AI can accelerate climate modeling by generating computationally cheap emulators of existing simulations. The design of climate emulators is an exciting and open challenge in AI+Climate. In this work , PhD student Adam Subel and Laure Zanna focus on two fundamental questions to facilitate the creation of ocean emulators: 1) the role of the atmosphere in improving the decadal skill of the emulator and 2) the representation of variables with distinct timescales (e.g., velocity and temperature) in the design of any emulator. This work was accepted in the ICLR 2024 Workshop: Tackling Climate Change with Machine Learning , and selected as a Spotlight Talk! Make sure to catch Adam’s talk in Vienna if you are there, or on our YourTube Channel in May.