The Forest Tent Caterpillar in Minnesota: Detectability, Impact, and Cycling Dynamics

Abstract

If periodically outbreaking forest insects are a generic source of forest decline, then why do outbreaks recur more periodically than decline episodes? Do standard field survey data and proxy data systematically underestimate the complexity in herbivore population dynamics? We examine three sources of previously un-analyzed time-series data (population, defoliation, and tree-ring radial growth) for the forest tent caterpillar, Malacosoma disstria Hübner (Lepidoptera: Lasiocampidae) feeding on trembling aspen, Populus tremuloides Michx. (Salicaceae), in Minnesota, in order to answer these questions. Spatial pattern analysis of defoliation data indicated not only that outbreaks are roughly periodic, with a 10–13-y cycle, but also that important deviations from periodic led to large-scale episodes of aspen decline starting in the 1950s and 1960s, near Duluth and International Falls, respectively. By using additional data from Alberta, Canada we identify critical population and defoliation thresholds where defoliation becomes aerially detectable and impactful on tree growth. The threshold where defoliation becomes aerially detectable was found to be ~50% defoliation, corresponding to a population density of ~12 egg bands per 20 cm DBH tree (or ~20 cocoons per 3 min of collection time, or ~10 male moths per pheromone trap), and which implies a radial growth reduction on the order of 40%. We found that not all moth population peaks occur above the threshold level where defoliation is aerially detectable. Asynchronous pulses of defoliation—which are difficult to detect—produce asynchronous signatures of outbreak in tree-ring data. When these pulses occur in close conjunction with regular cycling, it can lead to outbreaks of prolonged duration that result in anomalously high tree mortality.