Published online by Cambridge University Press: 21 December 2019
This study examined the use of degree-day models to predict alfalfa weevil Hypera postica (Gyllenhal) (Coleoptera: Curculionidae) population development on the Canadian prairies. Air temperatures, alfalfa weevil abundance, and instar data were collected in 2013 and 2014 from 13 alfalfa (Medicago sativa Linnaeus; Fabaceae) fields across Alberta, Saskatchewan, and Manitoba. We coupled three alfalfa weevil population prediction models with three temperature data sources to determine which combination most closely aligned with results observed. Our objective was to find the best prediction of peak occurrence of second instar alfalfa weevils, the optimum time for management decisions. Of the parameters analysed, prediction model had the greatest effect on the accuracy of peak instar prediction, with Harcourt and North Dakota models better at predicting population peaks than the Guppy–Mukerji model. Interactions between temperature source and prediction model significantly affected prediction accuracy. The probability of accurate prediction of population peaks to within 3.5 days of actual occurrence using in-field and multiple-site temperature data sets, combined with Harcourt and North Dakota development models, was 0.45–0.70. Lower predictability was found from fields in the Mixed Grass Ecoregion than in other ecoregions. The use of the recommended models can assist growers in timing their monitoring activities and deciding if pest management action is warranted.
Subject editor: Cécile Le Lann