Development and statistical evaluation of models in forecasting moth phenology of major lepidopterous peach pest complex for Integrated Pest Management programs

A 3-parameter Boltzman and a 4-parameter Logistic non-linear regression model were constructed to simulate the emergences and seasonal dynamics of the major moth pest complex of peach including Anarsia lineatella (Lepidoptera: Gelechiidae), Grapholitha (Cydia) molesta (Lepidoptera: Tortricidae) and Adoxophyes orana (Lepidoptera: Tortricidae). Model development used 4 yr (2004–2007) climatic data and field observations in peach orchards located in two separate regions in northern Greece. Model performance was evaluated by using the Adjusted r² and the Akaike and Bayes-Schwartz information criteria. Applied models had a very high prediction capability in most of the cases. The Akaike and Bayes–Schwartz information values of the 4-parameter Logistic model provided better estimates in most data sets. According to the Logistic equation, 50% of the cumulative number of male moths for the 1st, 2nd and 3rd generations of A. lineatella occurred at 151, 785 and 1513 degree-days respectively, for G. molesta at 654, 785 and 1251, and for A. orana at 406, 1260 and 2141 degree-days (Biofix: 1^st March in all cases). In addition, a simplified theoretical degree-day model adjustment was attempted to generate simultaneously a multispecies moth phonological model for the above species, in order to outline perspectives in Integrated Pest management (IPM). The unique proposed model revealed differences in the phenology of the three species, and can simplify calculations and recommendations to pest control advisors, based on a common base temperature and biofix.