Design Factors

Sourcing, Seeding, and Energy Matching

Where a microcosm is designed to represent a model of a system that already exists, it should contain all of the characteristic features of the ecosystem that are necessary in the context of the problem to be described or solved. It is usually recommended that components used in the microcosm (e.g., soil, water, plants, animals) are sourced from the natural ecosystem. It is also important to try and expose the microcosm to the same physical, chemical, and biological inputs or 'energies' (e.g., light, temperature, nutrients, turbulence, species immigration) as the natural ecosystem. This is termed the energy signature approach to microcosm design. It can be more difficult to match some inputs for laboratory microcosms. For example, artificial light is a poor substitute for natural light. Constructing microcosms by isolating parts of the natural ecosystem in situ can minimize disturbance and enables matching of light and temperature inputs; however, other energies such as turbulence may not be equivalent (Figure 2).

In synthesized, gnotobiotic microcosms, the researcher has the challenging role of system organiser while self-organization is prevalent in derived microcosms. A multiple seeding approach, where innocula from several natural assemblages are mixed together and left to self-organize, is a technique recommended by H. T. Odum to develop a more stable and sustainable microcosm system. Reinoculation can sometimes be necessary to maintain important species that do not develop sustainable populations.

Figure2 J. M. Quinn (National Institute of Water & Atmospheric Research) stands beside recirculating stream microcosms that have been used to investigate nutrient uptake and transfer in stream food webs and the effects of lighting and fine sediment deposition on stream biogeochemistry (Parkyn etal.).

Figure2 J. M. Quinn (National Institute of Water & Atmospheric Research) stands beside recirculating stream microcosms that have been used to investigate nutrient uptake and transfer in stream food webs and the effects of lighting and fine sediment deposition on stream biogeochemistry (Parkyn etal.).

Spatial Scaling, Wall, and Isolation Effects

Temporal Scaling

Microcosm size affects the amount of diversity that the system can accommodate, with larger microcosms being able to support a greater diversity and more trophic levels, than smaller ones. Microcosm shape also has the potential to strongly impact on microcosm functioning and it can be useful to incorporate testing of microcosm size and/or shape into experimental design (Figure 3). In particular, microcosm designs with a large wall surface area to volume ratio should be used with caution. The metabolic activity of microbial or periphyton biofilms ('edge communities') attached to these walls can be substantial and highly unrepresentative of natural conditions. To avoid these effects, larger microcosm volumes in relation to wall surface area are recommended. The composition of microcosm walls should also be considered. Wall materials should be inert and not leach or absorb substances that may affect the experiment. Gases such as oxygen can diffuse through more flexible plastics, which may or may not be desirable depending on the ecosystem being modeled. Consideration should also be given to the effects of artificial isolation, which restricts the movement of mobile organisms. The small size of microcosms also typically excludes higher trophic levels. However, the activities of some higher organisms or mobile species (e.g., grazing of vegetation, removal or replacement of individuals of a species by a predator or migration) may be simulated by human actions.

Figure 3 Design of an experiment testing the effect of microcosm size and the pesticide chlorpyrifos on macroinvertebrate colonization of estuarine sediments. Average taxa richness was significantly higher in larger microcosms although average animal density was higher in smaller microcosms. In both large and small microcosms, animal density was significantly higher near the perimeter, indicating an 'edge' effect. Reproduced from Flemer DA, Ruth BF, Bundrick CM, and Moore JC (1997) Laboratory effects of microcosm size and the pesticide chlorpyrifos on benthic macroinvertebrate colonization of soft estuarine sediments. Marine Environmental Research 43: 243-263.

Figure 3 Design of an experiment testing the effect of microcosm size and the pesticide chlorpyrifos on macroinvertebrate colonization of estuarine sediments. Average taxa richness was significantly higher in larger microcosms although average animal density was higher in smaller microcosms. In both large and small microcosms, animal density was significantly higher near the perimeter, indicating an 'edge' effect. Reproduced from Flemer DA, Ruth BF, Bundrick CM, and Moore JC (1997) Laboratory effects of microcosm size and the pesticide chlorpyrifos on benthic macroinvertebrate colonization of soft estuarine sediments. Marine Environmental Research 43: 243-263.

A critical consideration for microcosm studies is the duration of experiments. Most microcosm experiments are generally conducted over a period of only weeks to months. However, the duration of microcosm experiments needs to be sufficient to assess effects on slow-responding organisms or processes.

As with natural ecosystems, conditions within microcosms can change over time and these changes should be evaluated during the course of a microcosm study. As the duration of a microcosm experiment increases, so does the likelihood of greater variability developing between replicates as a result of natural divergence. Time series sampling can be incorporated into experimental design to monitor changes. However, careful consideration of the impact of any repetitive sampling of components is required. A large number of microcosm replicates can be established at the outset of an experiment to enable complete (i.e., destructive) sampling of a subset of replicates at designated time intervals during the course of the study (Figure 4).

Natural ecosystems are also subject to diurnal and seasonal variations as a result of light, temperature, and other climatic effects. In the laboratory, natural diurnal variations can be simulated to some extent by the use of controlled light-dark cycling of artificial lights (Figure 5). For experiments of short duration, seasonal variability may be taken into account by repeating experiments on a seasonal basis.

Studies of longer-term ecological processes such as succession, predator-prey cycles, extinction, can be studied in microcosms with short real-time duration using organisms with very short generation times (e.g., microorganisms). This has been termed the biological accelerator approach.

Figure 4 Simple aquatic microcosms (41 pails in a 1 m depth flow-through freshwater tank) used to investigate the effects of different sediment types on the growth responses of selected submersed macrophyte species (Matheson etal.).
Figure 5 Riparian wetland soil microcosms set up in a climate-controlled laboratory and used to investigate the fate of nitrate and the effect of wetland plant growth on nitrogen transformation processes (Matheson etal.).

Replication, Variability, and Divergence

Microcosm replication is an experimental design issue. The more complex the system to be studied, the more replicates are generally required to adequately describe and account for the associated variability in test results. Even when they are started similarly, microcosms often develop differently, particularly over longer periods of time. While this divergence can be problematic for microcosm replicability, this phenomenon does offer opportunities to test ecological theories and models about how different community structures can develop with, for example, different sequences of seeding (e.g., chaos and assembly theories, lottery and random models). The cross-seeding technique, where some of the contents of one replicate are regularly transferred to another, can be a useful strategy to reduce variability and divergence among replicates. This is often done in the initial, setup stage but may also be incorporated into the experimentation period.

Sufficient replication in scientific experiments is critical to enable robust statistical evaluation of results. Analysis of variance (ANOVA) experimental designs are most commonly used in microcosm studies. Regression designs are sometimes employed which may enable testing of a broader range of treatments but results from these are more robust if some replication of treatments is also included. Replication enables a mean value for test results to be calculated along with the standard deviation and error of the mean. Three replicates is a recommended minimum for scientific investigations but higher numbers of replicates will provide more robust results. Power analysis and sample size estimation can be useful statistical techniques to employ to ensure that there is sufficient replication.

F¡eld Greenhouse microcosm

F¡eld Greenhouse microcosm

Figure 6 Ilustration of the different components examined in a study comparing the community structure of methane-oxidizing bacteria in microcosms to the natural ecosystem. In both systems, the main factors controlling the population size and activity of methane-oxidizing bacteria were plant growth and availability of nitrogen. Community diversity, activity patterns, and the population structure in both systems were comparable, although different quantities were detected. Reproduced from Eller G, Kruger M, and Frenzel P (2005) Comparing field and microcosm experiments: A case study on methano- and methylo-trophic bacteria in paddy soil. FEMS Microbiology Ecology 51: 279-291.

Figure 6 Ilustration of the different components examined in a study comparing the community structure of methane-oxidizing bacteria in microcosms to the natural ecosystem. In both systems, the main factors controlling the population size and activity of methane-oxidizing bacteria were plant growth and availability of nitrogen. Community diversity, activity patterns, and the population structure in both systems were comparable, although different quantities were detected. Reproduced from Eller G, Kruger M, and Frenzel P (2005) Comparing field and microcosm experiments: A case study on methano- and methylo-trophic bacteria in paddy soil. FEMS Microbiology Ecology 51: 279-291.

Similarity to Natural Ecosystem

Designing a microcosm or any model of a natural ecosystem (macrocosm) is a test in itself of how much is known about the ecosystem. A study using microcosms should include measurement of the characteristic biotic and abiotic features and functions of the natural ecosystem it is trying to model. Selection of features and functions to measure should be based on how critical they are to the natural ecosystem. These measurements enable an assessment to be made of how closely the microcosm represents the natural ecosystem. Any extrapolation of results from microcosm studies to natural ecosystems should be based on sound evidence of close matching of key features and functions between these systems (Figure 6). However, ideally, results from microcosm studies should be confirmed with further field scale testing.

Project Earth Conservation

Project Earth Conservation

Get All The Support And Guidance You Need To Be A Success At Helping Save The Earth. This Book Is One Of The Most Valuable Resources In The World When It Comes To How To Recycle to Create a Better Future for Our Children.

Get My Free Ebook


Post a comment