The project will investigate how better forecasts of extreme weather events, such as tropical cyclones (TCs), can be achieved via unprecedented simulations which represent the Earth's atmosphere and oceans with very fine detail. The outcome of this project will contribute to solving the top global threats in the next decade—extreme weather and climate action failure.

Performing the world’s first-ever GPU-enabled global subseasonal nonhydrostatic simulations of the atmosphere at 1 km resolution coupled to an eddy-permitting ocean, this work seeks to answer fundamental questions: How does the meso-scale permitting ocean, when coupled to a convection permitting atmosphere, impact TC and Madden Julian Oscillation (MJO) representation? What is the impact of relaxing the hydrostatic approximation on the simulated convection, and thus on the TC and the MJO representation? What advantages, if any, does the 1 km resolution present over the more affordable 3-4 km resolution for representing the Earth’s weather and climate?

Further, the project will address whether explicitly resolving deep convection while simultaneously coupling to the ocean significantly improves forecast skill of the MJO and TCs in comparison to current models, as well as remote influence of the MJO on extremes and teleconnections between the MJO and other atmospheric features.