25 TABLE 9. MODELLING EXERCISES CARRIED OUT FOR THE PACIFIC OCEAN RADIOLOGICAL SCENARIO Exercise Features Source Radionuclide 1 Own circulation and parameters Hypothetical Tracer 2 Same circulation, own parameters Hypothetical Tracer; 137Cs 3 Same circulation and parameters Hypothetical Tracer; 137Cs 4a Same circulation and parameters Realistic 137Cs 4b Own circulation and parameters Realistic 137Cs TABLE 10. MODELS APPLIED TO SIMULATE FUKUSHIMA RELEASES IN THE PACIFIC OCEAN Model Country Circulation SELFE/IMMSP/KIOST Ukraine/Rep. of Korea Own KAERI LORAS* model Rep. of Korea NCOM** and JCOPE2*** NTUA Greece Own JAEA SEA-GEARN model Japan Univ. of Kyoto USEV 3-D model Spain JCOPE2 Sisbahia model Brazil Own Notes: The origin of water circulation is also given. ‘Own’ means that the circulation is calculated by the model; in other cases the name of the ocean forecasting model is given. * Lagrangian Oceanic Radiological Assessment System ** Navy Coastal Ocean Model *** Japanese Costal Ocean Predictability Experiment 3.2. EXERCISE 1: INITIAL MODEL COMPARISON 3.2.1. Modelling exercise A very simple exercise was initially carried out to compare the performance of models under simple conditions. A constant source of a perfectly conservative radionuclide (tracer) was considered. The magnitude of the input/source term was arbitrarily defined as 1.0 × 106 Bq/s of a long lived radionuclide (i.e. radioactive decay can be omitted). The release is modelled to start on 26 March 2011, and the time frame of calculations extends until 30 May 2011. Each model is run using varied water circulation as summarized in Table 10 and set or required parameters. The output of the models is a time series of tracer concentration in surface water at three points aligned with the Fukushima Daiichi NPP (P1, P2 and P3), and at various distances from it. The locations of these points are presented in Figure 11. 3.2.2. Results The time series of calculated concentrations in the ocean surface for the three considered points are illustrated in Figures 12–14. JAEA has applied two models, i.e. a finite di?erence (JAEA FDM) and a particle-tracking model (JAEA PT). KAERI has run the same model using circulation from two hydrodynamic models, i.e. JCOPE2 and NCOM. Point P1 is some tens of kilometres o?shore from Fukushima. The arrival of the signal to P1 is similar for all models, although in the case of the SELFE/IMMSP/KIOST model (denoted by the abbreviation I/K or I/K-E when running in Eulerian form) there is a delay of approximately 20 days with respect to the others. Predicted concentrations extend over two to three orders of magnitude. Even at the closest point to Fukushima (P2, Figure 13), predictions expand over