ATLSS:

Across Trophic Levels System Simulation

Network Analysis of Trophic Dynamics in South Florida Ecosystems

University of Maryland
Chesepeake Biological Laboratory
Solomons, MD 20688-0038

Personnel Details

Contract 1445-CAO9-95-0093


United States Geological Service


Biological Resources Division


A primary goal of coordinated research on South Florida's environmental resources is to understand those communities as whole ecosystems. Toward that end, the Across Trophic Levels System Simulation (ATLSS) project is an attempt to simulate the interactions of the various elements of wetland biotic communities within the framework of a single, encompassing computational scheme.


The resulting ATLSS model is one of the most complex and sophisticated models ever attempted. It will consist of simulation modules of varying, and often very high complexity that represent the important components of all the wetland ecosystems of South Florida. It follows that the output from ATLSS will be exceedingly complicated, and it may not be a straightforward task to elucidate the causal origins of any particular model behavior. Such uncertainty could become problematical, especially if the initial trials of ATLSS should behave "pathologically" (as is highly probable during the initial runs of such a complex model). Even should outputs not appear unrealistic, the difficulty remains that there exist no precedents for evaluating how well such a "multi-model" performs as an analog of the real system. That is, there are no set protocols for "calibrating" such complex simulations. The following pages describe only a component of the entire project. For further information about the entire ATLSS simulation, visit the ATLSS Home Page.

ATLSS, therefore, requires a partially independent benchmark against which one may gauge the plausibility of its outputs; and, towards that end, ATLSS investigators have chosen to create a suite of trophic flow networks that estimate material exchanges in the ecosystems being modeled. These will serve as calibration standards. In addition, these networks will be analyzed by a set of quantitative methods called Network Analysis (NA) that will serve as a guide for calibrating and debugging the initial modeling trials.

Trophic flow networks are graphical and mathematical depictions of the answers to the questions, "Who eats whom, and by how much?" Typically, diagrams of flow networks are comprised of boxes that represent the major components of the ecosystem. The boxes are connected by arrows, which indicate the transfers of material or energy between the components. Usually, each arrow is labelled with the magnitude of its transfer as averaged over a prescribed period of time.

Accordingly, the University of Maryland contingent of ATLSS investigators is assembling very detailled networks of carbon exchanges as they normally occur in the ecosystems of the South Florida wetlands. Networks ranging from 68 - 125 important compartments have been estimated for each of four habitats, using existing data and ongoing field work. Separate networks for wet and dry seasons are being created for the ecosystems of the forested wetlands, the gramminoid marshes, the mangrove estuaries, and the shallows of Florida Bay. Each network will be a snapshot of the trophic flows and biomasses as averaged both over the hydroperiod in question and over the spatial domain of that particular biotope.

Analyses were performed on the resulting networks at several scales. The key questions that can be answered for any fully- quantified trophic network include: (1) To what extent does each taxon depend upon (or contribute to) all other taxa over all trophic pathways, both direct and indirect? (2) What are the efficiencies with which material is being transferred up the trophic ladder? (3) What are the pathways by which material is being recycled within the system? and (4) What is the current organizational status of the ecosystem? Any or all of these answers can be used to debug an ATLSS model that is not performing realistically.

As of Fall 2000, we have constructed and analyzed the trophic networks for all four ecosystem types, and our final reports have been submitted. The results of our efforts are catalogued within this website, detailing both what is in these systems and what is occuring within them.

The following pages are best viewed using Netscape 4.0, or any other browser that supports style sheet formatting (HTML 4).