Simulation models of the European red mite, Panonychus ulmi and its predator Typhlodromus pyri in New Zealand apple orchards

Abstract

The objectives of this study were to develop a phenology model of the European red mite (ERM), (Panonychus ulmi), and a predator-prey model of Typhlcdromus pyri and ERM, to aid in decision making in integrated mite control programmes in New Zealand apple orchards. An existing simulation model of a mite predator-prey system (Rabbinge, 1976) was investigated for its suitability as a starting point for a T. pyri / ERM predator-prey model. This model was tested against field data from New Zealand and sensitivity analysis was carried out. Although it was eventually abandoned, the results of the sensitivity analysis indicated where most research effort was required for a new model. A phenology model (Logan and Anman, 1986) was adapted for a phenology model of ERM, and was also used as the framework for a new predator-prey model. The predator-prey model developed, is a discrete-time deterministic simulation model, which has temperature as its only driving variable. It is age structured, and incorporates predation by all predator stages on multiple prey stages. The effect of the disparate spatial distribution of predator and prey on predation, was incorporated by a correction factor. The dynamics of ERM were modelled by fitting functions to published data. Information on the dynamics of the predator, and on the predator-prey interaction was obtained from laboratory and field experiments, and theoretical or empirical models were fitted to these data. The predator-prey model, and the phenology model were validated against field data, and sensitivity analysis was carried out. The predator-prey model gave reasonably good predictions of the densities of predator and prey mites in the field, during early season. However, the model predicted extinction of the prey population, and cessation of predator oviposition which were not observed in the field populations. A graphical spray decision chart, derived from repeated runs of the model suggested that the present spray action thresholds will result in use of miticides against ERM when they are not necessary. The phenology simulations were compared to the phenology of ERM in field populations from Central Otago, and Nelson, and gave good predictions of the optimum December spray time for miticides. However, the model was inconsistent in predicting population peaks and troughs later in the season. The sensitivity analysis indicated that both models were sensitive to the changes in temperature expected from season to season. The predator-prey model was also sensitive to the estimation of predation rates on adult female ERM, and the density estimates of predator and prey required to initialise the model. The insight into the predator-prey interaction, gained in this study, and the potential use of the models in integrated mite control was discussed. Suggestions were made for further experimental work required to refine and improve the models.