Climbing down Charney’s ladder: machine learning and royalsocietypublishing.org/journal/rsta the post-Dennard era of computational climate science Discussion V. Balaji1,2 Cite this article: Balaji V. 2021 Climbing down 1Princeton University and NOAA/Geophysical Fluid Dynamics Charney’s ladder: machine learning and the Laboratory, Princeton, NJ, USA post-Dennard era of computational climate 2Institute Pierre-Simon Laplace, Paris, France science. Phil.Trans.R.Soc.A379: 20200085. https://doi.org/10.1098/rsta.2020.0085 VB, 0000-0001-7561-5438 The advent of digital computing in the 1950s sparked Accepted: 30 July 2020 a revolution in the science of weather and climate. Meteorology, long based on extrapolating patterns One contribution of 13 to a theme issue in space and time, gave way to computational ‘Machine learning for weather and climate methods in a decade of advances in numerical weather forecasting. Those same methods also gave modelling’. rise to computational climate science, studying the behaviour of those same numerical equations Subject Areas: over intervals much longer than weather events, artificial intelligence, computational physics, and changes in external boundary conditions. atmospheric science, climatology, Several subsequent decades of exponential growth in meteorology, computational mathematics computational power have brought us to the present day, where models ever grow in resolution and Keywords: complexity, capable of mastery of many small-scale computation, climate, machine learning phenomena with global repercussions, and ever more intricate feedbacks in the Earth system. The current juncture in computing, seven decades later, heralds Author for correspondence: an end to what is called Dennard scaling, the physics V. Balaji behind ever smaller computational units and ever e-mail:
[email protected] faster arithmetic.