Elevational gradient in the cyclicity of a forest‐defoliating insect

Abstract

Observed changes in the cyclicity of herbivore populations along latitudinal gradients and the hypothesis that shifts in the importance of generalist versus specialist predators explain such gradients has long been a matter of intense interest. In contrast, elevational gradients in population cyclicity are largely unexplored. We quantified the cyclicity of gypsy moth populations along an elevational gradient by applying wavelet analysis to spatially referenced 31-year records of defoliation. Based on geographically weighted regression and nonlinear regression, we found either a hump-shaped or plateauing relationship between elevation and the cyclicity of gypsy moth populations and a positive relationship between cyclicity and the density of the gypsy moth's preferred host-tree species. The potential effects of elevational gradients in the density of generalist predators and preferred host-tree species on the cyclicity of gypsy moth populations were evaluated with mechanistic simulation models. The models suggested that an elevational gradient in the densities of preferred host tree species could partially explain elevational patterns of gypsy moth cyclicity. Results from a model assuming a type-III functional response of generalist predators to changes in gypsy moth density were inconsistent with the observed elevational gradient in gypsy moth cyclicity. However, a model with a more realistic type-II functional response gave results roughly consistent with the empirical findings. In contrast to classical studies on the effects of generalist predators on prey population cycles, our model with a type-II functional response predicts a unimodal relationship between generalist-predator density and the cyclicity of gypsy moth populations.

Population cycles produce periodic range boundary pulses

“…Moreover, determining invasion rates from a field measurement of population size, as we have done, provides a more accurate picture of spread dynamics than county quarantine records that are coarser in spatial scale and most importantly cannot show retraction of range boundaries. Prior research has demonstrated spatial (Haynes et al 2009(Haynes et al , 2012(Haynes et al , 2013 and temporal (Allstadt et al 2013) variation in the periodic behavior of gypsy moth population dynamics, so the dynamics in the time period and area studied here may differ from the overall, spatiotemporally averaged dynamics of the system. These prior studies of gypsy moth population cycles have focused on New England, New York, and Pennsylvania (where gypsy moth has been established longer, leaving a longer record for analysis), whereas our analysis focused on Virginia and West Virginia given our interest in linking defoliation to spatially detailed trapping records that quantify spread dynamics.…”

“…The modifications accord with the timing of these mortality sources in field populations. Prior studies have shown the model produces an Allee effect at low densities (Bjørnstad et al ) and periodic fluctuations in gypsy moth density across 4 orders of magnitude (Dwyer et al , Bjørnstad et al , Haynes et al ) as observed in field populations of the gypsy moth (Berryman ).…”

“…One exception to the 'data problem' is the invasion of North America by the gypsy moth Lymantria dispar, which has become a model system for understanding the dynamics of biological invasions due to extensive monitoring of range expansion (Tobin et al 2004(Tobin et al , 2007a and the ability to use aerial survey defoliation data as a proxy for population outbreaks to study the cyclical fluctuations of established populations (Liebhold and Elkinton 1989, Johnson et al 2006a, Bjørnstad et al 2010, Haynes et al 2012. Spatiotemporal patterns of gypsy moth spread are not adequately described by classical spread models (Liebhold et al 1992) due to stratified diffusion resulting from the long-distance transport of gypsy moth life stages (Tobin and Blackburn 2008, Bigsby et al 2011, Frank et al 2013.…”

Transient synchrony among populations of five foliage‐feeding Lepidoptera

“…It has been found that such measurements of area damaged correlates well with measures of population abundance (e.g., light trap data) and can therefore can be used as proxies for population density (Leskó, Szentkirályi, & Kádár, , , ). Furthermore, area of defoliation is a commonly used metric used in analyses of large‐scale population dynamics and spatial synchrony (e.g., Allstadt, Haynes, Liebhold, & Johnson, ; Haynes, Liebhold, & Johnson, ; Liebhold et al., ; Williams & Liebhold, ).…”

Optimal control of integrodifference equations in a pest-pathogen system

“…For previous work on discrete time models for gypsy moth population without spatial features, see [12,10,11,5,23].…”