Data-Driven Probabilistic Air-Sea Flux Parameterization

2025-03-06

AI Summary

Deterministic air–sea flux parameterizations fail to capture the strong spatiotemporal variability of turbulent fluxes, limiting coupled models’ representation of air–sea interaction. To address this, we propose the first probabilistic flux parameterization framework that integrates Gaussian conditional distributions with neural networks, enabling joint estimation of flux means and predictive uncertainties while supporting stochastic sampling for ensemble generation. Trained on eddy-covariance observations and evaluated within a single-column upper-ocean model, our framework significantly improves seasonal simulations of sea surface temperature and mixed-layer depth. Notably, ensemble spread peaks during spring restratification—a physically meaningful signature—validating the efficacy of uncertainty quantification. This work establishes a new paradigm for data-driven, uncertainty-aware air–sea flux parameterization.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurately quantifying air-sea fluxes is important for understanding air-sea interactions and improving coupled weather and climate systems. This study introduces a probabilistic framework to represent the highly variable nature of air-sea fluxes, which is missing in deterministic bulk algorithms. Assuming Gaussian distributions conditioned on the input variables, we use artificial neural networks and eddy-covariance measurement data to estimate the mean and variance by minimizing negative log-likelihood loss. The trained neural networks provide alternative mean flux estimates to existing bulk algorithms, and quantify the uncertainty around the mean estimates. Stochastic parameterization of air-sea turbulent fluxes can be constructed by sampling from the predicted distributions. Tests in a single-column forced upper-ocean model suggest that changes in flux algorithms influence sea surface temperature and mixed layer depth seasonally. The ensemble spread in stochastic runs is most pronounced during spring restratification.

Problem

• Develop probabilistic framework for air-sea flux variability
• Estimate mean and variance using neural networks
• Assess impact of flux algorithms on ocean models

Innovation

• Probabilistic framework for air-sea flux variability
• Neural networks estimate mean and variance
• Stochastic parameterization from predicted distributions