Limnol. Oceanogr., 44(5), 1999, 1331–1340

Several characteristics inherent to experimental ecosystems were examined for their influence on ecological processes in five cylindrical indoor benthic–pelagic enclosures of different size and shape. Ecosystem development diverged significantly among the mesocosm dimensions even though environmental parameters such as surface photosynthetically available radiation (400–700 nm), turbulence intensity, water exchange rate, and nutrient input were held constant across systems. Here we show that factors that lead to the development of different plankton assemblages can be related to mesocosm geometry. The ratio of light-receiving surface area-to-water column volume (As : V) was shown to control both the rate of NO consumption and gross primary productivity. The attenuation 3 of water column irradiance was positively correlated with the wall area-to-volume ratio (Aw : V). This was manifested in greater light attenuation in systems with a high Aw : V ratio. In addition, notably greater microalgal biomass developed on the walls in systems with a high Aw : V ratio. Finally, the total surface area-to-volume ratio (At : V) of the mesocosms influenced the rate of energy gain and dissipation and water column temperature. The differences in temperature among dimensions possibly affected biological parameters such as bacterial biomass. Although the influence of area-to-volume effects on biological components in artificial systems may be substantial, our analysis indicates that some of these effects may be predictable and that future experiments can be explicitly designed to minimize artifacts of enclosure. Experimental ecosystems (mesocosms) are often used to examine ecological interactions and they allow investigators to observe one discrete body of water for a sufficient length of time to characterize nutrient fluxes and trophic interactions (Lalli 1990; Oviatt 1994; Glibert 1998). Mesocosms are particularly useful in studies of energy and material transfer from one trophic level to another, in studies of interactions between plankton and benthic communities, and in studies of chemical or biogeochemical transformations of nutrients or pollutants (cf. Takahashi et al. 1975; Parsons et al. 1978; Elmgren et al. 1980; Grassle and Grassle 1984; Oviatt et al. 1987; Doering et al. 1989; Egge and Aksnes 1992; Heiskanen et al. 1996). Nevertheless, it is well recognized that artifacts and constraints of experimental enclosures pose limitations on how readily results can be extrapolated from artificial to natural systems (Pilson and Nixon 1980; Brockmann 1990; Oviatt 1994). Experimental systems are inherently limited in their ability to reproduce nature accurately through omission of 1 To whom correspondence should be addressed. Present address: Department of Marine Chemistry and Geochemistry, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543. 2 Present address: National Center for Ocean Research, P.O. Box 23-13, Institute of Oceanography, National Taiwan University, Taipei, Taiwan 10617, R.O.C. Acknowledgments This work was supported by a grant from the U.S. Environmental Protection Agency, Centers for Exploratory Research Program (grant 819640). We thank Elgin Perry, Jeff Cornwell, Tim Goertmiller, Laurie VanHeukelem, Debbie Hinkle, Chris Madden, Steve Suttles, Tom Wazniak, and Alison Sanford for assisting with sampling and analyses. We thank Larry Sanford, Alison Sanford, Mike Kemp, John Petersen, and Mike Roman for sharing unpublished data, and John Petersen, Norm Nelson, and Todd Kana for helpful discussion of the manuscript. This is contribution 3185 from the University of Maryland Center for Environmental Science. higher trophic levels or water column structure (Carpenter 1996). In addition, ecosystem couplings can be influenced by two types of enclosure effects, fundamental scaling effects and artifacts of enclosure (Petersen et al. 1997). Fundamental scaling effects can be attributed to characteristics such as depth that are common to all ecosystems, whereas artifacts of enclosure are characteristics of experimental systems that differ from natural ecosystems (Petersen et al. 1997). A commonly used criterion for the design of mesocosm systems is that some variables, e.g., the sediment area-tovolume ratio or water exchange rate, should be similar to the natural system to which data will be applied (i.e., Nixon et al. 1980; Oviatt et al. 1987). However, regardless of scaling similarities between enclosed systems and their natural counterparts, biological parameters in enclosures very often behave differently than the same parameters in natural systems (French and Watts 1989; Oviatt et al. 1980, 1989; Axler and Reuter 1996). In enclosures of differing physical scales, food web dynamics may be affected to varying degrees, creating trophic interactions that differ with dimension and differ from those of a natural system (Kuiper et al. 1983; Stephenson et al. 1984). A characteristic artifact that can produce such imbalances includes the large wall area-to-volume ratio of many enclosures; the growth of material on walls has been shown frequently to dominate rate processes of the pelagic phytoplankton (Kroer and Coffin 1992; Oviatt 1994; Chen et al. 1997). In an effort to understand whether results from experimental systems of very different physical proportions can be compared when treatments are similar, a study was undertaken to quantify variability in biological parameters resulting from differences in physical scale and associated enclosure artifacts. Environmental variables such as room temperature, photosynthetically available radiation (PAR), and turbulence intensity were held constant, whereas physical dimensions such as radius and depth of the enclo-