Survey methods for assessment of Citrus tristeza virus incidence in urban citrus populations

This article concerns survey methodology for pathogens in urban citrus populations, motivated in particular by the need for assessments of Citrus tristeza virus (CTV) incidence. We envisage a large area R not devoted primarily to the commercial cultivation of citrus, that nevertheless has a substantial population of citrus trees. It is desired to sample the citrus population of area R in order to be able to make a statement about the level of infection of the population with CTV, or with particular isolates thereof. We describe a two-stage acceptance sampling scheme in which area R is divided into N sampling units, of which n are inspected. The size of the sampling units, while much smaller than R, is still large, so subsampling is carried out, introducing the possibility of misclassification of sampling units. To account for misclassification of sampling units, a larger number must be inspected than if it were assumed that there were no misclassifications. We describe the calculation of sample sizes required for subsampling within sampling units and for the total number of sampling units to be inspected, using parameters that can be adjusted to meet different specified regulatory scenarios.