332
T. Soomere
model with a horizontal resolution of 0.5 nm and vertical resolution of 1 m covering
the transition area from Skagen to Bornholm (Fig. 10.1) was two-way nested into
the above Baltic Sea–North Sea model. A detailed description of this model and its
forcing and boundary conditions is presented by Lu et al. (2012). Some interesting
results from this model are also described in Chap. 5.
10.4.4 Trajectory Simulations Using TRACMASS
The test elements used to evaluate the risk of coastal hit by pollution released at a
particular sea point and then transported by surface currents to the nearshore were
numerically simulated Lagrangian trajectories of water particles passively carried
by the currents. The trajectories were locked in the uppermost layer (with a depth
of 3 m in the RCO model, 2 m in the OAAS model and 1 m in the DMI/BSHcmod
model) and exerted only horizontal advection. They were thus not truly Lagrangian
and basically represented the motion of items persistently located in the surface
layer and completely following its motion. Such items and substances are generally
also affected by wind drag or wave impact. The impact of these drivers is addressed
in Chap. 11. Here we only consider the current-driven advection of such items.
‘Off-line’ simulation of the pathways of selected particles (in which the circulation modelling was separated from the trajectory calculation, see Chap. 7) was used
to build the trajectories from the output of the RCO model for the Gulf of Finland
(Soomere et al. 2011c, 2011d) and the northern Baltic Proper (Viikmäe et al. 2011),
and from the data of the DMI/BSHcmod model for the SW Baltic Sea (Lu et al.
2012). The trajectories were calculated with the TRACMASS code (Blanke and
Raynard 1997; Döös 1995; de Vries and Döös 2001) using precomputed 3D velocity fields to reconstruct the motion of the particles. See Chap. 7 for the underlying
theory and several examples of the use of this model.
The results depend to some extent on the temporal resolution of both the circulation data and the accuracy of the trajectory calculation. The resulting inaccuracies of
the latter may lead to a large divergence of a few simulated trajectories from the real
ones (Gräwe and Wolff 2010). Given the large number of trajectories, it is natural
to assume that the potential impact of the time step of the circulation data and the
trajectory reconstruction scheme on the resulting statistics is minor.
In the marine environment the particles are not only transported with the instantaneous simulated velocity in each grid cell but are also affected by local motions
such as diffusive processes, small-scale turbulence, local vertical motions of water
particles, etc. (so-called subgrid-scale turbulence). The latter is only partially accounted for in the circulation models and, even theoretically, cannot be adequately
represented in these trajectory models that exclusively rely on precomputed velocity
fields. It is only possible to replicate certain statistical features of trajectories in such
models.
The version of the TRACMASS code used in the described calculations did not
account for the effects of subgrid turbulence. In such cases the initially close trajectories have an overly tendency to stay close while in the real atmospheric and marine
T. Soomere
model with a horizontal resolution of 0.5 nm and vertical resolution of 1 m covering
the transition area from Skagen to Bornholm (Fig. 10.1) was two-way nested into
the above Baltic Sea–North Sea model. A detailed description of this model and its
forcing and boundary conditions is presented by Lu et al. (2012). Some interesting
results from this model are also described in Chap. 5.
10.4.4 Trajectory Simulations Using TRACMASS
The test elements used to evaluate the risk of coastal hit by pollution released at a
particular sea point and then transported by surface currents to the nearshore were
numerically simulated Lagrangian trajectories of water particles passively carried
by the currents. The trajectories were locked in the uppermost layer (with a depth
of 3 m in the RCO model, 2 m in the OAAS model and 1 m in the DMI/BSHcmod
model) and exerted only horizontal advection. They were thus not truly Lagrangian
and basically represented the motion of items persistently located in the surface
layer and completely following its motion. Such items and substances are generally
also affected by wind drag or wave impact. The impact of these drivers is addressed
in Chap. 11. Here we only consider the current-driven advection of such items.
‘Off-line’ simulation of the pathways of selected particles (in which the circulation modelling was separated from the trajectory calculation, see Chap. 7) was used
to build the trajectories from the output of the RCO model for the Gulf of Finland
(Soomere et al. 2011c, 2011d) and the northern Baltic Proper (Viikmäe et al. 2011),
and from the data of the DMI/BSHcmod model for the SW Baltic Sea (Lu et al.
2012). The trajectories were calculated with the TRACMASS code (Blanke and
Raynard 1997; Döös 1995; de Vries and Döös 2001) using precomputed 3D velocity fields to reconstruct the motion of the particles. See Chap. 7 for the underlying
theory and several examples of the use of this model.
The results depend to some extent on the temporal resolution of both the circulation data and the accuracy of the trajectory calculation. The resulting inaccuracies of
the latter may lead to a large divergence of a few simulated trajectories from the real
ones (Gräwe and Wolff 2010). Given the large number of trajectories, it is natural
to assume that the potential impact of the time step of the circulation data and the
trajectory reconstruction scheme on the resulting statistics is minor.
In the marine environment the particles are not only transported with the instantaneous simulated velocity in each grid cell but are also affected by local motions
such as diffusive processes, small-scale turbulence, local vertical motions of water
particles, etc. (so-called subgrid-scale turbulence). The latter is only partially accounted for in the circulation models and, even theoretically, cannot be adequately
represented in these trajectory models that exclusively rely on precomputed velocity
fields. It is only possible to replicate certain statistical features of trajectories in such
models.
The version of the TRACMASS code used in the described calculations did not
account for the effects of subgrid turbulence. In such cases the initially close trajectories have an overly tendency to stay close while in the real atmospheric and marine