Survival in two neotropical frog species prior to a disease outbreak

Background/Question/Methods

While many amphibian species continue to experience catastrophic declines worldwide, we still have little understanding of the population dynamics of the majority of species. Estimating survival rates and population densities can help us understand the spatial and temporal dynamics of healthy populations, assess how populations may be impacted by threats, and monitor the recovery of diminished populations. Mark-recapture techniques represent a powerful approach for estimating such vital rates, and have rarely been used in studies of tropical amphibians. We use five years of mark-recapture data (2000-2004) to assess survival rates of two stream-breeding frog species, Centrolene prosoblepon and Atelopus varius, prior to the arrival of the amphibian chytrid fungus. We studied the two species on four streams (Cascada, Guabal, Loop, and Silenciosa) in Parque Nacional G. D. Omar Torrijos, a mid-elevation, wet tropical forest in Panama. We estimated density and monthly and annual survival probabilities for male C. prosoblepon and monthly survival rates for male A. varius using Bayesian Cormack-Jolly-Seber models. These two species differ in natural history and habitat use, allowing us to explore what habitat features and life history characteristics may drive differences in survival.

Results/Conclusions

Across all streams and years, mean annual survival for male C. prosoblepon was 0.45 (0.43-0.51 CI) and mean monthly survival was 0.93. However, frogs on Guabal had lower annual survival rates (0.15-0.40) than frogs on the other three streams (0.42-0.70) in most years. Densities of C. prosoblepon varied substantially among streams, with higher densities on Loop (71+/-14 SE) and Cascada (63+/-6), and lower densities on Guabal (24+/-2) and Silenciosa (27+/-5). For A. varius, monthly survival was generally higher and more constant among years for individuals on Guabal (0.45-0.83) and Silenciosa (0.56-0.83) compared to Loop (0.26-0.83) and Cascada (0.32-0.68). The majority of A. varius individuals (62.5%) were captured on Guabal. The higher survival rates seen in C. prosoblepon compared to A. varius probably indicate a longer life span in the former species. Differences in survival rates and densities may reflect fine-scale habitat differences among streams, and species-specific breeding requirements. Finally, we show that streams with very different densities of individuals can have similar survival rates, suggesting habitat limits to population size. These survival estimates are the first that exist for these species. They provide a critical baseline for examining survival in other healthy areas and areas affected by disease and other threats.