380
J. Murawski and J. Woge Nielsen
spreading of oil, and subsurface processes (vertical dispersion). Evaporation is
treated fraction-wise as a function of mass, molecular weight, vapour pressure and
wind velocity according to Audunson (1980), Dippner (1984). More volatile fractions eventually evaporate completely within a relatively short time interval. The
loss in the oil mass is compensated by entrainment of sea water into water-in-oilemulsion. The intensity of this process depends on wind speed and is modelled as
a saturation process leading eventually to a maximum uptake of ∼70 % of water
(Mackay et al. 1982, 1983).
Oil slicks in sea water undergo advection, spreading and dispersion due to currents and winds. First of all, oil parcels are carried by ocean currents (current velocity u) and also impacted by the wind. These (advection) processes are modelled
using an Eulerian time stepping scheme ∂s = u · ∂t. The direct impact of wind is
accounted for by means of adding 3.0 % of the wind velocity in the direction of
the wind at the surface to the surface current (Reed and Aamo 1994). Spreading is
modelled using modified Fay equations (Fay 1969, 1971) where spreading in the
wind direction is enhanced by setting the spill radius in this direction 1 + 0.5U 10
times larger than the radius in the cross-wind direction.
Horizontal dispersion and turbulence are treated as a random subgrid process by
adding to u a random component Ru and a component equivalent to the Brownian motion R B . Vertical dispersion is modelled as an interplay of upwards buoyant
motion according to Stokes’ law and wave-induced downwards mixing, treated as a
turbulent diffusive process with wind- and wave-dependent eddy diffusivity (Sverdrup et al. 1942; Leibovich 1975). Turbulence effects are included as random droplet
size distributions of buoyant oil parcels (Johansen 1982; Forrester 1971) and random
contributions to the eddy diffusivity. Submerged oil parcels are staying submerged
in most cases and usually strand at the sea bottom. These parcels are removed from
the calculations and do not enter the drift analysis below and only those near-surface
oil parcels that reach the shoreline are accounted for.
11.3.2 Oil Spill Modelling in the Gulf of Finland
Circulation simulations for the subsequent analysis of oil drift in the Baltic Sea
employed DMI’s operational circulation model for storm surge warning and ocean
forecasts (the code is tagged as MyOcean V1). The applied set-up (Fig. 11.2) covers
the North Sea and Baltic Sea with a spatial resolution of 3 nautical miles (nm)
constructed using the 1 nm bathymetry of Seifert et al. (2001), and the transition
area in a higher resolution of about 0.5 nm. The target area of the Gulf of Finland
was also covered with a 0.5 nm grid that was two-way nested into the Baltic Sea
grid. Doing so made it possible to model 3D ocean currents with a high spatial
resolution for oil drift simulations in this area. The bathymetry of the Gulf of Finland
was constructed from the gridded depth information with a resolution of 0.25 nm
(Andrejev et al. 2010) and thus adequately replicates the actual small-scale features
of the seabed in this basin, except in the central deep area of the gulf. Owing to
J. Murawski and J. Woge Nielsen
spreading of oil, and subsurface processes (vertical dispersion). Evaporation is
treated fraction-wise as a function of mass, molecular weight, vapour pressure and
wind velocity according to Audunson (1980), Dippner (1984). More volatile fractions eventually evaporate completely within a relatively short time interval. The
loss in the oil mass is compensated by entrainment of sea water into water-in-oilemulsion. The intensity of this process depends on wind speed and is modelled as
a saturation process leading eventually to a maximum uptake of ∼70 % of water
(Mackay et al. 1982, 1983).
Oil slicks in sea water undergo advection, spreading and dispersion due to currents and winds. First of all, oil parcels are carried by ocean currents (current velocity u) and also impacted by the wind. These (advection) processes are modelled
using an Eulerian time stepping scheme ∂s = u · ∂t. The direct impact of wind is
accounted for by means of adding 3.0 % of the wind velocity in the direction of
the wind at the surface to the surface current (Reed and Aamo 1994). Spreading is
modelled using modified Fay equations (Fay 1969, 1971) where spreading in the
wind direction is enhanced by setting the spill radius in this direction 1 + 0.5U 10
times larger than the radius in the cross-wind direction.
Horizontal dispersion and turbulence are treated as a random subgrid process by
adding to u a random component Ru and a component equivalent to the Brownian motion R B . Vertical dispersion is modelled as an interplay of upwards buoyant
motion according to Stokes’ law and wave-induced downwards mixing, treated as a
turbulent diffusive process with wind- and wave-dependent eddy diffusivity (Sverdrup et al. 1942; Leibovich 1975). Turbulence effects are included as random droplet
size distributions of buoyant oil parcels (Johansen 1982; Forrester 1971) and random
contributions to the eddy diffusivity. Submerged oil parcels are staying submerged
in most cases and usually strand at the sea bottom. These parcels are removed from
the calculations and do not enter the drift analysis below and only those near-surface
oil parcels that reach the shoreline are accounted for.
11.3.2 Oil Spill Modelling in the Gulf of Finland
Circulation simulations for the subsequent analysis of oil drift in the Baltic Sea
employed DMI’s operational circulation model for storm surge warning and ocean
forecasts (the code is tagged as MyOcean V1). The applied set-up (Fig. 11.2) covers
the North Sea and Baltic Sea with a spatial resolution of 3 nautical miles (nm)
constructed using the 1 nm bathymetry of Seifert et al. (2001), and the transition
area in a higher resolution of about 0.5 nm. The target area of the Gulf of Finland
was also covered with a 0.5 nm grid that was two-way nested into the Baltic Sea
grid. Doing so made it possible to model 3D ocean currents with a high spatial
resolution for oil drift simulations in this area. The bathymetry of the Gulf of Finland
was constructed from the gridded depth information with a resolution of 0.25 nm
(Andrejev et al. 2010) and thus adequately replicates the actual small-scale features
of the seabed in this basin, except in the central deep area of the gulf. Owing to