Abstract
Seasonal pulsing of hydrology in the Greater Everglades watershed produces dynamic, annual cycles of wetland flooding and drying, and of environmental conditions. Biota have adapted to exploit these pulses through movement and foraging behaviors. A primary ecological concern, pertinent to Everglades restoration, is to understand how sufficient concentration of aquatic biomass, particularly that of small fish and invertebrates, is produced under nutrient limited conditions, to support diverse and abundant communities of piscivorous predators, such as wading birds and herpetofauna. This question is investigated through computer simulation modeling of small−bodied fishes in the freshwater Everglades, and analyses of empirical monitoring data of mangrove−associated fishes along the shoreline of southern Florida. In the freshwater Everglades, concentration of biomass, through interactions between movement behaviors, spatial heterogeneity of the landscape, and declining water levels, has been hypothesized to be a key function for making biomass available in useable densities. Fish avoid drying conditions by moving to adjacent refugia, and in the process become highly concentrated in localized areas. Although this conceptual picture is well known, spatially explicit prediction of the timing, magnitude, and duration of fish concentration across the landscape has not been possible. Here, a computer modeling approach, aimed at complementing monitoring, is developed for simulating and analyzing landscape and food web dynamics of pulsed wetland ecosystems, by explicitly tracking growth and concentration of fish biomass, and accounting for dynamic movement behaviors in response to spatiotemporally shifting environmental conditions. In coastal mangrove habitats, fluctuating and extreme salinity conditions have been hypothesized to constrain fish populations, and avoidance of hypersaline areas through movement is possible. Fish densities were related to frequency of hypersalinity over defined periods preceding surveys. Complementary quantile regression techniques were applied to make inferences on behavioral, population, and predator−prey dynamics, which can be tested with modeling. Ecological responses at multiples spatial and temporal scales were investigated in both the empirical and simulation studies. Spatially explicit maps of biomass concentrations, across these scales, were generated from simulation models, and effects of hydrology and landscape connectivity on fish biomass were shown. This methodology offers a new approach for using empirical analyses to generate hypotheses and theory which can be investigated with computer simulation modeling, and for using computer modeling to inform decisions on monitoring strategy and resource management.