The record of sea ice extent and concentration for the Arctic reveals significant interannual variability superimposed on a net trend of decreasing ice cover. This interannual variability as well as the longer-term trends result in large part from relatively complex interactions among atmospheric and oceanic dynamics and thermodynamics that regulate the ice cover. Climatologies are useful for assessing the basic performance of models, but evaluating the ability to simulate these detailed processes requires new types of data sets that provide consistent, multi-parameter information at improved temporal and spatial resolution.
Here, we make use of newly-available products generated by the NOAA/NASA Pathfinder project in combination with other data sets to document the performance of stand-alone ice models and a regional climate model of the Arctic (ARCSyM). The basic characteristics of the AVHRR-Based Polar Pathfinder data sets are reviewed, including a discussion of applications for model assessments as well as for observing regional and interannual variability in ice transport, cloud cover, surface temperature, and albedo. Error sources and potential limitations of the Pathfinder data are also reviewed. Particular climate situations that are considered include the mechanisms involved in record reductions in ice extent in 1990 and the conditions leading up to the anomalous ice conditions in the region of the Surface Heat Budget of the Arctic (SHEBA) experiment in 1997. Specific questions addressed include the ability of the models to reproduce regional variations in ice transport and net radiation that likely affect the unusual sea ice conditions. In the 1990 case for example, the record ice-extent reductions observed at the end of summer appear linked to unusual ice transport in late spring followed by above-normal net radiation during summer. Intercomparisons of observations, stand-alone modeling and coupled modeling output are used to assess the relative importance of these individual forcing mechanisms and their interactions. The results highlight some of the critical issues involved in achieving accurate simulations of trends in sea ice cover using regional and global climate models.