Atmosphere Ocean Science Friday Seminar

---

Data-driven probabilistic air-sea flux model using in-situ direct measurements

Speaker: Jiarong Wu, CAOS

Location: Warren Weaver Hall 1314

Date: Friday, May 10, 2024, 4 p.m.

Synopsis:

The exchange at the air-sea interface is important for many weather and climate systems. Still, large uncertainty remains associated with the widely used bulk formulae of air-sea fluxes. The surface fluxes (momentum, latent and sensible heat) are predicted given state variables (wind speed, air temperature, ocean temperature, humidity, etc.). In bulk formulae, simple parametric forms are assumed, and the coefficients are empirically determined from limited in-situ observations. We propose to rethink bulk formulae using machine learning tools. I will discuss our recent work of building a probabilistic model (as opposed to a deterministic one) of air-sea fluxes based on in-situ flux measurements, and the challenge that comes with using small and noisy observational data sets.