Transport modelling in coastal waters using stochastic differential equations

Abstract

In this thesis, the particle model that takes into account the short term correlation behaviour of pollutants dispersion has been developed. An efficient particle model for sediment transport has been developed. We have modified the existing particle model by adding extra equations for the suspension using a probabilistic concepts (the Poisson distribution function) to determine the actual number of particles to suspend in each cell. The deposition is modelled by an exponential decaying ordinary differential equation. In order to get accurate results from Monte Carlo simulations of sediment transport, a large number of particles is often needed. However, computation time in a particle model increases linearly with the number of particles. Thus, we have developed a high performance particle model for sediment transport by considering three different sediment suspension methods. Parallel simulation experiments are performed in order to investigate the efficiency of these three methods. We conclude that the second method is the best method on distributed computing systems (e.g., a Beowulf cluster), whereas the third maintains the best load distribution. Using variable time stepping to integrate the particle track in this thesis, has also proved to be efficient.