Advancements in satellite data assimilation in TOPAZ coupled ocean and sea ice model system

Assimilation of sea ice observations has been limited by the difficulty to define adequate forecast errors, key challenges include the non-Gaussianity of sea ice variables and a chronic underdispersion of forecast ensembles. In this presentation, we show two advancements in the TOPAZ coupled ocean and sea ice model system, where a 100-member ensemble is utilized to provide flow-dependent estimates of uncertainties. Firstly, the representation of atmospheric uncertainties is improved by an updated perturbation scheme. A comparison between the operational TOPAZ perturbation scheme with the ECMWF ensemble forecast revealed the deficiencies in sea ice drift spread at large scales. The new scheme takes a multiscale approach to compensate for this deficiency. Secondly, a radiative transfer model was implemented as a new observation operator in TOPAZ, which enabled the direct assimilation of brightness temperature measurements from the Advanced Microwave Scanning Radiometer 2 (AMSR2) satellite mission. Compared to the conventional assimilation of sea ice concentration products retrieved from AMSR2, the direct assimilation approach better accounts for the uncertainties from the sea ice and the atmospheric forcing, avoiding the biases due to assumptions made in the traditional retrieval process. We will also discuss the future direction to integrate these advancements in the TOPAZ system, in preparation for the upcoming Copernicus Imaging Microwave Radiometer (CIMR) mission. This work is funded by the European Horizon project ACCIBERG.