SNAP is coordinating a new set of controlled numerical experiments, designed to isolate and quantify the contribution of the stratosphere to forecast skill on subseasonal time scales. These experiments target three recent stratospheric events: two major Northern Hemisphere sudden stratospheric warmings in February 2018 and January 2019, and the unusual near-major sudden warming in the Southern Hemisphere that occurred in September 2019. Each of these events was followed by a surface extreme thought to be connected to the stratospheric anomalies, though the timescale and intensity of the downward propagation differed among the events. 

The basic experimental protocol consists of a set of forecast ensembles: (1) a standard, free running forecast ensemble, (2) a ‘perfect stratosphere’ forecast in which the stratosphere is relaxed towards the observed evolution, and (3) a ‘control’ forecast in which the stratosphere is relaxed towards climatology. Further details of the experimental protocol will be described in an article soon to be submitted to a peer-reviewed journal. To date, twelve modeling groups at eleven centers are planning to contribute integrations following this protocol. This will allow for an unprecedented, multi-model comparison of the dynamics underlying the surface responses to sudden stratospheric warmings. Moreover, by including ‘counterfactual’ forecasts in which the stratospheric circulation remains in a climatological state, the experimental protocol will allow for formal attribution statements to be made regarding the surface extremes that followed the stratospheric anomalies.

The goal is to have the experiments completed by summer of 2021, and the initial analysis will be carried out by a set of community working groups.

Anyone interested in participating in the community analysis of these experiments is encouraged to contact Peter Hitchchock (), Amy Butler (), and Chaim Garfinkel () for  further information. We expect initial results to be reported towards the end of 2021 through the first half of 2022.   After an initial embargo period, the dataset will be made available to the broader community in the fall of 2022.