Prediction model for cabbage stem weevilCeutorhynchus pallidactylusMrsh. occurrence on winter rape based on an artificial neural network

Abstract

Cabbage stem weevil Ceutorhynchus pallidactylus Mrsh. is a important pest of oilseed rape. The impacts of weather conditions and developing a prediction model are key prerequisites for making decisions about chemical plant protection. Based on data from long-term monitoring occurrence of C. pallidactylus (2002–2012), those meteorological parameters with the most significant effects on spring raid intensity were selected and a prediction model based on an artificial neural network was developed. The model was trained using data on the capture of C. pallidactylus between 10 March and 26 April and on weather conditions during January/February and March/April. The winning neural network provides 97% predictive reliability based on mean air temperature for the third March pentad, mean air temperature in the last week of March, mean soil temperature at a depth of 10 cm in the last decade of March, and mean soil temperature at depth of 10 cm for January/February. The occurrence of C. pallidactylus decreased with increasing soil temperature during January/February; in March, the opposite effect was observed. The effect of air temperature on the occurrence of C. pallidactylus uring March has a peak form with a maximum at 3–4 °C and 6–7 °C in mid-March and at the end of March, respectively.