A new parameterization for integrated population models to document amphibian reintroductions

Abstract

Managers are increasingly implementing reintroduction programs as part of a global effort to alleviate amphibian declines. Given uncertainty in factors affecting populations and a need to make recurring decisions to achieve objectives, adaptive management is a useful component of these efforts. A major impediment to the estimation of demographic rates often used to parameterize and refine decision-support models is that life-stage-specific monitoring data are frequently sparse for amphibians. We developed a new parameterization for integrated population models to match the ecology of amphibians and capitalize on relatively inexpensive monitoring data to document amphibian reintroductions. We evaluate the capability of this model by fitting it to Oregon spotted frog (Rana pretiosa) monitoring data collected from 2007 to 2014 following their reintroduction within the Klamath Basin, Oregon, USA. The number of egg masses encountered and the estimated adult and metamorph abundances generally increased following reintroduction. We found that survival probability from egg to metamorph ranged from 0.01 in 2008 to 0.09 in 2009 and was not related to minimum spring temperatures, metamorph survival probability ranged from 0.13 in 2010-2011 to 0.86 in 2012-2013 and was positively related to mean monthly temperatures (logit-scale slope = 2.37), adult survival probability was lower for founders (0.40) than individuals recruited after reintroduction (0.56), and the mean number of egg masses per adult female was 0.74. Our study is the first to test hypotheses concerning Oregon spotted frog egg-to-metamorph and metamorph-to-adult transition probabilities in the wild and document their response at multiple life stages following reintroduction. Furthermore, we provide an example to illustrate how the structure of our integrated population model serves as a useful foundation for amphibian decision-support models within adaptive management programs. The integration of multiple, but related, data sets has an advantage of being able to estimate complex ecological relationships across multiple life stages, offering a modeling framework that accommodates uncertainty, enforces parsimony, and ensures all model parameters can be confronted with monitoring data.

Survival cost to relocation does not reduce population self‐sustainability in an amphibian

“…This pattern is congruent with the results of a study focusing on the relocation of Rana pretiosa (Duarte et al. ). Interestingly, by the time locally born descendants reached adulthood, there was no difference in survival rates compared to the other populations in any life stage.…”

“…, Duarte et al. ), including ours, highlighted a survival loss of released individuals at different life stages. Those demographic estimates are critical to fixing the values of the parameters considered in adaptive management models (McCarthy et al.…”

“…However, the lack of knowledge about the demography of amphibians and the technical limitations to monitor their populations have often prevented the use of those tools so far. Several studies (Muths et al 2014, Duarte et al 2017, including ours, highlighted a survival loss of released individuals at different life stages. Those demographic estimates are critical to fixing the values of the parameters considered in adaptive management models (McCarthy et al 2012, Canessa et al 2016, which could be precious tools to increase the chance of relocation success in amphibians.…”

Integrated population models: powerful methods to embed individual processes in population dynamics models

“…, Duarte et al. ), and also more accurate estimates of parameters influencing demographic rates (Plard et al. ).…”

“…The first insight is the estimation of stage-specific demographic rates for which no explicit data have been collected. By combining data on different stages of an amphibian species (egg, metamorph, and adult), Duarte et al (2017) showed that survival from the egg to the metamorph stage and from the metamorph to the adult stage, respectively, and fecundity, could all be estimated separately in an IPM. However, one associated drawback with the additional parameters that can be estimated only from IPMs is that they are usually estimated with low precision and possibly also with bias.…”