FY 2012

Climate model projections for the Arctic

Overland, J.E., M. Wang, J.E. Walsh, J.H. Christensen, V.M. Kattsov, and W.L. Chapman

Chapter 3 in Snow, Water, Ice and Permafrost in the Arctic (SWIPA): Climate Change and the Cryosphere, Arctic Monitoring and Assessment Programme (AMAP), Oslo, Norway, 3.1- 3.18, Full report published at AMAP (2012)

Projections of future changes in the cryosphere are dependent on interpretation of results from multiple climate model forecasts at regional scales. In setting the stage for discussing the various physical and biological components of the SWIPA project, this chapter surveys the major causes of the range of results provided by multiple atmosphere-ice-ocean general circulation models and recommends a methodology for summarizing the available projections in practical applications such as this assessment. Strategies are assessed for reducing the uncertainties from multiple model projections while acknowledging that multiple model simulation results (termed ‘ensemble members’) must be retained because forecasts should include a range of results caused by natural variability present in the real world and represented in the climate models. Results from international modeling centers form an ‘ensemble of opportunity’ from which composite projections can be made. However, experience suggests that these models need to be subjected to a selection methodology and independent interpretation.

Interpretation of results from 18 international modeling centers developed for the Fourth Assessment of the Intergovernmental Panel on Climate Change, suggests an increase in Arctic-wide autumn and winter surface air temperatures of 3 to 6 °C by 2080, a nearly sea-ice free September by 2050, and a general increase in precipitation. Loss of sea ice and increased temperatures are a coupled process with the greatest temperature increases seen in the autumn. The Arctic shows the largest future temperature changes on the planet.

Considerations that guide in the initial screening of a set of models are model resolution, comprehensiveness of processes included in the model, and sophistication of the parameterizations. A main procedure for selecting a subset of the models is based on comparison to observations during the model hindcast period. The quality of individual model performance varies for different regions, variables, and evaluation metrics. For this reason, the use of a single model forecast is not recommended. The coarse resolution of most current climate models dictates caution in their application on smaller scales in heterogeneous regions such as along coastlines or in locations with rugged orography.