Atmosphere Ocean Science Colloquium

---

At the Interface of Atmosphere and Ocean: Physics and Parameterizations

Speaker: Jiarong Wu, NYU

Location: Warren Weaver Hall 1302

Date: Wednesday, November 5, 2025, 3:30 p.m.

Synopsis:

Accurately quantifying air–sea fluxes is important for understanding air-sea interactions and improving coupled weather and climate models. Many microphysical processes at the air–sea interface remain unresolved at current model resolutions and sophistication, and must therefore be parameterized. In this talk, I will first give an overview of these processes and ongoing efforts to represent them in air-sea flux parameterization and surface layer modeling. I will then present an alternative approach that frames air–sea flux parameterization as a regression problem with explicit uncertainty quantification given sparse direct observations. This probabilistic framework captures the highly variable nature of air-sea fluxes that is absent in traditional deterministic bulk algorithms. Neural networks (NNs) are employed to estimate conditional distributions, and the trained NNs provide a reasonable estimate of fluxes (with noticeably improved skill for latent heat flux) together with their uncertainty. Stochastic parameterizations can then be constructed by sampling from the predicted distributions. Tests using a single-column forced upper-ocean model suggest that changes in flux formulations can directly influence sea surface temperature and mixed-layer depth seasonally. I will conclude with a discussion of the broader implications of air–sea flux parameterizations for coupled general circulation models.