Assessing spatial and temporal population dynamics of cryptic species: an example with the European pond turtle.

Within the current context of biodiversity loss a number of biodiversity indicators have been developed to help measure the state of nature and how it is changing. However, most indicators are derived from bird populations. Reptiles and amphibians could be useful ecosystem indicators, but this requires obtaining precise and unbiased population parameters. This is a particularly challenging task for these two groups of species, because individuals are extremely difficult to detect for various reasons. We illustrate the use of a sampling and analytical method that explicitly takes into account imperfect detection to assess the population dynamics of a reptile species and its temporal and spatial variation. European pond turtles (Emys orbicularis) were sampled at two different locations during a 10-year study. The two sites differed by their water management regimes and number of livestock. At each site and for each sex, the data were modeled using a robust design capture-mark-recapture framework to obtain and compare estimates of survival, temporary emigration, time-specific abundance, density, and population growth rate, while taking into account effects of individual heterogeneity, trap-response, and time on capture probabilities. Temporary emigration was higher in males (0.344 +/- 0.046) (mean +/- SE) than in females (0.228 +/- 0.071), did not differ between sites for each sex, and was mainly Markovian. Apparent adult survival was higher in females (0.966 +/- 0.031) than in males (0.864 +/- 0.024), and was negatively related to the proportion of individuals with repaired shell fractures in the population. Average density was 63.7 +/- 6.6 turtles/km2 but was higher for females (39.2 +/- 14.0 females/km2) than for males (23.5 +/- 5.7 males/km2). The population sizes of males and females decreased where long periods of artificial drought and the highest density of livestock occurred. This research highlights differences in demographic parameters according to sex and site in a turtle species, and it provides evidence that significant human-induced disturbance represents a potential risk to European pond turtle populations. The sampling and analytical approaches illustrated in this study are applicable to many other species of reptiles and amphibians, and estimated population parameters could be combined to produce population indicators useful for conservation and management.