Arctic Extended-Range Predictions with MPAS and MPAS-CESM

Arctic sea ice exhibits considerable year to year variability, with occasional abrupt reductions in sea ice exhibiting limited prediction capability by numer- ical models. Variability of sea ice has potential to become amplified as sea ice thins. Extremes of summer Arctic sea ice loss are intimately connected to the atmospheric forcing from seasonal circulations to individual cyclones. Hy- pothesized to be key components of this atmospheric forcing, tropopause polar vortices (TPVs) are common, coherent upper level potential vorticity anomalies with typical radii of 100 to 1000 km and lifetimes of days to months. Limited prediction of year to year sea ice extremes and systematic errors in modelled TPVs motivate exploring the sensitivities of their evolution to model design, towards extending understanding and predictions.

The non-hydrostatic atmospheric core of the Model for Prediction Across Scales (MPAS) falls into a new class of models bridging gaps between coarse general circulation models and higher resolution limited area models. Further- more, MPAS has been implemented as a dynamical core within the Community Atmosphere Model (CAM) of the Community Earth System Model (CESM). Leveraging the flexibility of the fully-coupled (atmosphere, land, ocean, and sea ice) MPAS-CESM, and the local refinement capability of the MPAS atmospheric component, we discuss extended-range simulations of the summers of 2006 and 2007, which are two years characterized by strongly contrasting seasonal circula- tions, TPV trajectories, and sea ice extents. An ensemble of simulations samples varying regions of refinement, physics parameterizations, and component cou- pling. TPV ensemble sensitivities demonstrate the significance of representing local Arctic processes. Implications for applying MPAS-CESM as a regional prediction model over the Arctic are discussed.