Assessing Regional Sea Ice Predictability in the US Arctic: A multi-Model Approach

Enrique Curchitser -  Rutgers, The State University of New Jersey

 Abtract

Predictability of pan-Arctic sea ice extent and volume has been examined in a number of studies. In comparison, regional sea ice predictability is less explored despite its significance for a suite of stakeholders — coastal communities, shipping and oil industry, and marine ecosystem management. The goal of this study is to assess regional sea ice predictability in the US Arctic: the Beaufort Sea, Chukchi Sea, and Bering Sea.This effort is in part motivated by PMEL/EcoFOCI (Ecosystem and Fishery Oceanography Coordinated Investigations) group’s in situ knowledge of the role of sea ice on the regional oceanography and marine ecosystem. To assess predictability, we will utilize existing output from the NCAR CESM Large Ensemble simulations and the NOAA CFSv2 forecast system. In addition, we will carry out ocean-sea ice experiments using a pan-Arctic grid of the Regional Ocean Modeling System (ROMS) at nominal 10-km horizontal resolution, driven by these global models. We will analyze both diagnostic (lagged correlation) and prognostic (from initialization ensemble runs) sea ice predictability on regional scales. For the latter, root-mean-square error, anomaly correlation coefficient, and potential prognostic predictability will be calculated. We will begin our analyses on the regional sea ice area and thickness and, time permitting, expand to other properties such as ice edge locations and spring retreat timing. Moreover, we will examine cross-model differences in regional predictability and assess what factors (e.g., spatial resolution and coupling) and physical mechanisms contribute to such differences. Results from this study are anticipated to improve our understanding of processes controlling regional sea ice predictability in the US Arctic, a necessary step towards developing cross-scale earth system prediction capability. This effort will be in close coordination with scientists working on the ESPC Arctic sea ice focus project, directly contributing to the goals of the latter.

As part of this effort, Enrique Curchitser at Rutgers University will be responsible for a high-resolution regional implementation of the coupled sea-ice/circulation model in the Arctic.  The model is based on the Regional Ocean Modeling System (ROMS) framework, enhanced with a dynamic and thermodynamic sea ice model.  Together with colleagues at NOAA-PMEL, Curchitser will work on carrying out the coupled model simulations and the interpretation of results as well as preparation of manuscripts resulting from the proposed work