Spartina anglica is an autogenic ecosystem engineer. At a local (within-vegetation) scale, it improves plant growth by enhancing sediment accretion through attenuation of hydrodynamic energy with its shoots. This constitutes a short-range positive feedback. The vegetation also shows a long-distance negative feedback through formation of erosion troughs around the tussock, which are restricting lateral expansion. There is a growing recognition of the important role of such scale-dependent feedbacks for landscape structuring and ecosystem stability. By a series of flume studies, we provide direct experimental evidence that 1) both the local positive and the long-distance negative feedback are strongly densitydependent with clear thresholds, and 2) that both feedbacks are inherently linked via their density dependence, partly as the result of conservation laws. The observed thresholds occurred at vegetation densities commonly found in the field. We demonstrate that threshold densities are affected by external hydrodynamic forcing for the long-distance negative feedback. We expect the same to be true for the local positive feedback (sediment accretion), so that under mild hydrodynamic forcing, this local positive feedback may occur in the absence of a long-distance negative feedback (formation of erosion troughs). The density dependence of nearby positive and the long-distance negative feedback and how external conditions affect the linkage of both feedbacks are important factors shaping the vegetated salt-marsh landscape.

Polychaete tube lawns with high population densihes are frequent in marine soft-bottom environments. The influence of single tubes on near-bed flow dynamics has been q u t e well studied, but the critical population density that separates sediment destabilising effects from stabilising effects remains uncertain. This article presents results obtained with artificial tubes in a recirculating flume at a current velocity of 5 cm S-'. Four population densities were tested for their passive effects on the flow dynamics: expressed as percentage of the total surface area covered by tubes, they were 1.1, 2.0, 4.5 and 8.8 %. Using a high-resolution 3-dimensional current sensor, horizontal and vertical flow velocity profiles were recorded wthin the artificial tube lawns. An important deceleration of the current velocity was observed a t all population densities, ranging from 38.2 % at the lowest population density to 83.8 % at the highest. This deceleration, the shape of the vertical profiles, the calculated Reynolds stress values and the direct observation of sediment displacement led to the conclusion that the flow field is modified to gradually raise the effective level of the bottom towards the tube tips, resulting in skimming flow conditions at 8.8% surface coverage. Compared with field conditions, this is still a relatively low population density and thus means that many natural tube lawns have sediment stabilising effects, conditioning the substratum for further benthic succession.

Successful management and restoration of coastal vegetation requires a quantitative process-based understanding of thresholds hampering (re-)establishment of pioneer vegetation. We expected scouring to be important in explaining the disappearance of seedlings and/or small propagules of intertidal plant species, and therefore quantified the dependence of scouring on plant traits (flexibility, size) and physical forcing by current velocity. Flume studies with unidirectional flow revealed that scouring around seedlings increased exponentially with current velocity and according to a power relationship with plant size. Basal stem diameter rather than shoot length controlled scouring volume. Flexible shoots caused far less scouring than stiff shoots, provided that the bending occurred near the sediment surface as was the case for Zostera, and not on top of a solid tussock base as we observed for Puccinellia. Therefore, shoot stiffness is likely to strongly affect the chances for initial establishment in hydrodynamically exposed areas. Plant traits such as shoot stiffness are subject to a trade-off between advantages and disadvantages, the outcome of which depends on the physical settings.

Flume tanks are becoming increasingly important research tools in aquatic ecology, to link biological to hydrodynamical processes.There is no such thing as a ''standard flume tank'', and no flume tank is suitable for every type of research question. A series of experiments has been carried out to characterise and compare the hydrodynamic characteristics of 12 different flume tanks that are designed specifically for biological research. These facilities are part of the EU network BioFlow. The flumes could be divided into four basic design types: straight, racetrack, annular and field flumes. In each facility, two vertical velocity profiles were 2006 measured: one at 0.05 m s -1 and one at 0.25 m s -1 . In those flumes equipped with Acoustic Doppler Velocimeters (ADV), time series were also recorded for each velocity at two heights above the bottom: 0.05 m and 20% of the water depth. From these measurements turbulence characteristics, such as TKE and Reynolds stress, were derived, and autocorrelation spectra of the horizontal along-stream velocity component were plotted. The flume measurements were compared to two sets of velocity profiles measured in the field.Despite the fact that some flumes were relatively small, turbulence was fully developed in all channels. Straight and racetrack flumes generally produced boundary layers with a clearly definable logarithmic layer, similar to measurements in the field taken under steady flow conditions. The two annular flumes produced relatively thin boundary layers, presumably due to secondary flows developing in the curved channels. The profiles in the field flumes also differed considerably from the expected log profile. This may either have been due the construction of the flume, or due to unsteady conditions during measurement. Constraints imposed by the different flume designs on the suitability for different types of boundary layer research, as well as scaling issues are discussed.

Filter-feeding bivalves may have a pronounced grazing impact on the phytoplankton biomass in many shallow marine areas. The blue mussel Mytilus edulis, which lives in dense beds, can filter more than 100 m 3 m -2 d -1, but the grazing impact is highly influenced by hydrodynamic processes. Without externally generated currents or turbulent mixing, only a thin layer of near-bottom water would be subject to the down-mixing that causes an important supply of food to the mussels. Here, food-depleted jets of water expelled through the exhalant opening of a mussel may not only prevent the water, once filtered, from re-entering the animal, but the substantial speed of the jet may also help to mix the near-bottom water. The extent of such biological mixing -'biomixing' -caused by a dense population of M. edulis (200 cm long bed, filtering at 147 m 3 m -2 d -1) was studied experimentally at 2 flow speeds (4 and 8 cm s -1 ) in a laboratory flume channel at natural (low) algal concentrations in order to determine its relative importance compared to current-generated turbulence. Distributions of algal cells a distance of 162 cm from the start of the bed showed a near-bottom depletion of about 58 and 45%, respectively, for the 2 flow speeds, which indicate the degree of refiltration. In addition, flow structures were quantified in terms of distributions of velocity, turbulent shear stress and turbulent kinetic energy in the benthic boundary layer at 3 levels of mussel filtration-activity (maximal, reduced and zero). A description is given of this filtration activity of M. edulis with and without added suspensions of algal cells, which influence its valve-opening degree and filtration rate. It is concluded that biomixing enhances the flow-induced down-mixing of phytoplankton and can be identified as peaks in profiles of turbulent kinetic energy and turbulent shear stress. The associated increase in friction velocity over the length of the mussel bed at maximal filtration activity amounted to 56 and 49% for the 2 flow speeds studied. This shows the functionality of biomixing to be most helpful at low speed, where it is most needed due to the low levels of flow induced turbulence contributing to down-mixing of phytoplankton.