Multimodeling: new approaches for linking ecological models

The Everglades region of South Florida presents one of the major natural system management challenges facing the United States. With its assortment of alligators, crocodiles, manatees, panthers, large mixed flocks of wading birds, highly diverse subtropical flora, and sea of sawgrass, the ecosystem is unique in this country (Davis and Ogden 1994). The region is also perhaps the largest human-controlled system on the planet in that the major environmental factor influencing the region is water, and water flows are managed on a daily basis--subject to the vagaries of rainfall--by a massive system of locks, pumps, canals, and levees constructed over the past century. The changes brought about by such control have led to extensive modifications of historical patterns and magnitudes of flow, causing large declines in many native species, extensive changes in nutrient cycling and vegetation across south Florida, and great increases in pollutants such as mercury. Constrained by the conflicting demands of agriculture, urban human populations, and wildlife for control of water resources, and the varying agendas of hosts of government agencies and nongovernmental organizations, there is now an ongoing effort to plan for major changes to the system with expenditure estimates of eight billion dollars or more over the next several decades (USACOE 1999). Carrying out such planning, particularly as it impacts the natural systems of the region, provides one of the major challenges to the new field of computational ecology.