13 2.3.2. Model descriptions Four models were configured and run for the Baltic Sea intercomparison exercise. The models used include two box models (POSEIDON and NRPA model) and a three dimensional (3-D) model accounting for density gradients and incorporating ice thermodynamics (THREETOX). Lastly, an intermediate approach, consisting of a two dimensional (2-D) depth averaged model forced with annual mean winds (USEV model), has also been used. These are described in Annex I (NRPA model), II (POSEIDON), III (THREETOX) and V (USEV model). It is clear therefore that the intercomparison included a wide range of modelling approaches. For THREETOX and the USEV model, hydrodynamic models were used to calculate the current fields and used as inputs in the simulation of the transport of radionuclides. An advection–di?usion equation is solved for this purpose, which incorporates additional terms accounting for radionuclide exchanges between water, suspended matter and bed sediments. In the case of box models, water ?uxes between boxes are used in the conventional way. The main characteristics of the applied models are summarized in Table 7. 2.3.3. Modelling endpoints Simulations were started in October 1986 and ran for a duration of five years. This start time is around 6 months after the Chornobyl accident deposition event took place and when the first measurements of the fallout distribution in the Baltic Sea were carried out (see Figure 4 above; which was prepared using various measurements [23–25]). The same endpoints were calculated with each model in order to facilitate model–model and model–measurement comparisons, the latter with data from the HELCOM database. Also, estimates of annual inventories in the water column and bed sediments have been made from various measurements [32]. The modelling endpoints are: ? Time series of total 137Cs inventories in the Baltic water column and in bed sediments (a single value at the end of each month) (see Figure 5); ? Time series of 137Cs activity concentrations in surface water and bed sediments at locations indicated in Table 8 (daily values) (see Figures 6 and 7); ? Time series of mean activity concentrations of 137Cs in the water column and bed sediment in the regions described in Figure 2 above (a single value at the end of each month) (see Figures 8 and 9); ? Maps of 137Cs activity concentration in surface water (Bq/m3) and bed sediments (Bq/kg) at the end of the simulation time (31 October 1991) (see Figure 10). No calibration using the results of measurements was performed for the POSEIDON, THREETOX and NRPA models. Instead, default parameters were used. In the case of the USEV model, data on 137Cs inventories in the water column and seabed were used to calibrate uptake/release processes, as described in Annex V.