Arctic ocean–sea ice reanalysis for the period 2007–2016 using the adjoint method

We present an Arctic ocean–sea ice reanalysis covering the period 2007–2016 based on the adjoint approach of the Estimating the Circulation and Climate of the Ocean (ECCO) consortium. The spatiotemporal variation of Arctic sea surface temperature (SST), sea ice concentration (SIC), and sea ice thickness (SIT) is substantially improved after the assimilation of ocean and sea ice observations. By assimilating additional World Ocean Atlas 2018 (WOA18) hydrographic data, the freshwater content of the Canadian Basin becomes closer to the observations and translates into changes of the ocean circulation and of transports through the Fram and Davis straits. This new reanalysis compares well with previous filter‐based (TOPAZ4) and nudging‐based (PIOMAS) reanalyses regarding SIC and SST. Benefiting from using the adjoint of the sea ice model, our reanalysis is superior to the ECCOv4r4 product considering sea ice parameters. However, the mean state and variability of the freshwater content and the transport properties of our reanalysis remain different from TOPAZ4 and ECCOv4r4, likely because of a lack of hydrographic observations.

Arctic sea ice has declined rapidly and reached a record minimum in September, 2012. Arctic ocean–sea ice reanalyses are invaluable sources for understanding the Arctic sea ice changes. We produce an Arctic ocean–sea ice reanalysis of the years 2007–2016 using the adjoint method. The reanalysis is dynamically consistent without introducing unphysical mass and energy discontinuities as in filter‐based data assimilation methods.