Sea water temperature

The temperature of sea water in the surface layer and in deep waters is calculated using the 3D CEMBS ecosystem model of the Baltic Sea, from an  assimilation of remote-sensing Sea Surface Temperature (SST) data. In computer models SST is taken to mean the temperature of the model’s first layer. In  the case of the 3D CEMBS model this layer is 5 m thick. The SO SatBałtyk system gives distributions of SST and deep water temperatures in the Baltic in  degrees Celsius [^oC], four times a day, in the form of maps with a horizontal resolution of 1 km.

The methodology of determining sea water temperature

The temperature of sea water is calculated using the 3D CEMBS ecohydrodynamic prognostic model. The boundary conditions at the sea surface are  atmospheric data taken from the UM prognostic model (ICM University of Warsaw). The model has an open boundary with the North Sea so that changes  ensuing from inflows of North Sea water can be better mapped. The 3D CEMBS model is equipped with a river inflow module, which supplies information  on  the amount of fresh water supplied by 72 rivers and the quantities of nutrients entering the sea with it.  KPP parameterization is used to ensure that the  vertical mixing has been mapped as accurately as possible. The model has a module for assimilating satellite data from the MODIS sensor on the Aqua (EOS  PM) satellite. The model results have a horizontal resolution of ca 2 km. In the vertical the model is divided into 21 layers. Each of the first four layers is 5 m thick. The thickness of the other layers increases with depth. The results supplied to the SO SatBałtyk system are interpolated on a grid with a resolution of 1 km. Cressman’s method is used for data assimilation. This involves calculating, for every point on the model grid, the temperature on the basis of data from the previous model forecast and satellite data taken into account with a weight dependent on the distance between these data and on the accuracy of the results. The algorithm ignores overcast areas and remotely-sensed SSTs widely divergent from the model values. More information on this method and the results of determining sea water temperature can be found in the article Assimilation of the satellite SST data in the 3D CEMBS model (Nowicki et al. 2014).

Validation (assessment of accuracy)

The 3D CEMBS model was validated with respect to the data gathered during cruises of R/V Oceania, as well as those from ICES databases for the years  2010-2014. SSTs measured and calculated by the model are compared on the figure alongside. The statistical error, expressed as the standard deviation of  the differences between values of T_S calculated by the 3D CEMBS model and measured in situ, is estimated at 1.53°C. The systematic error (the mean  difference between measured and calculated values) is -0.06°C. The correlation coefficient r = 0.97. More information can be found in the article mentioned  in the previous paragraph.