Modeling and Simulation Group Meeting Old

 

Abstract: In the Earth and planetary sciences, accurately estimating rare event probabilities is challenging yet scientifically significant. To estimate these rare probabilities, direct numerical simulations is often a limited tool with an unacceptably low benefit-cost ratio. As an alternative, we want to develop a non-intrusive Monte Carlo method that can be applied to a black box simulation model to improve rare probability estimates. Yet despite mathematical evidence that such a method can exist, developing this method for real-world applications is an ongoing challenge. In this talk, I introduce this open problem in applied mathematics and describe three examples involving heat waves, hurricanes, and Mercury's potential to be ejected from the solar system.