EVALUATION OF ALGORITHMS FOR FORECASTING OF INSECT POPULATIONS

摘要

Pests in Greenhouses can cause major economic damage. A modern method of minimizing this damage is integrated pest management. In this process various chemical and biological pest control methods are combined and their use is supported by a decision support system. These decision support systems are usually based on simulations of the future development of pests and the effects of control measures. A frequently used model for the simulation of insect populations is the escalator boxcar train model, a continuous simulation model based on partial differential equations. In this work, discrete population-based simulation models have been developed as an alternative. A time-discrete and a discrete event simulation model have been implemented for the simulation of the population development of an insect species and for the predator/prey interaction of two insect species. The population development of the greenhouse whitefly Trialeurodes vaporar-iorum has been simulated with these simulation models and validated with data from greenhouse experiments. The simulated population developments for the first four weeks are close enough to the experimental data, so that the simulations can be used in the context of a decision support system for the grower. Based on the validated models for the single species, the interaction of the greenhouse whitefly with the parasitic wasp Encarsia formosa has been modeled and simulated. It has been shown that with a suitable selection of parameters the actual population trend can be approximated, so that the algorithms can be used in a decision support system for pro-active organic pest control.