Skilful predictions of near-term climate extremes are key to a resilient society. However, standard methods of analysing seasonal forecasts are not optimised to identify the rarer and most impactful extremes. The 2022 Pakistan floods, an extreme event, were widely predicted by seasonal forecasts, yet standard tercile probability maps failed to convey the magnitude. This paper argues that a strong summer La Niña created a window of opportunity for a much more confident extreme rainfall forecast than average skill estimates suggest. The authors explore methods for building forecast confidence, including physical understanding of dynamical mechanisms, perturbation experiments, and empirical relationships. They emphasize the need for improved tools and routine monitoring of forecasts to identify regional climate extremes and utilize these windows of opportunity for trustworthy early warnings.

Nick Dunstone, Doug M. Smith, Steven C. Hardiman, Paul Davies, Sarah Ineson, Shipra Jain, Chris Kent, Gill Martin, Adam A. Scaife