Scientists use supercomputer to understand 'extreme' turbulent events

US researchers at Georgia Tech have performed the largest-ever 3D direct numerical simulation of turbulence using the Frontier supercomputer, mapping fluid chaos across 35 trillion grid points. This unprecedented computational effort allowed them to reach a Reynolds number of 2,500, achieving a level of physical fidelity comparable to real-world laboratory experiments. The study decodes the core mechanics of turbulent fluctuations, offering new insights into phenomena ranging from global ocean circulation to jet engine environments. These findings have practical implications for improving weather forecasts and designing more efficient vehicles and airfoils. A key focus of the research is on “extreme events”—rare, intense fluctuations that classical turbulence theories often fail to predict but are responsible for high-impact occurrences like Category 5 tornadoes and engine auto-extinction. By shifting from deterministic predictions to stochastic statistical models, the team better characterizes the probability distributions of these events, enhancing preparedness for their real-world effects. The study also distinguishes between energy dissipation and enstrophy (localized swirling intensity),