🌀 Hurricane help from #NVIDIAResearch: StormCast is a new #generativeAI model for emulating high-fidelity atmospheric dynamics. This model can enable reliable weather prediction at mesoscale which is critical for disaster planning and mitigation. ➡️ ⛈️

Zeyuan Hu (Harvard University) and his colleagues have been working on developing machine-learning (ML) subgrid parameterizations for convection and radiation using the ClimSim dataset. By integrating microphysics constraints into the ML emulator, their work achieves stable and skillful multi-year hybrid climate simulations with near operational-level complexity including geography, seasonality, and explicit cloud condensate coupling. If this sounds interesting then please consider our AGU session where this and more like it will be being presented as talks: and as posters: Please get in contact if there’s anything I can help you with. Look forward to seeing you there! :) Conveners: Simon Driscoll (Reading - primary convener), Sara Shamekh (NYU/Courant), Aakash Sane (Princeton - @aakashsane), Laura Mansfield (Stanford/now Oxford - @lau_mansfield), Michael Bosilovich (NASA), Arun Kumar (NOAA/NCEP); ECR: Will Gregory (Princeton - @willjgregory_); Chairs on the day: Simon Driscoll, Michael Bosilovich, Will Gregory, Aakash Sane. #ml #ai #reanalyses #climate #environment #AGU24

🔍 Superresolution misalignment? Conditional diffusion/flow models falter when low-res and high-res data don’t align. Meet Stochastic Flow Matching—a framework mapping low-res to latent space, bridging high-res targets effectively. 📝 Paper:

📢🔥 Check out our new work on modeling heavy-tailed data (e.g., extreme weather events) using diffusion models.

Excited to share our new review paper on the use of ML for climate science led by Veronika Eyring and Bill Collins and based on a great workshop in Colorado in 2022