Meeting Documents

Evaluating Wind-Ice Coupling in ECCO

Abstract

Sea ice modulates the atmosphere-ocean energy exchange. Due to anthropogenic activities, the Arctic is warming, leading to reduction trends in sea ice extent, thickness, and the fraction of multiyear ice. In turn, the average sea ice drift speeds are increasing. Although observations in the Arctic are sparse, numerical ocean models have proven to be a powerful tool for understanding multiple aspects of the Arctic wind-ice-ocean coupled climate system. In this study, we evaluated the seasonal and regional variability of the wind-ice coupling in version 4 of Estimating the Circulation and the Climate of the Ocean (ECCO v4r4). ECCO v4r4 is a dynamically consistent ocean state estimate synthesizing a vast number of oceanographic observations using the MITgcm ocean model. While ECCO v4r4 has been frequently used and assessed for the midlatitude oceans, we present the first assessment of the large-scale Arctic Sea ice circulation in the model. Since wind stress is a key driver of the Arctic Ocean circulation, an accurate representation of wind-ice coupling in the model is critical. We evaluated the ECCO v4r4 sea ice circulation relative to the National Snow and Ice Data Center (NSDIC) Polar Pathfinder 25 km EASE-grid ice motion product and the ERA-Interim wind field between 1992-2017. Overall, broad features of the Arctic Sea ice circulation (e.g., Beaufort Gyre, Transpolar Drift) are well represented in ECCO v4r4. However, we found that ECCO drift speeds are consistently higher than observed values, with summer drift speeds nearly twice as high as in the NSIDC. The strength of the correlation varies by region and is highest in the central Arctic. In the marginal seas, the correlation is highest in winter and lowest in summer. We investigated the underlying physics of this difference by examining the relationship between the monthly ice drift speed and the monthly wind field. Compared to the observations, the correlation between wind speeds and ice drift in ECCO v4r4 is weaker in summer when the ice pack is not in contact with the coasts. This correlation increases during the winter, suggesting the modeled sea ice pack is more responsive relative to the wind than it is in the observations. Going forward, our goal is to understand how a model bias in drift speed is likely to affect the upper ocean circulation and the air-ice momentum transfer.