Evaluation of sea-ice thickness reanalysis data from the coupled ocean-sea-ice data assimilation system TOPAZ4

Yongwu Xiu,Jiping Xie,Longjiang Mu,Bo Han,Chao Min,Qinghua Yang

doi:10.1017/jog.2020.110

Abstract

AbstractWith the assimilation of satellite-based sea-ice thickness (SIT) data, the new SIT reanalysis from the Towards an Operational Prediction system for the North Atlantic European coastal Zones (TOPAZ4) was released from 2014 to 2018. Apart from assimilating sea-ice concentration and oceanic variables, TOPAZ4 further assimilates CS2SMOS SIT. In this study, the 5-year reanalysis is compared with CS2SMOS, the Pan-Arctic Ice-Ocean Modeling and Assimilating System (PIOMAS) and the Combined Model and Satellite Thickness (CMST). Moreover, we evaluate TOPAZ4 SIT with field observations from upward-looking sonar (ULS), ice mass-balance buoys, Operation IceBridge Quicklook and Sea State Ship-borne Observations. The results indicate TOPAZ4 well reproduces the spatial characteristics of the Arctic SIT distributions, with large differences with CS2SMOS/PIOMAS/CMST mainly restricted to the Atlantic Sector and to the month of September. TOPAZ4 shows thinner ice in March and April, especially to the north of the Canadian Arctic Archipelago with a mean bias of −0.30 m when compared to IceBridge. Besides, TOPAZ4 simulates thicker ice in the Beaufort Sea when compared to ULS, with a mean bias of 0.11 m all year round. The benefit from assimilating SIT data in TOPAZ4 is reflected in a 34% improvement in root mean square deviation.

Highlights

Arctic sea ice is an important component of the global climate system (Serreze and Meier, 2019) and plays a crucial role in the global ocean circulation (Sévellec and others, 2017)
A 5-year TOPAZ4 sea-ice thickness (SIT) reanalysis product has been generated with the assimilation of sea-ice concentration (SIC) from Ocean and Sea Ice Satellite Application Facility (OSISAF) and SIT from CS2SMOS
Despite assimilating SIT reduced the SIT errors by ∼28% from 19 March 2014 to 31 March 2015 (Xie and others, 2018), quantified statistics remains to be made to figure out how TOPAZ4 perform based on more observations and over longer period

Summary

Introduction

Arctic sea ice is an important component of the global climate system (Serreze and Meier, 2019) and plays a crucial role in the global ocean circulation (Sévellec and others, 2017). Among all the sea-ice properties, SIT is important, because changes in its distribution can affect the ocean-atmosphere heat fluxes especially during the fall period (Kurtz and others, 2011). With respect to the Ice Thickness Regression Procedure (Lindsay and Schweiger, 2015) data in February–March, most of these products underestimate SIT to the north of the Canadian Arctic Archipelago (NOCAA) and the north of Greenland and the Fram Strait, but overestimate SIT over the Eurasian shelves. This exposes the lack of direct constraints on SIT, which is a common deficiency of all reanalysis products

Methods

Results

Conclusion