WP2 Sea Ice Forecasts
Work in the WP has mostly progressed as planned in the RP. The transfer of regional ice-ocean ensemble forecasts was slightly delayed but we avoided detrimental consequences for the downstream WP3. The software and implementation of prototypes for sea ice forecast products on the EDITO cloud computing platform have been delivered on time (Tasks 2.1 and 2.2) and good progress has been made with developing more reliable sea ice forecasting systems (Tasks 2.3 and 2.4).
WP3: Data Assimilation
The assimilation of raw (Level-1b) satellite EO data for improving the forecasts of sea-ice concentration (SIC) and sea-ice type is able to successfully correct the background model field of SIC. We have performed standalone analysis in a period of six months in 2021, which consistently presents positive results when assimilating AMSR2 brightness temperature (Tb) data. This work has led to the submission on time of a draft report on the direct assimilation of Tb in ocean/ice forecasting applications. Currently the assimilation of Tbs does not perform better than the assimilation of SIC, providing similar results in both assimilation setups. This might be due to the current setup of the Radiative Transfer Model (RTM) which provides simulated Tb that are directly linear to model SIC. In this context, we are testing a stochastic version of the RTM which should set the assimilation system in a non-Gaussian regime. This work is ongoing. The RTM has successfully been interfaced into the NEDAS data assimilation (DA) system, with initial DA tests already realized and reported in the "Open-source contribution to the EnKF software, including the satellite simulator". This is a first step towards the Arctic pre-operational forecast runs together with the NEDAS DA system, which will lead to a final report on the direct assimilation of Tb (planned for September 2026). This work will provide insights on the impact of assimilating Tb in the forecast ice fields. An independent task has been the computation of ice emissivities using machine learning. These emissivities did not provide better results in the assimilation of Tbs for updating SIC, but might help in updating other sea ice variables like sea-ice type or snow-depth. This will be investigated in the last year of the project.
WP4: Iceberg Forecasts
WP4 has attained mixed progress: the remote sensing product from DMI was successfully integrated in Copernicus Marine Services, the iceberg forecast-on-demand is now available as a software on two different web-based platforms (Værio for operational users and EDITO for the public at large). However the statistical risk forecast was not started, and was redefined in dialogue with the users as a GIS-compatible extension of the forecast-on-demand. Demonstrations with ships-of-opportunity have been initiated with low-cost iceberg tags while the more performant buoys were delayed in a public procurement procedure. A successful near-operational demonstration was given with an iceberg tagged “Juliette” which drifted in an area previously void of observations. The demonstration on “Juliette” helps uncover bugs in the OpenBerg software package.
WP5: Dissemination and Communication activities
The communication and dissemination activities in the project have ramped up over the past reporting period, and will continue to do so over the last year of the project. There has been good engagement with the key stakeholders, with the Ice Services beginning to test and use the ACCIBERG products. A large number of dissemination activities have taken place, to both academic and end user audiences, and the number of communication activities has increased.
Comments from the last reporting period have been taken on board - a communication group has been set up, the website has been updated, video produced. Project outputs are available on the website (not a data portal per se, as the data generated by the project - iceberg trajectories - has only just been produced). There’s been a good range of dissemination activities throughout the reporting period to scientific and end users. Attended EU Booster session but felt that it would be most beneficial for a wider/more general audience, which would be useful later in the project.
