2026-08-12 –, Room 3
Traditionally, climate models are difficult to run for end users, and even harder to customize or interface with machine learning. We want to change that. Here, we present the ongoing development of SpeedyWeather.jl: A general circulation model that’s differentiable, GPU-capable and ready for machine learning integration. SpeedyWeather.jl is a spectral atmospheric general circulation model with an everything-flexible attitude. In this talk, we will give an overview of SpeedyWeather.jl’s development of the last year, in which we worked towards differentiability with Enzyme, GPU-capability with KernelAbstractions and Reactant and rewrote our parametrizations for better performance and more customisability.
The current generation of usually Fortran-based climate models presents a high entry barrier to climate modelling. Julia on the other hand gives us the tools to write climate models that are performant, but at the same time easy to use for both end users and developers. This extends to differentiable programming and GPU programming as well. In the past year, we spent a considerable effort on making SpeedyWeather.jl differentiable with Enzyme, GPU-capable with KernelAbstractions and Reactant. We also extended our process implementations by adding new parameterizations of radiation, and simple sea ice, land and snow models to run climate simulations with SpeedyWeather.jl. In this talk, we will give an overview on the changes we had to make for this. These changes also enabled us to redesign parts of our model for even better customisability and composability, as demonstrated for example by a new parametrization system. Furthermore, we will give an outlook on using SpeedyWeather.jl differentiability for sensitivity analysis and ongoing work on including machine-learning-based parametrizations and coupling SpeedyWeather.jl to other Earth system component models.
Authors
Maximilian Gelbrecht (1,2), Milan Klöwer (3), and SpeedyWeather.jl contributors*
- Potsdam Institute for Climate Impact Research, Germany
- Technical University of Munich, Germany
- University of Oxford, UK
*SpeedyWeather.jl is a community project with valuable contributions from a large group of contributors, we are sorry that we can’t list everyone here by name
Researching differentiable programming and machine learning for Earth system models and dynamical systems
- The GPU acceleration of SpeedyWeather.jl, the friendly and flexible climate model
- Differentiable Climate Modeling: Calibrating SpeedyWeather with Enzyme
- A learned surface roughness scheme for climate prediction in SpeedyWeather.jl
- Terrarium.jl: Fully differentiable and GPU-accelerated land modeling at all scales in Julia
Milan Klöwer is a NERC Independent Research Fellow at the University of Oxford. He did his postdoc at the Massachusetts Institute of Technology (MIT) working on climate model development in Julia. He started SpeedyWeather.jl, a global atmospheric model designed as a research playground to develop prototype ideas on machine-learned representations of climate processes and computationally efficient climate models. He also works on low precision computing, data compression and information theory, predictability of weather and climate, and software engineering.