Integrated eco-hydrological modeling of forage fish resource pulse accumulation through changes in landscape connectivity and isolation
Seasonally-pulsed wetland ecosystems that expand and contract in area intra-annually have dynamic patterns of hydroscape connectivity that alternately enable and impede dispersal of seasonally migrating fish. When areas become disconnected as water levels recede, some fraction of a population becomes trapped and forms super-concentrations, or “pulses.” In oligotrophic ecosystems, these pulses are necessary for concentrating prey in sufficient densities to make them available for top predators. We developed a computer simulation model, GEFISH, to quantify how seasonal hydrologic patterns, landscape connectivity, food web interactions and fish movement behaviors combine to affect forage fish biomass production and accumulation in Everglades freshwater wetlands. Forage fish are key food web links that supply large amounts of biomass energy to top predators, and high concentrations are essential for detection by tactile feeders such as wood storks. These concentrations were historically formed through gradual landscape drying over a complex, braided ridge-and-slough topography that trapped fish sequentially. This structure has been degraded in many areas due to human-induced alterations in hydrology. We hypothesized that changes in landscape connectivity due to degraded ridge-and-slough structure and altered hydrology result in reduced landscape-scale fish dispersal, decreased production and stranding, and overall worse foraging conditions for wading birds.
We simulated a reciprocal translation experiment using two different intact and degraded Everglades ridge-and-slough landscapes derived from empirical plots, and two respective twelve year time series of water level data (2001-2012). Three forage fish species were given different life history traits and movement strategies based on empirical relationships. Fish were assumed to disperse between shallow water habitats seeking foraging opportunities and predator refuges. The three fish groups competed for the same resources with different parameters for movement speed, probability of movement, stranding and feeding. The landscapes, hydrology patterns, and fish movement behaviors used were highly realistic, and the model was spatially-explicit on a continuous 100 x 100 grid of cells representing 4 km2 plots. Twelve years of simulated biomass were analyzed to proxy different scenarios for restoring ridge-and-slough habitat and water flow. Overall, the intact ridge-and-slough landscape was better for generating fish biomass for wading birds than the degraded landscape, producing greater biomass availability over longer time periods. The intact landscape had interconnected networks and seldom dried out completely, resulting in a high persistence of directional connectivity and net fish movement, while the degraded landscape had isolated pools and longer drought periods, resulting in low persistence of connectivity.