Uncertainties in ocean mixing parameterizations are primary sources for ocean and climate modeling biases. Due to lack of process understanding, traditional physics-driven parameterizations perform poorly in the tropics. Recent advances in deep learning method and new availability of long-term turbulence measurements provide an opportunity to explore data-driven approaches to parameterize oceanic vertical mixing processes. Here, we describe a novel parameterization based on an artificial neural network trained using a decadal-long time record of hydrographic and turbulence observations in the tropical Pacific. This data-driven parameterization achieves higher accuracy than current parameterizations, demonstrating good generalization ability under physical constraints. When integrated into an ocean model, our parameterization facilitates improved simulations in both ocean-only and coupled modeling. As a novel application of machine learning to the geophysical fluid that employs turbulence measurements, these results show the feasibility of using limited observations and well-understood physical constraints to construct a physics-informed deep learning parameterization for improved climate simulations.

Compound wind and precipitation extremes (CWPEs) can severely impact multiple sectors and regions, often causing critical infrastructure failure and fatalities, especially in the Indo‐Pacific region, which is a hotspot for CWPEs of various synoptic origins. Results show that the northwestern Pacific Ocean and its coasts have experienced the most frequent, strongest, and longest‐lasting CWPEs in summer in recent decades, which are induced by cyclones. Landfalling atmospheric rivers are one of the main drivers for frequent occurrences of CWPEs in central and western China and the northwestern Indo‐China Peninsula in both boreal summer and winter. The frequency of CWPEs over southern China exhibits significant decreasing trends in contrast to increasing trends in equatorial tropical areas. Moreover, the magnitude and pattern for the observed changes in the frequency of CWPEs result primarily from the variations in the dependence between univariate extremes, when evaluating the Indo‐Pacific as a whole.