Parameterization

NORi combines a physics-based ocean boundary-layer mixing scheme with neural networks that learn the entrainment missing from a local diffusive closure. It trains against the evolution of temperature and salinity in large-eddy simulations while conserving tracers by construction. Tests examine unfamiliar forcing conditions and seasonal mixing at Ocean Weather Station Papa, where performance is comparable to established closures. An idealized double-gyre experiment remains numerically stable for 100 years despite training on two-day trajectories, although this does not establish accuracy in a realistic global ocean. Readers get a detailed example of designing, training, and evaluating a hybrid physics-and-ML parameterization that targets missing processes while retaining a simple physical foundation.

This paper introduces CATKE, a parameterization for small-scale ocean mixing that predicts turbulent kinetic energy and diagnoses a mixing length. Its convective-adjustment component represents both the depth reached by convective plumes and the time required for them to mix the water column. This allows the model to represent competition between convection, shear-driven mixing, and processes that restore stratification. Ensemble Kalman inversion calibrates the scheme against large-eddy simulations at several vertical resolutions, followed by tests on idealized and realistic cases. Readers get the closure formulation and calibration strategy for a relatively economical mixing model, together with evidence of its performance against established alternatives.

This paper studies passive-tracer mixing in an idealized, convecting ocean mixed layer. A two-dimensional direct numerical simulation is used to compute an eddy-diffusivity kernel relating the mean vertical flux at one depth to tracer gradients at other depths. The kernel captures local and nonlocal contributions within a single transport representation. The experiment shows how multiple passive tracers can diagnose that representation directly from the simulated flow. Readers get a concrete method for measuring nonlocal mixing and a physical explanation of why a diffusivity defined only at the local depth can be insufficient.