Hydrological variation is believed to be the main density-independent factor that controls fish recruitment in floodplain ecosystems. However, our ability to fully understand these controls is greatly impeded by the size-selective nature of sampling gear. To illustrate the benefits of estimating the effects of size-selective bias on population parameters, we used a cohort analysis to reconstruct a 20-year time series of larval/neonate abundance for five species in the Order Cyprinodontiformes along a hydrological gradient in the Florida Everglades. We applied generalized linear models to estimate recruit density and analyze size-selectivity for our sampling gear. The adjusted data resulted in a 7 to 40-fold increase in estimated recruit density, which varied seasonally at regional and local spatial scales and was greatest at the end of the wet season (October, December) for most species; no consistent seasonality in recruitment of any species was apparent in the raw data. Using the adjusted data, we detected a positive relationship between recruit density and recovery time following marsh drying events at short and intermediate-hydroperiod sites for all species. However, depth was the major hydrological driver of recruitment at long-hydroperiod sites. Within sites, we observed interspecific variation in species responses to changes in annual hydroperiod. We suggest that fisheries models can be applied to data from any meshed sampling gear to improve abundance estimates and reveal seasonal recruitment dynamics. Our use of such models revealed seasonal recruitment dynamics that were previously undetected, with implications for planning of restoration and management of the Everglades.
