The basic information for any model that projects future population dynamics should be good empirical studies. A large number of empirical studies have been done on the Florida Snail Kite. Such studies provide basic information on the biology of the species. They also provide the correlative relations between specific aspects of the snail kite life-history and behavior with the hydrology of the system. These relations form the building blocks of any hydrology-driven population-dynamical kite model.

Opinions differ on whether the best approach to modeling the life history of a population should be by means of a system-wide deterministic matrix model or, alternatively, a spatially-explicit stochastic individual-based model. We argue that rather than choosing among these two approaches, it is better to implement both concurrently. Next to the system-wide deterministic matrix model and the spatially-explicit stochastic individual-based model two other versions were implemented: a system-wide stochastic matrix model and a spatially explicit deterministic individual-based model. With these four tools in hand, we approached the challenge of making reliable projections of future population development of the snail kite under various hydrological scenarios.

Next to having a rigorous and transparent model structure, two issues are central to getting a reliable kite model: how to parameterize the model and how to set ranges of uncertainty to its output. The preferred statistical framework to solve both issues would be maximum likelihood estimation. In principle, the Maximum Likelihood Method provides a formal and rigorous approach to deal with the five sources of uncertainty: structural uncertainty, uncertainty in the hydrological input, uncertainty in the biological parameters, uncertainty due to demographic stochasticity, and finally uncertainty due to errors in the empirical data on basis of which the model is parameterized.

There is optimism that the kite model will be among the first applied population dynamical models that succeeds in disentangling and integrating these sources of uncertainty in a formal way. The current implementation of the model (Everkite 3.00), that has already been applied to evaluate hydrological scenarios and is ready to analyze new scenarios as they come, provides an excellent starting point for these new developments.

Contact: Wolf M. Mooij, Netherlands Institute of Ecology, Centre for Limnology, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands; Phone: +31 294 239352, Fax: +31 294 232224, w.mooij@nioo.knaw.nl

(This abstract was taken from the Greater Everglades Ecosystem Restoration (GEER) Open File Report 03-54)

Back to Project Homepage