RESEARCH PROJECTS

MEERC Research Projects 1998-2001

Scale-Dependent Relations Governing Extrapolation from Mesocosm to Nature in Coastal Ecosystems

Victor S. Kennedy, Center Director
W. Michael Kemp, Research Coordinator

Background

References to 'scaling' ideas in ecology and oceanography have increased steadily over the last fifteen years, with numerous examples of 'scale-dependent' patterns being reported. There is, in fact, a growing awareness among ecologists that most quantitative observations made in the field will vary with the scale of the measurement (e.g., Schneider 1994, Giller et al. 1994). Estimates of mean and variance of most ecological properties will differ depending on whether the observations are made at 1 or 100 stations, and whether they are made at monthly or hourly intervals. By increasing the resolution (e.g., more frequent observations) and extent (e.g., over larger spatial scales) of observations, estimates of "reality" may be improved, but finite availability of time and resources limit the scales of measurement. Some properties may be distributed in fractal patterns, so that there is no exponential approach to reality. Field ecologists frequently talk about 'matching the scales of the observations to the scales of the questions.' While this is a reasonable goal, there are few principles available to guide this search for a match.

Integrated Themes of Our Workplans

In our proposed research, we have been exploring the possible relationships between scale-dependent ecological patterns in time and space as well as scaling relations that may guide extrapolation of results from controlled experiment to natural ecosystems. During the first phase of MEERC, we examined how changes in the scales of time, space, and complexity of experimental systems influenced ecological dynamics and responses to nutrient and contaminant perturbations. Empirical and modeling studies produced several important findings, including: 1) predictive relations wherein ecosystem properties scale with water depth and container width; 2) temporal scaling relations for both input frequency and residence time of perturbants; 3) preliminary complexity scaling relations demonstrating how trophic structure modulates system responses to nutrient inputs. However, we concluded that this empirical approach was too limited in its ability to detect and explain fundamental scaling relations in aquatic ecosystems. Hence in 1997 we began the second phase of this effort with a plan to lay a theoretical foundation for the continuing sequence of experiment and modeling studies.

We hypothesize that fundamental ecological scaling relations arise from two basic sources: 1) dimensional matching of critical time and length scales for biological-physical interactions, and 2) spatial heterogeneity in distributions of key ecosystem properties and processes. It is these sources of scale-dependence that produce discrepancies between controlled experimental and natural conditions, as well as differences between ecological dynamics in large and small habitats. In the first case we propose that experiments can be designed to retain the dimensional similitude occurring in nature between the important scales of biological interaction (e.g., predator-prey) or scales of physical-biological coupling (e.g., residence-time for biological processes). Several studies in the first phase of MEERC have taken this approach, with experiments examining: 1) depth-scaling of ecosystem photosynthesis; 2) width-scaling of periphyton growth; 3) scaling water residence-time and associated effects on zooplankton grazing and on nutrient-SAV interactions; 4) scaling input frequency effects on SAV-nutrient interactions; and 5) scaling boundary-layer constraints on benthic filter-feeding. In the second phase of MEERC, we have continued research that considers dimensional scaling of physical-biological coupling, but a major focus has been on the question of how spatial heterogeneity produces scale-dependent patterns and dynamics in nature that are not typically reproduced in mesocosms. Here we postulated that many of the reported scale-dependent dynamics seen in aquatic ecosystems emerge from characteristic exchanges of material, energy, and motile organisms across space.

Summary of Workplan Since 1998

Modeling and synthesis included development of a 3-dimensional spatial simulation program that accommodates modeling of multiple habitats (marsh, SAV, plankton, benthos) with flexible spatial orientations and gridding densities. From the large spatial grid, a single ecosystem volume can be isolated and simulated apart from the surrounding spatial array. This isolated volume can be used to simulate dynamics of experimental mesocosms with limited (or regulated) exchanges. Simulated responses of such isolated-volume ecosystems (the "simulated mesocosm") to treatments can be compared to those of the integrated spatial model. This produces a suite of 'simulated comparisons' of responses in 'mesocosms' versus 'nature.' We believe that these kinds of modeling experiments will move us closer to a fundamental understanding of scaling relations for extrapolation from mesocosm experiment to estuarine ecosystem.

The experiments, modeling studies, and syntheses in Projects 1 through 3 clearly complement this approach. In these studies, effects of scaling exchanges of water, material, and organisms between adjacent habitats were examined, as were spatially-scaled top-down effects of fish movement and top predator predation on processes at lower trophic levels. We linked mesocosms in multicosm arrays. We examined how spatial patterns of marsh plant diversity and distribution (and temporal scaling of water-movement) affect ability of intertidal ecosystems to use and transform groundwater-delivered chemicals. Contaminant studies (Project 4) have examined how variations in trophic structure influence mobility and accumulation of toxic chemicals. This effort also depends on analysis of spatial and temporal patterns in data from natural estuarine ecosystems. While we continued to examine scale-dependent patterns in nature and their implications for extrapolation from mesocosms, we also conducted a range of modeling experiments to test the dependence of model behaviors on their inherent time, space, and complexity scales (Project 5).

Project 1: Scaling Trophic Interactions in Pelagic Estuarine Ecosystems: The Role of Predators in Mesocosms and Nature

Principal Investigators: E Houde, W M Kemp, T Miller, J Purcell, M Roman
Visiting Scholars: J Petersen
Graduate Students: B Muffley, X Ma
Faculty Research Assistants: D Hinkle

Project 2: Multi-Habitat Studies of Ecosystem Interactions (Multicosms)

Principal Investigators: L Sanford, W M Kemp, L Murray
Faculty Research Assistants: S Suttles
Graduate Student: R Bartleson, J. Melton, K. Schulte

Project 3: Scale-Dependent Behavior in Marsh Mesocosms

Principal Investigators: J C Stevenson, J C Cornwell
Visiting Scholars: P Kangas, C Madden, J Stribling
Faculty Research Assistant: D Hinkle
Graduate Students: J Dewar, T Coley, S. Swartwood

Project 4: Ecotoxicology and Issues of Scale

Principal Investigators: R P Mason, J E Baker
Post-doctoral Fellow: E Porter
Graduate Students: E H Kim, J Leaner, K McAloon, A Merten, A. Sveinsdottir

Project 5: An Integrative Modeling Approach to the Analysis of Scale-Dependent Effects of Spatial Heterogeneity

Principle Investigators: R H Gardner, W M Kemp, W Boynton, R Ulanowicz
Visiting Scholars: J Petersen
Faculty Research Assistants: D Fiscus
Graduate Students: R Bartelson, D Scheurer

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