Consistent Air-Ice-Sea Data Assimilation of Satellite Observations (CAISA)

Research Project, 2022 – 2024

Processes close to the Earth's surface play a central role for both
oceanographic and meteorological forecasting. Most socio-economic processes that are affected by weather and the sea are situated close to the Earth's surface. Moreover, the exchange of heat, water and water vapour between the surface (water, ice, snow) and the atmosphere are important drivers for the meteorological as well as the oceanographic state. To make good forecasts, it is not sufficient to have good forecast models, it is also crucial to have as good initial conditions as possible. Due to the superior coverage of satellites compared to in-situ observations, it is very important to be able to make use of existing and future satellite data.
Regarding today’s ice-ocean forecasts in the Baltic Sea, no satellite-derived ice products are currently assimilated into the forecasting models, due to limited quality and the lack of data in coastal regions. The ice-ocean surface is important also for weather forecasts as it serves as the lower boundary for atmospheric models. Further, low-peaking channels are today not used in weather forecasts in the Baltic region.
To reap the full benefit of the available satellite data, we propose to develop methods that will make it possible to assimilate much more satellite observations into the forecasting models than has been possible before. This will be accomplished by developing methods to directly assimilate raw satellite data that are sensitive to near-surface atmospheric fields as well as ice-ocean fields. Instead of relying on satellite retrievals of ice and snow, which are not well developed for the Baltic Sea region, we will instead use different Radiative Transfer Models for the open ocean and ice and snow
surfaces to calculate model equivalents of the satellite radiances. This will enable us to assimilate satellite radiances directly into the forecast models, leading to a more optimal use of satellite data and increased consistency between the models, with the potential of improving the forecasts.