Population models for threshold-based control of Tetranychus urticae in small-scale Kenyan tomato fields and for evaluating weather and host plant species effects.

The spatial distribution of motile life stages of the two-spotted spider mite Tetranychus urticae Koch in Kenyan small-scale tomato fields was described by Taylor's power law and an enumerative sampling plan was designed for research purposes. The exponential increase of T. urticae populations during three growing seasons permits the development and use of a simple exponential model for the design of a threshold-based chemical control system. For this purpose, a critical threshold of 440 motile mites per sample unit at the end of the growing season (12 weeks after transplanting) was translated into a proportion of 0.83 infested units in sample 3, i.e. five weeks after transplanting. A sequential binomial sampling plan with respect to the proportion of 0.83, five weeks after transplanting, was designed. The exponential model was extended to account for the influence of weather and host plant species. Model development and parameter estimation were based on three data sets (Kenyan tomato fields, Italian and Californian strawberry fields). The model satisfactorily predicted a positive influence on growth rates by (i) changing the host plant from tomato to strawberry, and (ii) temperature, while a negative effect resulted from (iii) rainfall; both (ii) and (iii) are controlled by temperature-rainfall interactions. In contrast, the estimated parameter values did not satisfactorily describe the expected responses at specific temperature and rainfall values. Nevertheless, the model allowed the rating of host plant species in the field. A fourth data set from Swiss apple orchards was used to test the model, and population build-up on the apple host plant appeared to be higher than on tomatoes but lower than on strawberry.