Linking phenology models and monitoring data to improve planning of pest management programs

Planning to optimize timing and selection of pest control inputs requires having some idea of both the future status of pests, and the efficacy of control tactics. While data on pest abundance is routinely collected at regular intervals in many crop systems to monitor population trends, these data are rarely used for forecasting pest status or assessing controls. This is a key knowledge gap especially for crop systems in which phenology models are used to predict the proportion of pests of a certain stage based on degree-days, because data collected by growers could feed models to produce site-specific recommendations. Here we present a framework that produces in-season predictions on pest populations and assesses control programs based on scaled versions of phenology models to field counts. We used a 20-year dataset of codling moth Cydia pomonella (L.) pheromone trap captures for parameter estimation and model validation. Model validation revealed that > 75 % of the tested moth capture trajectories fell within prediction bands when they were produced after 350 degree-days. Also, simulations revealed that adult counts may be used to classify pest populations as treated or non-treated with control programs, based on the shifts in temporal patterns of adult emergence caused during the residual period of insecticides applied to kill larval stages. This framework is a powerful tool to optimize planning of pest management programs and could be applied to virtually any pest that can be sampled regularly and whose phenology can be modeled as a function of degree-days.