Eruptive Insect Dynamics

Apr 13, 2018 · 1 min read
projects

Several of Canada’s most economically and ecologically important forest pests – forest tent caterpillar (Malacosoma disstria), spruce budworm (Choristoneura fumiferana), and mountain pine beetle (Dendroctonus ponderosae) – share a defining trait: their populations are governed by eruptive, nonlinear density-dependent processes that periodically push local densities from endemic to outbreak levels. Understanding how these dynamics unfold across space and time is central to anticipating where and when outbreaks will occur, how long they will persist, and under what conditions they will collapse.

Working with Barry Cooke and collaborators, we use long-term defoliation and aerial-survey records to characterize the periodicity, synchrony, and triggers of insect outbreaks across Canada. Time-series analyses reveal that broad-scale regional cycles often emerge from asynchronous local eruptions rather than from coherent climatic forcing, and that outbreak timing is inherently difficult to predict from environmental drivers alone. Recent work extends these methods to mountain pine beetle, where intensive monitoring data have allowed us to diagnose the density-dependent feedbacks underlying outbreak expansion – and, in Alberta, the engineered collapse of an ongoing outbreak through sustained green-tree removal.

This research complements the mountain pine beetle spread simulations and the broader boreal ecosystem forecasting work, providing the population-dynamic foundation on which spatially explicit forecasts of insect disturbance are built.

Alex M Chubaty
Authors
Ecologist and simulation modeller

Dr. Chubaty is an ecologist and simulation modeller who builds spatial models of forest landscapes and the species that live in them. He completed his PhD at Simon Fraser University modelling host selection in mountain pine beetle (MPB), followed by postdoctoral research at Université Laval and Natural Resources Canada developing forecasting models of MPB spread.

He runs FOR-CAST Research & Analytics in Calgary, Canada. His recent work includes LANDIS-II forest landscape simulation studies, landscape connectivity modelling, and analyses of insect outbreak dynamics, to inform land management and species-at-risk decisions. He writes open-source R packages for this work, including grainscape for connectivity analysis, landisutils for LANDIS-II simulations, and co-developed the SpaDES simulation platform. He advocates for open data and reproducible workflows.