Field and Numerical Studies of Near-Bed Aggregate Dynamics
203
distribution of aggregates in the BBL may provide information to interpret the
vertical distribution patterns of epibenthic vagile animals (Brandt 1996; Childress et al. 1989; Angel 1990). These organisms would be predicted to live and
feed at heights where their preferred aggregate size spectrum exists or accumulates.
In the presented experiments the different physicochemical properties of the
aggregates sizes were not taken into account. For simplicity and in contrast to
the available knowledge (e.g. Alldredge and Mc Gillivary 1991) the stickiness, a
factor influencing the probability of aggregate formation, was set to unity. Especially the effect of various different aggregate sources may have strong effects on
the stickiness and consequently on the near-bed aggregate dynamics. Further
on, resuspension of fine-grained clay particles will influence the aggregation behaviour of the suspension (Muschenheim et al. 1990). Likewise, increasing the
relative amount of freshly settled, most likely senescent phytoplankton cells
(Riebesell1992) will result in the abundance of transparent exopolymere particles (TEP; Alldredge et al. 1993; Passow et al. 1996) close to the seafloor. Their
presence will also alter the aggregation but also disaggregation rates in the hydrodynamically energetic BBL. Thus, experiments with varying size-dependent
stickiness values would simulate aggregate dynamics more realistically.
13
Conclusions
The model potentially improves the understanding of aggregate dynamics close
to the seafloor. In contrast to the layer-averaging model of Hill and Nowell
(1995), the presented results suggest that aggregate interaction influences both
concentration profiles of the three investigated aggregate size classes and their
overall size distribution close to the seafloor. Thus, it supports the idea of hydrodynamical sorting, a process which cannot be resolved by measuring bulk parameters of aggregate composition alone. Hydrodynamical sorting was observed in a more tranquil hydrodynamic regime 1 (u lOO = 0.1 m s-1). In the more
vigorous hydrodynamic regime 2 (u lOO = 0.5 m S-1), a c-shaped distribution was
obtained near the bed, which was more pronounced when aggregate interaction
was included. Aggregate interaction strongly influenced the concentration distribution of different aggregate size classes. Especially the smallest aggregate
class was strongly reduced when aggregates were allowed to interact. However,
when aggregate interaction is taken into account, both aggregate formation
(turbulent aggregation and differential settling) and disaggregation have to be
calculated. Without disaggregation, the aggregate distribution was unrealistically dominated by the largest aggregate class.
The major goal was to introduce a new tool to investigate the near-bed aggregate dynamics using a simple diffusion advection approach including aggregate
interactions. However, further experiments should focus on lower current velocities and varying roughness heights. The presented results suggest that the model has to be extended to more interacting aggregate size classes. Future experi-
203
distribution of aggregates in the BBL may provide information to interpret the
vertical distribution patterns of epibenthic vagile animals (Brandt 1996; Childress et al. 1989; Angel 1990). These organisms would be predicted to live and
feed at heights where their preferred aggregate size spectrum exists or accumulates.
In the presented experiments the different physicochemical properties of the
aggregates sizes were not taken into account. For simplicity and in contrast to
the available knowledge (e.g. Alldredge and Mc Gillivary 1991) the stickiness, a
factor influencing the probability of aggregate formation, was set to unity. Especially the effect of various different aggregate sources may have strong effects on
the stickiness and consequently on the near-bed aggregate dynamics. Further
on, resuspension of fine-grained clay particles will influence the aggregation behaviour of the suspension (Muschenheim et al. 1990). Likewise, increasing the
relative amount of freshly settled, most likely senescent phytoplankton cells
(Riebesell1992) will result in the abundance of transparent exopolymere particles (TEP; Alldredge et al. 1993; Passow et al. 1996) close to the seafloor. Their
presence will also alter the aggregation but also disaggregation rates in the hydrodynamically energetic BBL. Thus, experiments with varying size-dependent
stickiness values would simulate aggregate dynamics more realistically.
13
Conclusions
The model potentially improves the understanding of aggregate dynamics close
to the seafloor. In contrast to the layer-averaging model of Hill and Nowell
(1995), the presented results suggest that aggregate interaction influences both
concentration profiles of the three investigated aggregate size classes and their
overall size distribution close to the seafloor. Thus, it supports the idea of hydrodynamical sorting, a process which cannot be resolved by measuring bulk parameters of aggregate composition alone. Hydrodynamical sorting was observed in a more tranquil hydrodynamic regime 1 (u lOO = 0.1 m s-1). In the more
vigorous hydrodynamic regime 2 (u lOO = 0.5 m S-1), a c-shaped distribution was
obtained near the bed, which was more pronounced when aggregate interaction
was included. Aggregate interaction strongly influenced the concentration distribution of different aggregate size classes. Especially the smallest aggregate
class was strongly reduced when aggregates were allowed to interact. However,
when aggregate interaction is taken into account, both aggregate formation
(turbulent aggregation and differential settling) and disaggregation have to be
calculated. Without disaggregation, the aggregate distribution was unrealistically dominated by the largest aggregate class.
The major goal was to introduce a new tool to investigate the near-bed aggregate dynamics using a simple diffusion advection approach including aggregate
interactions. However, further experiments should focus on lower current velocities and varying roughness heights. The presented results suggest that the model has to be extended to more interacting aggregate size classes. Future experi-