Long term amphibian monitoring at wetlands lacks power to detect population trends
Abstract
Amphibians are declining worldwide due to habitat destruction, disease, and environmental stressors. Extremely variable breeding populations and a paucity of long-term monitoring data limits rigorous testing of amphibian population trends, or bias associated with sampling regimes. We used 24 years of continuous trapping data to compare annual probability of presence, and population trends and statistical power for six species among seven wetlands using five sampling scenarios (SS) based on the interval and span of years analyzed. Richness within a year and wetland ranged 29–89% of total species captured there (all years), and 27–82% of total species captured during the study (all years, pooled wetlands). SS had little effect on probability of presence for most common species but did for less common species. Population trends were inconsistently significant or nonsignificant among wetlands within SSs, and among SSs within the same wetlands. The direction (+/−) of trends among wetlands and scenarios for a species generally agreed, but not always. Low statistical power for virtually all population trend estimates, including the All-years SS indicated results were inconclusive. Juvenile recruitment was correlated with adult populations in some subsequent years for four of the six species. We illustrate how probability of presence and population trend estimates can differ among similar wetlands within a landscape, and according to the span, or subset of years sampled. Our results indicate that amphibian monitoring at wetlands cannot conclusively gauge population trends for breeding populations that fluctuate widely among wetlands and from year to year.