Temperature effects on long‐term population dynamics in a parasitoid–host system

Abstract

Long‐term environmental changes will likely alter the strengths of interactions between species and consequently their population dynamics, leading to changes in the stability of ecological systems. While an increasing number of empirical studies have shown that environmental changes can alter the strengths of species interactions, these studies are typically short (<1–2 generations) and therefore give only partial information about longer term population dynamics. To focus on longer term dynamics, we investigated population cycles of pea aphids and their most common parasitoid, Aphidius ervi, in Wisconsin, USA. Data collected over three years in alfalfa fields showed an apparent host–parasitoid population cycle. Furthermore, higher pea aphid population growth rates and increased parasitism were correlated with higher naturally occurring temperatures. While these effects were observed with seasonal fluctuations in temperature, they beg the question of how long‐term changes in mean annual temperature might change aphid–parasitoid population cycles, a question which we further pursued with laboratory experiments. To quantify temperature‐dependent demographic parameters, we used short‐term (<1 generation) experiments conducted at 20°C and 27°C. The higher temperature increased aphid and parasitoid development rates, adult aphid life span and fecundity, and parasitoid attack rates. We then conducted multi‐generation population‐level laboratory experiments to reveal the effects of temperature (20°C vs. 27°C) on population dynamics. We fit the resulting time series data using a nonlinear age‐structured state‐space model to estimate population‐level processes that could not be estimated in short‐term laboratory experiments. Using the model, we parsed out the demographic rates that had the largest impacts on aphid–parasitoid population cycles. This analysis showed that there were frequent contrasts in the effects of temperature operating through different demographic rates. For example, the temperature‐dependent increase in aphid development rate decreased cycle amplitude, while the increase in parasitoid attack rate increased cycle amplitude. There were also striking interactions among demographic rates. For example, the temperature‐dependent increase in aphid development rate could either increase or decrease the cycle period depending on the values of other demographic rates. Although these complexities make predictions difficult, overall they suggest that increasing long‐term mean temperature will decrease the period, increase the amplitude, and tend to destabilize pea aphid–A. ervi dynamics.