Both models operate on the same grid, so there are no problems with exchanging fluxes between them. In this paper, however, we focus only on the biological part of the 3D model. The 3D ecosystem model is based on the 1D biological model of Dzierzbicka-Głowacka (2005, 2006). In this model, phytoplankton is represented click here by one state

variable, and the model formulations are based on the simple total inorganic nitrogen (NO3 + NO2 + NH4) cycle. Initially, this nutrient serves to trigger the phytoplankton bloom but later to limit phytoplankton production. The set of CEMBSv1 equations with the biogeochemical processes and parameter values are given in Appendix A and Table 1. The model is conceived for a typical shallow sea, the mixed layer being replenished with nutrients from the bottom. The water column dynamics is implemented

in a three-dimensional frame, where phytoplankton and nutrient (nitrogen) are transported by advection and diffusion. The physical framework, including all the necessary forcing, is presented in Figure 2. The biological model incorporates formulations for the primary production and remineralization mechanisms in the mixed layer, in the lower layer and at the bottom. Primary producers are transported, die and are consumed by zooplankton (mesozooplankton). The grazed phytoplankton is divided into three parts: one contributes to zooplankton growth, another is deposited as faecal pellets, and the third is excreted by zooplankton as dissolved metabolites; thus, it replenishes the nutrient pool. A proportion of the material contributing to growth is assumed to be lost immediately – this represents

dying zooplankton. Proportions of PI3K inhibitor both STK38 faecal and excreted material are immediately regenerated (Radach & Moll 1993, Dzierzbicka-Głowacka 2005). Phytoplankton mortality is modelled in two ways: a) grazing by mesozooplankton, which form the bulk of the grazers in the Baltic Sea – here it is described by the mesozooplankton biomass; b) all other kinds of mortality, like cell lysis and grazing by zooplankton other than mesozooplankton, are assumed to be proportional to the phytoplankton standing stock, with a constant mortality rate, and therefore dynamically coupled to the phytoplankton dynamics. The assumed time scale for the sinking of faecal and dead material a few days old (Jickells et al. 1991) is much less than the time scale for benthic regeneration processes, which is from weeks to months (Billen et al. 1991). Therefore, most of the detrital material is deposited on the bottom, where it collects as a benthic pool. Only a small portion of detritus remains suspended in the water column (Postma & Rommets 1984), i.e. 20% of the remineralized dead phyto- and zooplankton and faecal material in the water column. The effect of the microbial food web (Azam et al. 1983) is parameterized by converting this portion of detrital material immediately into regenerated nutrients in the water column.