Particle Fluxes at the Edge of the Ross Ice Shelf: the Rôle of Physical Forcing
189
of particulate matter in the océan [39-41], by larger particles of biogenic origin.
Settling particles resulting from blooms of large dense cells or from aggregation
of small particles can constitute short-term puises rapidly reaching deep layers
and sédiments and scavenging fine minerai particles of the water column [42-45].
In our study, transmissometer and sédiment trap collection records generally
show a good corrélation, with the exception of the period May-August (i.e. from
120 to 240 Julian days). To try to understand the discrepancy between transmissometer and sédiment trap records in this period a simple Montecarlo model of
particle downfall was implemented.
Dunbar et al. [46] hâve found that deep currents are weaker in the investigated
area than in other zones of the Ross Sea continental shelf. A one-dimensional
model for the settling of biogenic material is thus likely to be successfully applicable there. According to the assumptions of Jaeger et al.[47], uninterrupted downward sinking of biogenic particles (produced by primary and secondary production processes mostly in the 0- to 100 m-layer) supplied from the surface is
assumed to originate at the bottom of the surface layer (100 m) and settle to the
seabed. The total downfall velocity is computed as the sum of intrinsic downfall
velocity of particles plus the actual vertical displacement of the water layer in
which they are contained. A set of pseudoparticles is numerically produced in our
simulation at 100-m depth, with a Gaussian spectrum of their intrinsic velocity.
The value of intrinsic downfall velocity is extracted randomly on the basis of a
gaussian shape with an average of 100 m d'1 and a variance of 20 m d 1 in order
to simulate a wider spectrum of real particle size and shape, consequently implying different downfall velocities. Particle intrinsic settling velocities are highly
variable. The chosen value of 90 ± 20 m day1 is in the range of the settling velocities of the particles mostly represented in our traps. For fecal pellets we refer to
measurements performed by Jaeger and collègues [47] in the study area. As far as
diatoms are concerned, it is known that Antarctic species tend to be larger and
with more robust frustules than those from silicate-poor waters and can reach,
when in assemblages (very common in our trap material), sinking speeds of up to
more than 100 m day'1 [42,48].
In our model, vertical displacement is updated every day. Due to the distribution spectrum of particle intrinsic velocities and to total vertical displacement
variability, total vertical settling velocity considerably changes after Julian day
120, when the average vertical displacement starts to increase, reaching values
comparable with particle intrinsic downfall velocity.
A pseudoparticle daily counter is set numerically in our model at 230-m depth
and normalized to simulate the transmissometer measurements. A numerical
integrator on a time interval of 15 days is situated at 423 m, to simulate the trap
measurements. We ran models with different daily production rates of
pseudoparticles at 100 m, considering the hypothetical production of any kind of
particles, not just those originated by primary production. Two model outputs
seem particularly interesting to observe: the first assuming a constant production, the second simulating a seasonal cyclic production with an amplitude going
from 100 to 10%. The cycle is almost a square function going from 100% during
December through April and decreasing to 10% during June through October;
during May-November a linear interpolation occurs. It is évident that the pro
189
of particulate matter in the océan [39-41], by larger particles of biogenic origin.
Settling particles resulting from blooms of large dense cells or from aggregation
of small particles can constitute short-term puises rapidly reaching deep layers
and sédiments and scavenging fine minerai particles of the water column [42-45].
In our study, transmissometer and sédiment trap collection records generally
show a good corrélation, with the exception of the period May-August (i.e. from
120 to 240 Julian days). To try to understand the discrepancy between transmissometer and sédiment trap records in this period a simple Montecarlo model of
particle downfall was implemented.
Dunbar et al. [46] hâve found that deep currents are weaker in the investigated
area than in other zones of the Ross Sea continental shelf. A one-dimensional
model for the settling of biogenic material is thus likely to be successfully applicable there. According to the assumptions of Jaeger et al.[47], uninterrupted downward sinking of biogenic particles (produced by primary and secondary production processes mostly in the 0- to 100 m-layer) supplied from the surface is
assumed to originate at the bottom of the surface layer (100 m) and settle to the
seabed. The total downfall velocity is computed as the sum of intrinsic downfall
velocity of particles plus the actual vertical displacement of the water layer in
which they are contained. A set of pseudoparticles is numerically produced in our
simulation at 100-m depth, with a Gaussian spectrum of their intrinsic velocity.
The value of intrinsic downfall velocity is extracted randomly on the basis of a
gaussian shape with an average of 100 m d'1 and a variance of 20 m d 1 in order
to simulate a wider spectrum of real particle size and shape, consequently implying different downfall velocities. Particle intrinsic settling velocities are highly
variable. The chosen value of 90 ± 20 m day1 is in the range of the settling velocities of the particles mostly represented in our traps. For fecal pellets we refer to
measurements performed by Jaeger and collègues [47] in the study area. As far as
diatoms are concerned, it is known that Antarctic species tend to be larger and
with more robust frustules than those from silicate-poor waters and can reach,
when in assemblages (very common in our trap material), sinking speeds of up to
more than 100 m day'1 [42,48].
In our model, vertical displacement is updated every day. Due to the distribution spectrum of particle intrinsic velocities and to total vertical displacement
variability, total vertical settling velocity considerably changes after Julian day
120, when the average vertical displacement starts to increase, reaching values
comparable with particle intrinsic downfall velocity.
A pseudoparticle daily counter is set numerically in our model at 230-m depth
and normalized to simulate the transmissometer measurements. A numerical
integrator on a time interval of 15 days is situated at 423 m, to simulate the trap
measurements. We ran models with different daily production rates of
pseudoparticles at 100 m, considering the hypothetical production of any kind of
particles, not just those originated by primary production. Two model outputs
seem particularly interesting to observe: the first assuming a constant production, the second simulating a seasonal cyclic production with an amplitude going
from 100 to 10%. The cycle is almost a square function going from 100% during
December through April and decreasing to 10% during June through October;
during May-November a linear interpolation occurs. It is évident that the pro