Canadian Forest Service Publications

An object-oriented, process-based stochastic simulation model of Bacillus thuringiensis efficacy against spruce budworm, Choristoneura fumiferana (Lepidoptera: Tortricidae). 1996. Cooke, B.J.; Régnière, J. International Journal of Pest Management 42(4): 291-306.

Year: 1996

Available from: Laurentian Forestry Centre

Catalog ID: 16731

Language: English

CFS Availability: PDF (request by e-mail)


An individual-based, process-oriented simulation model was designed to investigate the effects of Bacillus thuringiensis var. kurstaki (Bt) sprays on populations of spruce budworm. The model shows that the diameter of Bt droplets is the single most important determinant of efficacy. Fine sprays (numerous small droplets) are predicted to cause a considerable amount of larval feeding inhibition and subsequent protection of foliage. Post-spray temperatures are predicted to have an important impact on the level of final mortality. When susceptibility to Bt is assumed to be constant between instars, the model predicts that spraying late (instar V or VI) should kill more larvae and lead to somewhat lower defoliation, compared with spraying early (instar IV). Conversely, when susceptibility is assumed to be proportional to body weight, optimal spray timing in terms of foliage protection and budworm kill occurs during instar IV. Efficacy predictions were highly sensitive to some questionable assumptions regarding the expression of ingested doses of Bt. This underscores the need for more basic research into the mode of action of Bt. Full model validation could not be performed because of insufficient input information from field trials. Nevertheless, two comparisons of predicted and observed efficacy were attempted with more complete spray-trial data sets from Ontario, although in both cases estimates of several important model parameters were lacking. In the first comparison, model predictions were commensurate and consistent with observations. In the second, predicted efficacy grossly underestimated observed efficacy. Model responses to major input parameters and alternatives to critical assumptions are discussed.

Date modified: