Transition path theory analysis of noise-induced tipping in a stratospheric model

Nonlinear atmospheric dynamics produce rare events that are hard to predict and attribute due to many interacting degrees of freedom. Sudden stratospheric warming is an example where the winter polar vortex rapidly breaks down under wave disturbance, inducing midlatitude cold spells. Numerical simulations reproduce this phenomenon, but the complexity of the physics and data is a challenge for interpretation. We llustrate a description of rare weather events with Transition Path Theory, a mathematical framework that defines and relates extreme event statistics. Though the events that we focus on and their statistics involve very long timescales, we describe an approach to estimating them in high dimensional state spaces using only relatively short simulations of the system. Applying this methodology to a classical low-order stratospheric model with stochastic forcing, we compute optimal predictors, dominant pathways, and return times of noise-induced regime transitions, and relate them to physical observables. The study aims to motivate the broad potential meteorological utility of Transition Path Theory and of a computational framework to estimate TPT quantities for realistic models.