How reasonable is the Arctic/subArctic ocean in historical CMIP5 simulations?

The Arctic Ocean plays a key role in the heat budget of the highly sensitive polar climate through surface thermodynamic fluxes, surface radiative fluxes (through its impact on sea ice distribution), and advective exchanges with lower latitudes. Here we examine the mean, seasonal, and long-term trends in representation of these processes in historical runs of ensemble members from 14 CMIP5 climate models. We begin with an examination of surface meteorology, emphasizing those aspects that have the largest impact on ocean circulation and thermodynamic properties.

One key feature that impacts the ocean and differs significantly among the models is representation of the wintertime maximum in Beaufort sea level pressure (Polar High) and the corresponding Aleutian low. The several widely used models with weak or displaced Polar Highs have distorted ocean circulations and an inability to store ocean reserves of freshwater in the Beaufort gyre (an important aspect of the observed ocean). A second feature is the displacement of the storm tracks in the Atlantic sector and the frequency of blocking events in that sector. Model-to-model variations in these properties shift the relative importance of surface heat flux in the heat budget of the Arctic Ocean. Comparisons of the structure of blocking events in these models to observations will be presented. Averaged throughout the Arctic, net surface heating varies by more than a factor of two among models, mainly as a result of differences in sea ice distribution, with those models with the most intense seasonal net surface heat fluxes showing the most intense seasonal changes in ocean heat storage.

The second part of the talk focuses on representation of the Arctic and subarctic Ocean. The main source of warm water to the Arctic in the modern climate is the northward transport of some 8x10⁶ m³/s of Atlantic Water through the gap between Greenland and the UK. The heat and salt in this water becomes insulated from the sea ice and atmosphere by an overlying cool, fresh layer of water as it moves northward. Almost uniformly in these models the temperature and salinity characteristics of Atlantic Water are less extreme than observed. This systematic error further reduces the role of advective heat transport per unit volume transport and increases the rate of exchange of heat from the Atlantic Water to the sea ice and overlying atmosphere. For many of the 14 models we consider we have been able to compute an exact term-by-term seasonal heat budget (including transports through straits) and these will be presented in comparison to observations. As an extension of these budget calculations we discuss the ocean's thermodynamic contributions to the growing and shrinking of seasonal sea ice and, more broadly, the role that the high latitude ocean plays in the centennial trends in polar warming.