Validation of the RIMpro Decision Support System for Apple Sawfly (Hoplocampa testudinea) with field observations in The Netherlands; Belgium, Denmark and Austria 2010-2015

摘要

Apple sawfly is a key pest in organic and low-input apple production systems in Europe. Many organic orchards need an annual pesticide application to control the apple sawfly population at an economic level. The botanical larvicides used for this haveto be applied at or just before, the start of egg hatch. Consultants estimate the start of egg hatch by extrapolating the results of field observations on the embryonal development of the egg population. These microscopic observations are time consumingand have an unavoidable margin of error. We examined if the output of the RIMpro-Hoplocampa DSS is accurate enough to replace these field observations.The application dates as had been recommended tot fruitgrowers in The Netherlands, Belgium, Denmark and Austria between 2010 and 2015 were compared to the application dates as calculated by the simulation model for 123 orchards.Though treatments are aimed at first egg hatch, the bulk of the sawfly eggs hatch several days later. A foreseen long stretched period of egg hatch can force to the decision to postpone the first treatment, use a split-up treatment, of repeat the treatment. To examine if the variation in egg development observed in the field samples is explained by the model, the average development stage of the egg-population in each of 238 samples was compared to the average development as simulated by the model for that same date and location.In 86 % of the 123 cases the DSS advised date was close enough to the human expert advised date to expect the same level of efficacy (between 2 days before, and 1 day after). In cases were the difference was more important, this could be due to incompleteness of the simulation model, incorrect noted biofix date, unreliability of the on-farm weather station data, as well as inaccuracy of the human expert's estimation. The distribution in the egg development as simulated by the model matched the observedvariation in development stages close enough to allow advanced decisions on timing of control measures. (Correlation coefficients for the years examined range from 0.88 to 0.65).We concluded that the RIMpro-Hoplocampa DSS is a valuable tool for optimizing the management of apple sawfly populations. The model can reduce the field observations to a final check on the day before the DSS recommends the pesticide treatment, or even completely substitute the field observations.