Coastal oil spills are serious environmental disasters often leading to significant long-term impacts. From 1978 to 1995, there were in excess of 4,100 major oil spills of 10,000 gallons or more (Etkin and Welch, 1997). There has, however, been a downward trend in the number of major incidences from a peak in 1991, but with approximately 3 billion gallons of oil in daily use worldwide, a large proportion of which is transported at sea, the threat of coastal oil spills remains acute. On Friday, March 24, 1989, the Exxon Valdez, carrying 1.25 million barrels of crude oil, ran hard aground on Blight Reef, Prince William Sound, Alaska, spilling 258,000 barrels of oil mostly in the first few hours.
Since contingency disaster planning had only focused on spills of up to 2,000 barrels, the spill could not be contained and instead spread to nearly 2,000 km of coastline devastating fisheries, wildlife, and scenic beauty. Restoration cost over $10 billion.
The U.K. government’s estimate for cleanup, salvage, and losses of fisheries and tourism was up to £43 million (Harris, 1997). On Tuesday November 19, 2002, the Prestige broke up and sank 210 km off Spain’s northern coast. A few days earlier, the Prestige had been holed off Galicia, Spain, spreading oil to the nearby coast. The Spanish authorities had ordered the tanker towed away from the coast with perhaps inevitable consequences. For at least two months, oil continued to seep from the wreck to be spread by tide and wind onto the sandy beaches around the Bay of Biscay from Galicia to La Rochelle in France with again devastation to wildlife, fisheries, and tourism with severe social and economic consequences in many coastal towns and villages.
In a coastal oil spill incident, oil floats and spreads out rapidly across the water surface to form a thin layer—a slick. As the spreading process continues, the layer becomes progressively thinner, finally becoming a sheen. Complex interrelated physical, chemical, and biological processes depending on the type of oil, the hydrodynamics, and other environmental conditions govern the behavior of the slick and the sheen. Drifting of the oil is mainly by advection and diffusion due to currents (tide and wind).
Analysis and assessment of the risks, however, rely heavily on numerical simulation of coastal oil spill behavior so that the environmental impact assessment (EIA) and contingency planning can be pertinent in the protection of sensitive areas and installations. Simulation of oil spill behavior requires more than one model, each having a specific task: hydrodynamic modeling, trajectory modeling, and fate modeling. The structure of typical coastal oil spill modeling is given in Figure 6.13, which represents a dynamic 2D distributed parameter model.
We have already seen the hydrodynamic modeling in Chapter 5. Figure 5.25 showed an example of currents simulated using FEM for hydrodynamic modeling. The tidal currents calculated from the forced inputs at the open boundary were shown graphically with the arrow size proportional to the speed of the current. The arrow direction, as shown in Figure 5.25, is actually a resultant vector calculated from two components: a northerly component U and an easterly component V. Each of these components has itself two components, which describe it: a velocity in m/s Ua Va and a deflection in radians Ug Vg. For each tidal constituent in the hydrodynamic modeling, the current simulation can be represented by the following equations (Li, 2001):
U(x, t) = Ua · cos(ωt – Ug) (6.3)
V(x, t) = Va · cos(ωt – Vg) (6.4)
where Ua = amplitude of northerly component, Ug = deflection of northerly component, Va = amplitude of easterly component, Vg = deflection of easterly component, ω = angular frequency of the relevant tidal constituent, x = location, t = time step.
The purpose of the hydrodynamic modeling is to calculate a time series of currents from tidal constituents and bathymetry over the whole coastal study area and is in itself a means of calculating and distributing parameters to which are added further distributed parameters for the trajectory and fate modeling. Given these distributed parameters, many of which are in time series, the trajectory model becomes predominantly a routing simulation that requires an arithmetic solution across a grid. Consequently, the output from the hydrodynamic model is reinterpolated into a grid with additional inputs, such as mean current, wind direction (for wind-driven currents), and oil properties.