Estimating stage-specific demographic rates using a hidden Markov model for juvenile stages in Anaea aidea (Nymphalidae)

In butterflies, little is known about the dynamics of demographic rates (survival and stage transitions) among juvenile stages in nature. We quantified the progression of juvenile stages in a common butterfly of the southeastern U.S., by repeated censusing of 510 caterpillars in their natural environment, recording developmental stage (egg, 5 instars, pupa) and whether individuals were alive or dead every 3 days for one month. We ask how the probability of surviving and transitioning between stages varies in time, testing for effects of temperature and density. We modeled the process underlying our observations as a hidden Markov process because the true demographic process is partially hidden due to imperfect detection of caterpillars. Detection, survival, and stage transition probabilities were estimated given each individual encounter history. During our survey, caterpillar densities increased to a maximum then declined and attracted multiple insect and arachnid predators. Survival declined as per capita predation rate increased. Transition rates increased with warming spring temperatures, likely due to increased feeding and metabolism. These demographic rates will parameterize the juvenile stages of a periodic stage-structured matrix model that characterizes the life cycle of this bivoltine butterfly in 3-day time steps over one year. Analysis of this matrix reveals life history traits (e.g. life expectancy) and trait sensitivity to changes in demographic rates through time for Anaea aidea. Our next step will be to generalize this model to elucidate the life history of a closely related endangered butterfly, A. floridalis, for which there is very limited empirical data.