Evaluating large ensembles for persistent extreme events such as the 2020 temperature anomaly over Siberia

In 2020, Siberia experienced a persistent period of relative warmth, consisting of 11 consecutive months of positive temperature anomalies exceeding 1-3 standard deviations and resulting in an annual mean anomaly of 4°C relative to the 1951-1980 mean. Using dynamical adjustment, we dissect the event into contributions from atmospheric circulation anomalies, which dominated the warmth in winter and early spring, and its predominantly thermodynamic residual, which dominated in summer and fall. The residual temperature contribution is correlated significantly with antecedent factors such snow cover, soil moisture and sea ice. These factors were near or at record levels in 2020, thereby substantially increasing the survival probability of temperature anomalies from one month to the next. We evaluate several climate model large ensembles for their ability to produce similar events in terms of persistence, magnitude, and physical drivers. We also investigate how future warming changes the survival probabilities of temperature anomalies, conditional or unconditional on antecedent factors, and attempt to derive generalizable conclusions on the ability of climate models to simulate such persistent extreme events.