Print PageThe annual drying of the Florida Everglades wetlands creates a challenge for fish and macroinvertebrates that must find refuges for the dry season. We hypothesize that landscape features and species-specific patterns of refuge finding determine the makeup of communities in drying pools that attract wading birds for dry-season foraging. Small fish and macroinvertebrates were collected using throw traps to evaluate this hypothesis. From 2005 through 2009, we collected wet-season samples from over 150 sentinel sites located throughout the Everglades, while in the following dry seasons we sampled at a subset of the wet-season sentinel sites, random drying pools, and at locations where wading birds were observed actively feeding. We tested whether fish communities in drying pools could be predicted by local wet-season community data and the local habitat matrix. We compared wet-season community composition to composition in foraging pools and randomly selected sites in the dry-season. Euclidean distances between wet and dry sites, and least-cost-path distances, using both dichotomous (permeable or impermeable) and cost-weighted landscape grids, were measured using ARCGIS. This was done to measure theoretical swimming distances between sites through the heterogeneous Everglades landscape. These were incorporated into a multiple regression model to predict dry-season fish biomass
Results/Conclusions
We found that the spatial distribution of wet-season wading bird prey biomass changed dramatically from 2005 to 2009, with the highest numbers in 2008 and 2009 in short-hydroperiod areas. Prey biomass in random drying pools was usually higher than in dry-season sentinel sites. Drying pools where wading birds were observed actively feeding often had very elevated densities of prey species. Using spatial cross-correlation analysis, wet-season prey fish biomass was a predictor of dry-season biomass when collected within 10 km, after which, there was rapid distance decay in predictability. Community composition was more dissimilar between wet and dry-season sites when water levels dropped quickly and when greater disparities were present in total biomass. Dry-season sample type (sentinel, random pool, or foraging sites) also affected the dissimilarity between wet and dry-season samples. Akaike's information criterion indicated that cost-weighted, but not Euclidean or dichotomous-swimming, distance was important for predicting dry-season fish biomass. Differences in drying pool communities resulted from different species' recruitment, habitat use, dispersal, and risk aversion in a seasonally-drying heterogeneous landscape.
