A model for the simulation of particle movements in water should incorporate the mutual distance dependent correlation. As long as reliable data are accessible a model can be created of the dispersion in a given area from a statistical description of turbulence. Current measurements have been performed in an area north of the Swedish nuclear power plant Barsebäck, and statistical time series analysis have made it possible to estimate multivariate autoregressive moving-average (ARMA) models for these data using the Box-Jenkins method. The correlation structure for the area has been investigated in detail. Transport and dispersion models for the marine environment are used in estimating doses to the population from the aquatic food chain. Some of these models are described with special emphasis on the time and length scales they cover. Furthermore, to illustrate the background of the simulation model, short introductions are given to health physics, time series analysis, and turbulence theory. Analysis of the simulation model shows the relative importance of the different parameters. The model can be expanded to conditional simulation, where the current measurements are used directly to simulate the movement of one of the particles. Results from the model are also compared to results from a sampling of bioindicators (Pucus vesiculosus) along the Danish coast. The reliability of bioindicators in this kind of experiment is discussed.
|Place of Publication||Roskilde|
|Publisher||Risø National Laboratory|
|Number of pages||151|
|Publication status||Published - 1986|