Simulating the response of a threatened amphibian to climate-induced reductions in breeding habitat

Context Amphibians are declining worldwide due to disease, invasive species, and habitat loss. Climate change may exacerbate habitat loss by altering the availability or suitability of aquatic breeding habitat through changes in precipitation, temperature, and the biophysical factors they influence. Measuring biological vital rates and the environmental covariates that affect them are crucial to understanding amphibian responses to a changing climate. However, doing so can be difficult due to access constraints, funding limitations, and difficulty of measuring spatially structured populations at landscape scales. Simulation-based approaches may provide reliable approximations of amphibian responses to changing environmental conditions. Objectives In this study, we constructed a spatially explicit individual based model to simulate the response of the Arizona treefrog ( Hyla wrightorum ), a species of conservation concern, to reductions in breeding habitat availability. The Arizona treefrog metapopulation in the Huachuca Mountains and Canelo Hills of Arizona is known to breed at fewer than 20 breeding ponds that may become unsuitable due to climate change-induced shifts in hydroperiod, increased rates of invasion by non-native predators, or filling by sedimentation. Methods We simulated Arizona treefrog response to bioclimatically realistic scenarios that included reduced breeding habitat availability, failed recruitment, and a combination of both. Simulations included variable landscape configuration through space and time that were informed by an empirical hydroperiod dataset. Results We found that climate-driven reductions in breeding habitat alone resulted in mean population declines of nearly 65%, even when reductions targeted the ponds least likely to fill with water. However, scenarios with concurrent breeding habitat loss and recruitment failure or very high recruitment failure resulted in 79% and 83% population declines, respectively. Reduced breeding habitat also increased spatial synchrony of occupancy through time and among simulations, pointing to a potential transition from a metapopulation to multiple isolated populations. Conclusions In the face of anthropogenically driven climate change and continually emerging management challenges, individual-based models provide a useful, mechanistic tool to explore how a combination of biological and environmental factors may interact to influence the future of species of conservation concern. We found evidence of a potential transition to isolated populations that may lead to limited functional connectivity and an increased risk of regional extinction for this species of conservation concern. Sensitivity analyses suggest our model was robust to uncertainty in model parameters and point to the critical role of dispersal in maintaining demographic and landscape connectivity among spatiotemporally heterogeneous habitat. A spatially explicit approach also enabled identification of specific habitat patches that may be less sensitive to climate change as well as those that may require more intensive management to conserve key populations or preserve metapopulation dynamics.