326
T. Soomere
model is supposed to adequately resolve the current-driven transport in the area in
question and to provide (numerically simulated Eulerian) velocities at certain grid
points and with a certain temporal resolution over a long time period. Secondly,
the generated velocity data are used to compute Lagrangian trajectories of a large
number of water particles. These particles are interpreted as passive tracers representing potentially adverse substances of neutral buoyancy (e.g., single oil particles)
released at different locations and at different time instants.
Thirdly, the trajectories are analysed with respect to a given cost function and/or
problem setup using standard statistical methods. This analysis leads to a spatial
quantification of the offshore areas. Note that the resulting distributions characterize the points of release of adverse impacts rather than the vulnerable areas. Finally,
the spatial distributions of these quantities (optionally together with additional constraints such as the location of the ports) are used for decision-making, for example,
for the identification of the optimum location of the fairways, by choosing them as
close as possible to the local or global minima of the distributions.
Each of the steps requires a number of parameters, implicit time scales and options that may potentially affect the resulting 2D distributions (maps) and the further
decision-making process. The existing implementations have applied the steps consecutively. This made it possible to improve the particular methods used within steps
separately (Andrejev et al. 2010; Viikmäe et al. 2010). Some relevant issues are discussed in Chap. 9 and in Viikmäe et al. (2010). The dependence of the results on
the choice of the options was evident in the early analysis (Soomere et al. 2010)
where the equiprobability lines were calculated using two slightly different sets of
trajectories. The optimum fairways also depend substantially on the resolution of
the ocean model (Andrejev et al. 2010). The strong seasonal variation of net and
bulk transport patterns (Chap. 9; Soomere et al. 2011d) suggests that seasonally optimum fairways may be radically different for windy and calm seasons in the Gulf
of Finland as also discussed in Chap. 11. Lu et al. (2012) showed that extensive
variations in the optimum fairway correspond to the inflow and outflow situations
in the south-western (SW) Baltic Sea.
10.3.1 The Baltic Sea Under Pressure
There are several reasons why this technique has been first implemented for the
Baltic Sea (Fig. 10.1). This water body is under strong pressure from shipping and
other offshore activities. The concentration of ship traffic (including tankers of various kinds) is exceptionally high (HELCOM 2009). More than 70 large ports handle
more than 1 million tonnes of cargo per year. The number of ship voyages (excluding
ferry traffic) is estimated at 150,000 per year (Gollasch and Leppäkoski 2007), and
it is assumed that it will increase considerably in the future. Since 1980 the Baltic
Sea has experienced on average one major shipping accident per year resulting in
an oil spill larger than 100 tonnes (WWF 2010).
T. Soomere
model is supposed to adequately resolve the current-driven transport in the area in
question and to provide (numerically simulated Eulerian) velocities at certain grid
points and with a certain temporal resolution over a long time period. Secondly,
the generated velocity data are used to compute Lagrangian trajectories of a large
number of water particles. These particles are interpreted as passive tracers representing potentially adverse substances of neutral buoyancy (e.g., single oil particles)
released at different locations and at different time instants.
Thirdly, the trajectories are analysed with respect to a given cost function and/or
problem setup using standard statistical methods. This analysis leads to a spatial
quantification of the offshore areas. Note that the resulting distributions characterize the points of release of adverse impacts rather than the vulnerable areas. Finally,
the spatial distributions of these quantities (optionally together with additional constraints such as the location of the ports) are used for decision-making, for example,
for the identification of the optimum location of the fairways, by choosing them as
close as possible to the local or global minima of the distributions.
Each of the steps requires a number of parameters, implicit time scales and options that may potentially affect the resulting 2D distributions (maps) and the further
decision-making process. The existing implementations have applied the steps consecutively. This made it possible to improve the particular methods used within steps
separately (Andrejev et al. 2010; Viikmäe et al. 2010). Some relevant issues are discussed in Chap. 9 and in Viikmäe et al. (2010). The dependence of the results on
the choice of the options was evident in the early analysis (Soomere et al. 2010)
where the equiprobability lines were calculated using two slightly different sets of
trajectories. The optimum fairways also depend substantially on the resolution of
the ocean model (Andrejev et al. 2010). The strong seasonal variation of net and
bulk transport patterns (Chap. 9; Soomere et al. 2011d) suggests that seasonally optimum fairways may be radically different for windy and calm seasons in the Gulf
of Finland as also discussed in Chap. 11. Lu et al. (2012) showed that extensive
variations in the optimum fairway correspond to the inflow and outflow situations
in the south-western (SW) Baltic Sea.
10.3.1 The Baltic Sea Under Pressure
There are several reasons why this technique has been first implemented for the
Baltic Sea (Fig. 10.1). This water body is under strong pressure from shipping and
other offshore activities. The concentration of ship traffic (including tankers of various kinds) is exceptionally high (HELCOM 2009). More than 70 large ports handle
more than 1 million tonnes of cargo per year. The number of ship voyages (excluding
ferry traffic) is estimated at 150,000 per year (Gollasch and Leppäkoski 2007), and
it is assumed that it will increase considerably in the future. Since 1980 the Baltic
Sea has experienced on average one major shipping accident per year resulting in
an oil spill larger than 100 tonnes (WWF 2010).