Spill Science & Technology Bulletin, Vol.4, No.2, 71-88, 1997
Simulation of coastal oil spills using the random walk particle method with Gaussian kernel weighting
Oil spills, amongst the most common marine hazards, are a major community concern, particularly in coastal areas where there are competing requirements due to population centres, conservation of the natural environment, industrial development and transport. It has become common practice to include simulations and modelling exercises in applications for development, so that oil spill trajectory studies are regularly carried out for planning and assessment purposes as well as for making decisions in the event of actual incidents. In this paper we present a procedure for the simulation of coastal oil spills which uses the random walk particle method with Gaussian kernel weighting to simulate the movement of oil in the sea, and includes special beaching and refloating algorithms for shorelines. This procedure allows one to simulate coastal spills rapidly and accurately and to perform calculations at closed boundaries consisting of a variety of shoreline types such as exposed headlands, sand beaches, sheltered marshes, etc., and at open boundaries with an extended computational region. Moreover, the mass concentration of the spill can be calculated at any point of a finite diffusion domain and conveniently represented by contour and mesh plots. Simulations in a typical coastal region are presented to illustrate the procedure for spills from instantaneous point sources, instantaneous line sources, instantaneous area sources, continuous point sources and continuous moving sources.