Assessment of high latitude variability and extreme events in the Bering Sea as simulated by a global climate model

Atmospheric and Oceanic observations of the Arctic and Subarctic are relatively sparse and hinder our ability to analyze short term variability and long-duration anomalies of physical and biological variables over decadal time scales. Earth System Models (ESM's), such as the Community Earth System Model (CESM1), represent a useful tool to advance the understanding and the predictive potential of large-scale shifts in the climate and climate related impacts.;This thesis initially focuses on assessing the skill of the Community Climate System Model (CCSM4), to capture natural variability of the climate system. Subsequently, I examine the impacts of variability and seasonal-scale extremes of the physical environment on the marine ecosystem of the eastern Bering Sea as simulated by an earth system model, the CESM1, which includes the CCSM4 and earth system elements. A performance assessment of key atmospheric components (air temperature, sea level pressure, wind speed and direction) simulated by the CCSM4 over the Bering Sea and Arctic domains suggests a general improvement in model predictions at high latitudes relative to the model's predecessor, the CCSM3. However, several shortcomings, with possible implications for marine ecosystem modeling, still remain in this version of the CCSM. The most important of which includes an under-simulated Siberian High and a large northwest displacement of the Aleutian Low resulting in a negative bias of up to 8 hPa over the Bering Sea. The simulated inter-annual variability of surface air temperature and sea level pressure over the Bering Sea was found to exceed observed variability by ∼1.5 to 2 times. The displaced pressure systems and increased variability could have important ramifications for modeling efforts that use CCSM atmospheric output as drivers for marine ecosystem studies.;When the CCSM was combined with other earth system elements to form the CESM, the coupled model was found to simulate strong linear relationships between primary production and air temperature, and between primary production and sea ice area over the fifty-five year period examined (1950-2005). A trend towards warmer air temperatures and reduction in sea ice area was found in every season. With a simulated increase in air temperature, an increase in the occurrence of seasonal positive primary production extremes followed. There were several instances of extremes in the physical environment coinciding with primary production extremes. However, clear, discernable patterns relating seasonal extremes in the physical environment to extremes in the production were hard to come by, suggesting that the complex interaction between the biology and physics was not fully captured by the variables examined. However, it is perhaps more likely that the lack of correspondence was because the important interactions between the biology and physics of the eastern Bering Sea occur on sub-seasonal timescales.