Environmental Engineering Reference
In-Depth Information
switched off; see below) as well as vertical movements (fixed depth of 25 m, which
is approximately half the typical mixed layer depth in the area at the time), and all
cod larvae settled to the seabed when they reached a given age (90 days). The
particles initially represented cod eggs, and their hatching time and subsequent
(post-hatch) larval growth were modelled as functions of their individual tempera-
ture history. Egg development is largely dependent on temperature (e.g. Fox et al.
2003) and temperature, in addition to its direct effect, can be used as a proxy for
other co-varying factors influencing larval fish growth in the field, such as food
availability, day-length, water column stability, etc. (Gallego et al. 1999). Daily
temperature fields were linearly interpolated from monthly mean three-dimensional
temperature fields derived from a statistical model of observational data (ICES
statistical rectangle horizontal resolution: 0.5
latitude, 1.0
longitude and five
layers in the upper 150 m; Heath et al. 2003). In the baseline run, the particles
were not subject to mortality. Each particle was considered a “superindividual”, i.e.
an assemblage of identical individuals. The weighting of each particle was given as
a fraction of the sub-population's egg production (total egg production by the sub-
population divided by the number of particles released within its spawning area).
When applied, mortality was implemented by reducing the weighting of each
individual particle (see below).
The effect of mortality was investigated by applying a mortality rate parameterized
as a function of length whereby faster growing superindividuals experienced lower
cumulative mortality over the simulation period.
The effect of horizontal diffusion was tested by subjecting the particles to an
additional stochastic horizontal velocity component. This was computed at particle
positions as a function of horizontal velocity shear from the hourly residual plus
tidal current fields, following a method derived from Oey and Chen (1992). To
achieve statistical stability (see Brickman et al. 2009), multiple particles must be
released from each start position so I carried out runs with 10, 50 and 100 particles
per start position.
The effect of particle release pattern was tested by comparing the baseline,
where all annual egg production was released on the date of peak spawning for
each subpopulation, with a run where the start and end dates of the spawning season
were fixed for each spawning area, based on field observations. The proportion of
the annual egg production shed each day was represented by a Gaussian curve with
mean at the date of peak spawning and standard deviation equal to 1/6 of the
spawning season duration.
It is generally accepted that planktonic organisms are passively advected by the
horizontal flow but are capable of vertical migration. To test its effect, the baseline
simulation (vertical depth fixed at 25 m; see above) was compared with a simulation
where particles migrated with a sinusoidal diel vertical migration (DVM) pattern
between 5 and 45 m depth. It was also compared with a more complex pattern
where the vertical position of cod eggs and larvae varied ontogenetically, using a
model (Heath et al. 2003) derived from observational data reported in the literature.
At the end of their pelagic phase, cod juveniles become demersal. This transition
period is likely to be one when density-dependence can influence year-class strength
