Extreme weather events are of growing concern for societies because under climate change their frequency and intensity are expected to increase significantly. Unfortunately, general circulation models (GCMs)–currently the primary tool for climate projections–cannot characterize weather extremes accurately. Here, we report on advances in the application of a multi-scale deep learning framework, trained on reanalysis data, to remedy deficiencies in GCMs to replicate the location, frequency, and intensity of tail events. The proposed approach 1) corrects the low-order and tail statistics of the GCM output at coarse scales; and 2) enhances the resolution of the debiased GCM output by reconstructing the finer scales as a function of the coarse scales. The novelty of our approach is to transform the GCM output without constraining the freedom of the climate model to sample the full distribution of possible extreme events consistent with near-present climate. This has significant implications for probabilistic risk assessment of natural disasters.

Zoom Link: https://ncsu.zoom.us/j/97638681103?pwd=dDJrRkE3d3NQZEhrRlhOMDc4T0pRUT09