Multicosm Research:
Basic Science To The Real World

If successful, Multi-scale Experimental Ecosystem Research Center Studies could revolutionize the way environmental policy is formulated, implemented and evaluated.

By Merrill Leffler

Fom the Chesapeake Bay to Puget Sound to the Seto Inland Sea in Japan, coastal waters have been hit with declining health in recent decades - but just how much of that decline is related to human activities remains uncertain. Experiments to answer such questions generally take place in the laboratory, and laboratory findings often have a complex and unclear connection to actual aquatic ecosystems.

In the Chesapeake, for example, countless factors affect the impact of nutrient enrichment or toxic chemicals, factors such as water circulation, turbulence, dissolved oxygen levels, biological and chemical interactions, in both sediments and water. Can realistic techniques be developed that will make it possible to extrapolate laboratory findings to the real world?

That is the aim of an ambitious, long-term research project at the University of Maryland Center for Environmental Science (UMCES). Tom Malone, director of the UMCES Horn Point Laboratory and head of the Multiscale Experimental Ecosystem Research Center - or MEERC - calls it "an incredible challenge."

"Its funding by the Environmental Protection Agency," says Wayne Bell, UMCES Vice-President for External Affairs, "is, to put it simply, visionary. If successful," he adds, "it could revolutionize the way environmental policy is formulated, implemented and evaluated."

The goal is no less than trying to develop a fundamental understanding of how aquatic ecosystems function. "Our hope," says Malone, "is that from controlled experiments and simulation models we will eventually make it possible to predict how such systems respond to any type of environmental stress, whether it be nutrient enrichment, toxic chemicals, heavy metals, sea level rise or temperature change."

Over the next five years, investigators will undertake increasingly complex experiments using various-sized experimental systems called "mesocosms." However, measuring the ecological changes in these different mesocosm environments as they respond to stresses such as nutrient enrichment is only the first step. Using simulation models to analyze empirical results represents a second step. The "daunting goal," though, as UMCES scientist Mike Kemp says, is to discover if there are relational or scaling rules among the different mesocosm experiments that will enable them to extrapolate their results to real world settings.

Lab Experiments and Environmental Regulations

Natural aquatic systems, no matter how small or how large, are too messy for conducting finely-tuned experiments. Biological and chemical conditions in the sediments and overlying waters, for example, are in continual flux, whether from natural cycles or the impacts of human beings. Mesocosms, on the other hand, can be controlled, thereby serving as surrogate ecosystems - whether small indoor tanks or large outdoor enclosures, they contain water, sediment and living organisms. For an experiment on the effects of nutrients or toxic compounds on phytoplankton production or fish mortality, the mesocosm can be dosed with known concentrations of each and a range of ecosystem effects measured.

Regulatory limits on contaminants have not been based on such sophisticated understanding of how ecosystems affects contaminants. Those limits are generally based on dose-response reactions - in plain words, how a species reacts to a toxic compound in a small, controlled space. Typical experiments, for example, might use a species of fish in an aquarium with controlled conditions of salinity, temperature and light. A measured concentration of the chemical in question is diluted in the water and the fish is then left on its own. How long does it take for the fish to die or to exhibit a "sublethal response," for instance, damage to its reproductive or endocrine system? It is from such laboratory observations that regulatory agencies set toxic discharge limits.

These observations, however, do not take into account the more complex reactions a contaminant undergoes in a real aquatic system or an ecologically complex mesocosm, reactions such as the interactive effects of the contaminant with sediment particles in the water or uptake by phytoplankton, or recycling by benthic (bottom-dwelling) organisms, let alone the effects of turbulence or the role of species higher in the food chain. Consequently, it is nearly impossible to extrapolate with predictive confidence lab findings to the real world ecosystem itself. And as a result, environmental regulations must often rely on research that is augmented by best guesses.

It is because of such experimental limitations that research labs have been using mesocosms to try to simulate conditions in aquatic systems. For some years, the Environmental Protection Agency (EPA) has supported the University of Rhode Island's Marine Ecosystem Research Laboratory (MERL) - large enclosed mesocosm tanks located on the shores of Narragansett Bay. Conditions in these tall cylinders have been designed to mimic the biology, chemistry and turbulence common to the Bay's sediments and waters.

The 14 MERL tanks make it possible to replicate experiments and to compare the effects of different contaminant concentrations. As surrogates for Narragansett Bay, MERL mesocosms have been fairly effective in giving reasonable results. They do, however, have limitations - they are expensive to operate and they are unique to the Narragansett Bay system. According to most researchers, applications of MERL results to other ecosystems have significant limitations.

It is such limitations that the MEERC project with its differently scaled mesocosms, or multicosms, is trying to overcome.

This tank is one of 40 mesocosms at the UMCES Horn Point Laboratory set up to mimic the dynamic conditons of a natural ecosystem.

The Multiscale Approach

"Some people originally thought that UMCES was going to do mini-Chesapeake Bays," says Karen Morehouse, Director of EPA Centers and Special Programs in the Office of Exploratory Research. There was some thought that these mesocosms were going to be exact microscale duplicates." That is not the case at all - the multicosm concept, she emphasizes, is unique. The long-term goal is to develop models that can be used to account for the complex interaction and feedbacks of aquatic systems, whether they be Chesapeake Bay or Puget Sound. To do this means first trying to determine if there are clear physical and mathematical relationships in laboratory mesocosm experiments.

One set of mesocosms is looking at the biological and chemical responses to nutrient loading in tanks containing sediments and overlying waters. Two other sets of mesocosms are looking at the biological response of marshes and submerged aquatic vegetation (SAV) to nutrient loading. In a few years, the sediment-water mesocosms and the marsh-SAV mesocosms will be linked together physically; for now, scientists are working to understand each of these systems separately.

In this first year, the sediment and overlying water (called pelagic-benthic) experiments have focused on three sizes of mesocosms and replicates: 0.1, 1.0 and 10 cubic meters. With some forty different tanks, the researchers are watching to see how these three sizes respond to different types of nutrient loading.

Our initial aim, says Michael Kemp, is in detailing how tank sizes and configurations affect biological responses to various types of nutrient loading; for instance, they are dosing tanks with high and low concentrations of continuous nutrient enrichment, as well as with high pulses of nutrient enrichment. Of the forty tanks, there are two separate configurations with three different volumes that have the same depth, and three different volumes that have the same ratio of diameter-to-depth, though they each differ in depth.

A Beginning

The data collected from one mesocosm would be large in itself, but the data from three different volumes, two configurations, and replicate experiments are downright massive. Though researchers are currently analyzing this first year's data, they have, says Mike Kemp, preliminary results on how planktonic, bottom-dwelling and wall growth (periphyton) communities in the different mesocosms responded to nutrient dosing.

While it is difficult to generalize, Kemp and his colleagues have been learning about the scale-dependency of mesocosm behavior.

• Under nutrient-rich water conditions associated with springtime, the rate of algal production to water surface area was similar for all mesocosm sizes; however, algal production to water volume was least in the deepest mesocosms. In contrast,
  
  
• Under relatively low nutrient conditions of summertime, algal production per water volume was constant for all mesocosms, so that photosynthesis per unit of water surface area (and light energy) was highest in the deep tanks.

Kemp says these findings suggest that results from mesocosm experiments should be scaled to real world conditions by depth in spring but by volume in summer. The implication is that seasonal variations in nutrient and light availability may change the appropriate scaling criteria for mesocosm experiments.

UMCES scientist Mike Kemp is heading a team of researchers who are trying to determine how mesocosm size and shape affects the results of an experiment.

While these studies on the scaled behavior of sediment and overlying water communities are one major elements of the MEERC project, two others are marsh mesocosm studies, headed by Court Stevenson, and SAV ecosystem studies, headed by Laura Murray. Their research is exploring how mesocosm complexity influences marsh and SAV productivity and nutrient retention.

During this first year, Stevenson and Murray have been asking questions that surprisingly, they say, have not been asked before. For example, how does ecosystem complexity affect responses to nutrient loading on marsh plants and on submerged aquatic vegetation? Does the ability of plant communities to absorb and transform nutrients vary with the form (dissolved vs. particulate) and the route of delivery, for example, groundwater vs. surface water? How do water circulation and the abundance of animals at higher trophic levels affect the responses of plant communities to nutrients and other climatic effects?

As a first-step simulation of marshland, Stevenson is employing large rectangular tanks of sediments planted with one or several species of grass. Nitrate-rich groundwater slowly seeps through the sediments, which are also on the receiving end of water that replicates tidal effects. The initial version of these experimental systems works effectively, maintaining healthy plant communities over long periods of time. "This year we will vary plant species composition," says Stevenson, "and examine response to groundwater nitrate under different tidal regimes."

"We're excited about these experiments," he says, "because nobody has ever replicated tidal marsh systems with groundwater inputs being varied while varying species composition. We believe," he adds, "we are carving out new ground."

Research that Is Truly Exploratory

"UMCES scientists are asking very basic questions," says Karen Morehouse, "questions that have to be asked. Our applied research is a house of cards if we don't have the answers." While nobody has done this kind of research, Morehouse points out, the political and scientific risks in such work are considerable. Such research requires long-term funding, meaning there are no quick answers - not an inconsequential issue in a climate in which politics often requires short-term results. Although Congress required EPA to set aside funding for exploratory research in 1979, "it is often difficult," she says, "to justify research support that is not immediately applicable." MEERC, like three other EPA-funded research centers, is eligible for continued funding over ten years.

"We may find that you cannot extrapolate anything meaningfully from the lab to the real world," says Morehouse, "though we know that if UMCES comes up with anything, it will be valuable. Unless we understand the basic way that aquatic ecosystems operate, then we have no business setting standards and regulations." This kind of research is essential, she believes, for EPA and other regulatory agencies.

"Understanding the system and predicting effects gives us a better chance of setting accurate regulations and defending them. After all," she adds jokingly, "EPA is famous for making indefensible decisions."

It is years too soon to know whether UMCES researchers can successfully develop models to characterize the wildly complex dynamics of these mesocosms and then discover if there are rules that enable them to extrapolate among differently-scaled systems. The answers may depend on increasingly sophisticated mathematics and, perhaps, luck - but that is another story.

Home • Search • Our Program • Chesapeake Bay
Research  Policy • Extension Program • Marine Education • Book Store & Videos • Hot Topics

This page was last modified May 24, 2002

http://www.mdsg.umd.edu/MarineNotes/Jan95/

Accessibility Statement