RESEARCH HIGHLIGHTS

Depth regulates concentrations of algal biomass and the balance between production and respiration for the integrated ecosystem. These relationships vary depending on the degree to which light and nutrient availability control primary production. This discovery not only affects how we conduct experiments and extrapolate their results to nature, it also suggests conditions under which trophic efficiency will be greatest.

Trophic efficiency (ratio of secondary production (e.g., fish growth) to primary production) changes with level of primary production in highly non-linear ways. Under conditions where fish populations are actively exploited, trophic efficiency will decline with eutrophication. Thus, while eutrophication tends to increase primary production in aquatic ecosystems, it may lead to decreased secondary production.

Experimental artifacts tend to decrease with width of container for experimental ecosystems. This is related to a variety of "wall-effects" that can also change the partitioning of ecological rates among major habitats (e.g., plankton vs benthos).

Light attenuation within the water columns of experimental ecosystems depend on the width of the containers, the optical characteristics of wall material and the growth of attached periphyton communities. We have developed quantitative descriptions of these relationships that will provide useful guidelines for future experimentalists.

Spatial variability that occurs in nature affects non-linear ecological interactions (essentially all interactions) in ways that cannot be predicted in conventional experiments using the "mean field approximation" of environmental conditions. This is particularly important for systems dominated by mobile consumers (e.g., fish) that can contribute to further spatial heterogeneity. We have developed spatial modeling codes, in collaboration with Dave Scheurer and Bob Gardner, which provide a mechanism for extrapolating results from experiments conducted under mean conditions to natural spatially varying ecosystems.

Relative variance among replicate experimental ecosystems tends to decrease with size of containers. Relative variance within experimental ecosystems tends to increase with container size. Recent analyses of data collected in P/B mesocosms supports this hypothesized pattern.

Growth of fish is strongly scale-dependent. Planktivorous fish (bay anchovy and Atlantic silversides) grow faster in 10 m³ than in 1 m³ mesocosms. Growth rates of the fish in the 10 m³ mesocosms are essentially equal to growth rates of these species in Chesapeake Bay in experiments of 10 to 25-day duration.

Fish grow faster in the most pelagic mesocosms (highest volume to wall area ratio). The 10 m³ E mesocosms (relatively wide, shallow) supported the highest growth rates. This result suggests that walls and boundaries reduce foraging efficiency of enclosed pelagic fish.

Fish predation exhibited a consistent top-down control on zooplankton communities in experiments. Effects generally are expressed more in 10 m³ than in 1 m³ mesocosms, indicating that predator consumption is scale-dependent and more effective and efficient in the larger mesocosms.

Evidence for trophic cascades was observed, although it tended to be variable in the MEERC predation experiments. In general, mesocosms with fish predators not only had lower zooplankton abundances, but zooplankton diversity tended to be lower and sizes were smaller. Experimental treatments that included predators tended to support higher phytoplankton biomasses. The cascade progressed to the nutrient level in some MEERC experiments in which ammonia levels tended to be low in high-predation treatments.

A simulation model indicated that foraging efficiency of planktivorous fish declined as enclosure size was reduced. When combined with results from bioenergetics modeling, small reductions (~10%) in foraging efficiency were demonstrated to lead to large declines (~50%) in fish growth. The models suggest that enclosures >2 m diameter are required to support growth and efficient foraging of 50-mm planktivorous fish at levels near those in natural ecosystems. Ongoing modeling promises to extend our knowledge of mechanistic and behavioral constraints on scale-dependent foraging by planktivorous fish.

Vertebrate (fish) and invertebrate (lobate ctenophore) predation both produce strong top-down control over zooplankton communities in enclosed pelagic ecosystems. Although difficult to equate, fish exercise stronger effects than does the ctenophore. Trophic cascades, indicated by effects transmitted to primary producers, also can be attributed to both predators.

Average residence time for water in experimental and natural ecosystems tends to regulate the competitive balance between algae and submersed vascular plants for nutrients and light. Water residence time (i.e., the inverse of exchange rate) also regulates the balance between phytoplankton and zooplankton. Hence, because exchange rates are easily controlled in experimental ecosystem designs, care should be taken to match these with those in the natural system of interest.

Light requirements for survival of seagrasses and related submersed aquatic plants are critically related to water depth and nutrient levels. Models developed in MEERC have been used to relate submersed habitat suitability to water quality conditions as measured in routine monitoring programs. This is an important step in planning for restoration of these submersed plants.

In the past, turbulent mixing in experimental ecosystem studies has been too often ignored, inadequately characterized, or unreported. In our studies, design principles from the chemical engineering literature were applied; laboratory scale models were built and tested; full scale prototypes were built, tested, and characterized; turbulence scaling relationships were tested and shown to hold in both experimental systems and natural systems; a thorough review of the influences of turbulence in ecosystem dynamics was produced; field data on natural turbulence in Chesapeake Bay was collected and analyzed; physical controls of boundary fluxes were investigated; students were instructed on principles of scaling, dimensional analysis, and turbulent mixing; and an improved mixing device was developed that produces both realistic benthic boundary layer shear stresses and water column turbulence in a single tank. We have learned a great deal about turbulence in natural aquatic ecosystems during our MEERC studies. As a result of our work, other ecosystem scientists will be better informed about specific influences of turbulence and mixing, and they will be able to avoid many of the mistakes of past studies were physical mixing was not appropriately addressed.

"Wall effects" are the reality of mesocosm contaminant studies, so any endeavor using mesocosms must recognize this and design the experiments and sample appropriately, i.e., quantifying periphyton biomass and periphyton contaminant concentrations. Mesocosm design can be altered to minimize these effects. This is especially true for particle-reactive contaminants. Adsorption to periphyton or biofilms on walls or solid surfaces is an important process that must be quantified and the importance of periphyton or biofilms in the natural environment, even in oligotrophic lakes, in contaminant fate is now being realized.

Other artifacts of using mesocosms for contaminant study relate to the particular chemical under study. Clearly, air-water exchange processes need to be considered for volatile and semi-volatile contaminants because the dimensions of mesocosms (surface area/depth ratio) do not necessarily mimic reality. Also, mesocosm shape will effect photochemical processes. Further, the fact that sediment is typically only at the bottom of mesocosms, whereas natural environments can have sediment "walls", must be considered for those contaminants whose fate is strongly related to processes at the sediment-water interface, e.g., mercury methylation. Clearly, the ratio of sediment surface to water volume is an important consideration.

"Biomass dilution" of contaminant concentration under eutrophic conditions was demonstrated in the mesocosm studies and confirmed by the modeling studies. However, the effect on piscivorous fish concentration was less than expected because of "trophic dampening" - at higher trophic levels, various feedback mechanisms will lead to a decrease in the effect of eutrophication on bioaccumulation.

Bioaccumulation in benthic food chains is much less influenced by water column productivity. Because the benthos is an important source of contaminants to "pelagic" species in shallow ecosystems (consider, for example, the current discussions over striped bass consumption of juvenile blue crabs), this result suggests that benthic-pelagic coupling must be considered in any studies of contaminant cycling in productive and physically dynamic, shallow ecosystems such as estuaries.

Resuspension will have a measurable effect on contaminant cycling and bioavailability, especially for those contaminants whose fate is related to factors such as sediment redox status and sediment microbial activity. The kinetics of contaminant desorption is an important mechanism influencing the route of exposure of organisms in the water column to particle-reactive elements in shallow, energetic ecosystems. The importance of "bio-transfer" of contaminants from sediment to the water column via direct feeding, or by migration of invertebrates such as amphipods from the sediment surface to the water column has been recognized as being an important mechanism for the transfer of contaminants to higher trophic organisms and may be as important as other mechanisms (diffusion and physical advection).

Sediment chemistry and the matrix in which the contaminant resides are important mediators of contaminant bioavailability. The relative importance of these factors is such that they typically overwhelm the magnitude of the changes in total contaminant concentration in a particular environment. Thus, total contaminant load in sediment is a very poor predictor of bioaccumulation potential.

For most metals, food or sediment represent the most important accumulation route. Furthermore, the process of solubilization during benthic invertebrate digestion is a very important determinant of bioaccumulation. Indeed, the factors controlling solubilization are more important than the factors influencing transport across the gut membrane in determining the bioaccumulation of Hg, MMHg and other metals.

The integration of experimental methods, empirical studies, and modeling is required for extrapolation of results across temporal and spatial scales. We have integrated MEERC mesocosm results with measurements in estuarine systems by: a) analysis of temporally and spatially extensive data from the Chesapeake Bay; b) development of a spatially explicit model of ecosystem dynamics for pelagic systems; and c) estimation of model parameters from experimental systems and empirical observations.

Results of the analysis of chlorophyll, salinity and temperature data obtained from TIES (Trophic Interactions in Estuarine System - investigates mechanisms controlling secondary production in estuarine ecosystems) sampling within Chesapeake Bay, performed with both spectral and fractal analysis, have shown that:

• Spatial patterns can be described by power law and fractal relationships. Although the confines of the Chesapeake Bay were hypothesized to induce scale-dependent effects, results are near the b = 2.0 predicted from theory for scales of 200m to 15 km.
• Small but potentially significant changes in scaling pattern can be detected between years, seasons, and Bay sections.
• These results allow fractal methods to be used to extrapolate information across scales and, by statistical inference, the estimation of variables not directly measured by TIES transects (e.g., spatial variation in DIN).
• The spatial scaling relationships defined by these analyses have important implications for sampling protocols to characterize the mean and variation of ecological processes in the Chesapeake Bay. Optimal placement of limited sampling stations vary depending on the spatial structure (as defined by fractal analysis) of pelagic interactions.

A 2-D pelagic model has been developed to simulate ecosystem dynamics over a range of scales from mesocosms (grain = 1m) to the Chesapeake Bay (extent = 20 km). Physical processes considered include diffusion, advection, and turbulent mixing. Biological processes incorporated in the model include nutrient dynamics, trophic interactions, and top down predation (fish) effects. The formulation of this model allows explicit consideration of the spatial variation in ecosystem processes on predicted pelagic dynamics. The effects of spatial heterogeneity in forcing functions are now being examined by varying the frequency, duration, intensity, and spatial pattern of these inputs. Results to date show:

• Spatial sensitivity analysis is a promising tool to identify the scale-dependent effects of spatial variation of physical and biological processes. Although patchiness of chlorophyll concentrations is most affected by hydrodynamic effects (e.g. fronts, convergence zones), fine-scale variation in biological dynamics (at scales < 200m) may have significant effects on predicted patterns of pelagic systems.
• The analysis of simulation results by spectral and fractal techniques allows chlorophyll distributions, and subsequent trophic interactions, to be compared with the analysis of TIES data.

Return to top of this document.

The Center at a Glance || Facilities || Admin. & Councils || What's New ! || Research Projects
Research Highlights || Participants || Pubs & Reports || Workshops || Related Links || Models in MEERC