AI-boosted rare event sampling to characterize extreme weather
read the original abstract
Weather extremes pose major societal risks, especially in a changing climate, but due to their rarity, they are difficult to study using limited observations or complex climate models. We introduce AI+RES, a framework coupling fast AI weather forecasts with a high-fidelity physics model using a rare-event algorithm to efficiently characterize extremes. This approach enables the study of the statistics and physics of very rare events, such as once per millennium heatwaves at two orders-of-magnitude lower computational cost. AI+RES can be applied broadly across climate science and other fields concerned with rare events.
This paper has not been read by Pith yet.
Forward citations
Cited by 2 Pith papers
-
Conditional Tropical Cyclogenesis Rates via Rare-Event Sampling in a Neural Weather Emulator
Forward Flux Sampling applied to a 1-degree neural weather emulator resolves conditional tropical cyclogenesis rates spanning three orders of magnitude across 98 Atlantic initial conditions, with self-consistency rati...
-
Routes to rare events with optimally timed perturbations: a Tent Map is all you need
The optimal advance split time for rare event sampling in 1D chaotic maps equals the logarithm of the ratio of target rarity to initial perturbation size, generalized via maximum entropy.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.