The purpose of this study is to develop a method for optimizing the data assimilation system of the HIROMB-BOOS -model at the Finnish Meteorological Institute by finding an optimal time interval and an optimal grid for the data assimilation. This is needed to balance the extra time the data assimilation adds to the runtime of the model and the improved accuracy it provides.
Data assimilation is the process of combining observations with a numerical model to improve the accuracy of the model. There are different ways of doing this, some of which are covered in this work.
The HIROMB-BOOS -circulation model is a 3D-forecast model for the Baltic Sea. The variables forecast are temperature, salinity, sea surface height, currents, ice thickness and ice coverage. Some of the most important model equations are explained here.
The HIROMB-BOOS -model at the Finnish Meteorological Institute has a preoperational data assimilation system that is based on the optimal interpolation method. In this study the model was run for a 2-month test period with different time intervals of data assimilation and different assimilation grids. The results were compared to data from five buoys in the Baltic Sea.
The model gives more accurate results when the time interval of the data assimilation is small. The thicker the data assimilation grid is, the better the results. An optimal time interval was determined taking into account the time the assimilation takes. An optimal grid was visually determined based on an optimal grid thickness, for which the added time had to be considered as well.
The optimized data assimilation scheme was tested by performing a 12-month test run and comparing the results to buoy data. The optimized data assimilation has a positive effect on the model results.
