4 Studying the Baltic Sea Circulation with Eulerian Tracers
115
over sea, for instance, with slightly too warm air temperatures during summer and
slightly too cold temperatures during winter.
Nevertheless, many modelling studies for the Baltic Sea have applied this data
set successfully as atmospheric forcing (e.g., Lehmann 1995; Neumann et al. 2002;
Meier et al. 2003; Meier 2007; Andrejev et al. 2004a, 2004b; Myrberg and Andrejev
2003, 2006). Unfortunately, SMHI stopped its production in 2011.
The second alternative for the atmospheric forcing is the output of a regional
high-resolution atmosphere model. As the resolution of global reanalysis products
like ERA-40 (Uppala et al. 2005) or ERA-Interim (Dee et al. 2011) is too coarse
for Baltic Sea applications (Omstedt et al. 2005), surface fields of a regional highresolution model are needed. Samuelsson et al. (2011) evaluated surface fields of a
regional atmosphere model driven with ERA-40 data at the lateral boundary. Furthermore, Meier et al. (2011) assessed the quality of atmospheric surface fields over
the Baltic Sea of a regional coupled atmosphere-ocean model with respect to ocean
dynamics. In both studies it was concluded that the modelled atmospheric surface
fields have sufficiently good quality and the data might be used as forcing for ocean
models. This approach was applied, for instance, by Höglund and Meier (2012).
However, there are also disadvantages. The model domain of the regional atmosphere model by Samuelsson et al. (2011) is rather large allowing a regional weather
development that differs from the observed weather due to the chaotic nature of the
atmospheric circulation. Hence, the tracks of low and high pressure anomalies might
be shifted in time or space compared to observations. Due to this internal variability the quality of the atmospheric forcing might be limited and data assimilation is
needed. At present regional reanalyses utilizing the assimilation of observations into
regional high-resolution atmosphere models are under development (Luhamaa et al.
2011) and will soon replace earlier downscaling products that lack data assimilation.
The use of proper atmospheric data is crucial in Baltic Sea applications because
this water body is rather shallow with a mean water depth of 54 m (Seifert et al.
2001) and the variability of its dynamics is largely controlled by atmospheric variability. Further, as mixing in the ocean depends on the cubed wind speed, it is important for the stratification in the Baltic Sea that wind speed extremes are correctly represented. Otherwise the simulated vertical stratification might be biased.
Höglund et al. (2009) showed that wind speed extremes in a state-of-the-art regional
atmosphere model are underestimated. To compensate for this shortcoming they developed a bias-correction based upon the simulated wind gustiness (Höglund et al.
2009). There are many studies devoted to this problem. However a more detailed
discussion on simulated wind speed extremes is beyond the scope of this chapter.
4.2.2.7 River Runoff
Finally, runoff from rivers and other fresh water sources are needed to simulate the
water cycle of the Baltic Sea. River runoff is the main contributor to the fresh water
surplus of the Baltic Sea. Note that the Baltic Sea catchment area is four times larger
than the surface of the Baltic Sea (Sjöberg 1992). In the RCO model the 29 largest
115
over sea, for instance, with slightly too warm air temperatures during summer and
slightly too cold temperatures during winter.
Nevertheless, many modelling studies for the Baltic Sea have applied this data
set successfully as atmospheric forcing (e.g., Lehmann 1995; Neumann et al. 2002;
Meier et al. 2003; Meier 2007; Andrejev et al. 2004a, 2004b; Myrberg and Andrejev
2003, 2006). Unfortunately, SMHI stopped its production in 2011.
The second alternative for the atmospheric forcing is the output of a regional
high-resolution atmosphere model. As the resolution of global reanalysis products
like ERA-40 (Uppala et al. 2005) or ERA-Interim (Dee et al. 2011) is too coarse
for Baltic Sea applications (Omstedt et al. 2005), surface fields of a regional highresolution model are needed. Samuelsson et al. (2011) evaluated surface fields of a
regional atmosphere model driven with ERA-40 data at the lateral boundary. Furthermore, Meier et al. (2011) assessed the quality of atmospheric surface fields over
the Baltic Sea of a regional coupled atmosphere-ocean model with respect to ocean
dynamics. In both studies it was concluded that the modelled atmospheric surface
fields have sufficiently good quality and the data might be used as forcing for ocean
models. This approach was applied, for instance, by Höglund and Meier (2012).
However, there are also disadvantages. The model domain of the regional atmosphere model by Samuelsson et al. (2011) is rather large allowing a regional weather
development that differs from the observed weather due to the chaotic nature of the
atmospheric circulation. Hence, the tracks of low and high pressure anomalies might
be shifted in time or space compared to observations. Due to this internal variability the quality of the atmospheric forcing might be limited and data assimilation is
needed. At present regional reanalyses utilizing the assimilation of observations into
regional high-resolution atmosphere models are under development (Luhamaa et al.
2011) and will soon replace earlier downscaling products that lack data assimilation.
The use of proper atmospheric data is crucial in Baltic Sea applications because
this water body is rather shallow with a mean water depth of 54 m (Seifert et al.
2001) and the variability of its dynamics is largely controlled by atmospheric variability. Further, as mixing in the ocean depends on the cubed wind speed, it is important for the stratification in the Baltic Sea that wind speed extremes are correctly represented. Otherwise the simulated vertical stratification might be biased.
Höglund et al. (2009) showed that wind speed extremes in a state-of-the-art regional
atmosphere model are underestimated. To compensate for this shortcoming they developed a bias-correction based upon the simulated wind gustiness (Höglund et al.
2009). There are many studies devoted to this problem. However a more detailed
discussion on simulated wind speed extremes is beyond the scope of this chapter.
4.2.2.7 River Runoff
Finally, runoff from rivers and other fresh water sources are needed to simulate the
water cycle of the Baltic Sea. River runoff is the main contributor to the fresh water
surplus of the Baltic Sea. Note that the Baltic Sea catchment area is four times larger
than the surface of the Baltic Sea (Sjöberg 1992). In the RCO model the 29 largest