Evaluating amphibian occurrence models and the importance of small, isolated wetlands in the Delaware water gap national recreational area

Isolated wetlands lack permanent stream connections to adjacent bodies of water, and are ecologically important because they can be the most abundant type of wetland in a region, and can harbor unique biological communities. Despite this, isolated wetlands are rarely protected by federal and state legislation, and federal agencies do not assess their abundance when reporting national trends in wetland losses. In this dissertation, I present the results from a field study that illustrates how the presence of small and isolated wetlands is important in maintaining populations of pond-breeding amphibians across the Delaware Water Gap National Recreational Area located in New Jersey and Pennsylvania. I found that stream connections at wetlands strongly influenced the occurrence patterns of five out of nine amphibian species studied, and species that bred earlier in the year were the most likely to occur at small and isolated wetlands. I used occurrence patterns at surveyed wetlands to predict the presence of amphibian species across the entire study area, and estimated that two-thirds of all species would lose at least half of their breeding sites if either isolated wetlands or small wetlands (< 0.30 ha in size) were removed from this landscape. I also characterized the abundance and sizes of isolated and non-isolated wetlands that could serve as amphibian breeding sites, and modeled the probability these wetlands were detected on wetland maps as a function of their size. I found the majority of wetlands were small (< 0.5 ha) and isolated, and this trend was similar to several landscapes in the northeastern United States. Thus, if these landscapes were to follow national trends in wetlands loss, approximately 21 wetlands smaller than 0.5 ha could be degraded or destroyed for every wetland larger than 0.5 ha. Furthermore, I estimate this disparity could be twice as large because the rate small wetlands in this study area were included on wetland maps was considerably lower than inclusion rates for larger wetlands.