On the role of Eurasian autumn snow cover in dynamical seasonal predictions

Seasonal predictions leverage on predictable or persistent components of the Earth system that can modify the state of the atmosphere. The land surface provides predictability through various mechanisms, including snow cover, with particular reference to Autumn snow cover over the Eurasian continent. The snow cover alters the energy exchange between surface and atmosphere and induces a diabatic cooling that in turn can affect the atmosphere locally and remotely. Lagged relationships between snow cover in Eurasia and atmospheric modes of variability in the Northern Hemisphere have been documented but are deemed to be non-stationary and climate models typically do not reproduce observed relationships with consensus. The role of the snow in recent dynamical seasonal forecasts is therefore unclear. Here we assess the role of Autumn Eurasian snow cover in a set of five operational seasonal forecasts with large ensemble size and high resolution and with the help of targeted idealised simulations. Forecast systems reproduce realistically regional changes of the surface energy balance. Retrospective forecasts and idealised sensitivity experiments identify a coherent change of the circulation in the Northern Hemisphere. The main features of the atmospheric response are a wave-train downstream over the Pacific and North America and a signal in the Arctic. The latter does not emerge in reanalysis data but is compatible with a lagged but weak and fast feedback from the snow to the Arctic Oscillation.