. A., & Dwyer, J. G. ( 2018). Using machine learning to parameterize moist convection: Potential for modeling of climate, climate change, and extreme events. Journal of Advances in Modeling Earth Systems, 10, 2548– 2563.

Rasp, S., Pritchard, M. S., & Gentine, P. ( 2018). Deep learning to represent sub‐grid processes in climate models. Proceedings of the National Academy of Sciences, 39, 9684– 9689.