Evaluation of sequential presence-absence sampling plans for the diamondback moth (Plutellidae: Lepidoptera) in cruciferous crops in Australia.

Two sets of sequential presence-absence sampling plans for decision-making in the management of diamondback moth, Plutella xylostella (L.), were developed and evaluated. One set of sampling plans targeted the classification of proportions of infested plants, and the other set of sampling plans targeted the classification of larval density. The action thresholds investigated were 0.15, 0.25, 0.35, and 0.45 proportion of plants infested with larvae, and 0.2, 0.4, 0.6, and 0.8 larvae per plant. They are representative of the action thresholds currently practiced by Australian crucifer growers. For each sampling plan, the population range within which a minimal correct decision rate of 95% can be expected at a maximal average sample size of 50 plants (OC95ASN50) was specified. The closeness of an OC95ASN50 range to the target action threshold is a measure of the expected performance of the sampling plan. A closer distance reflects better performance. The OC95ASN50 ranges of the proportion-classification sampling plans were within 33-53% of the target action thresholds. The width of these OC95ASN50 ranges represents 73-87% of the entire population range (0-1). For the classification of larval density, an empirical proportion-density model was first established using data from different states and different cruciferous crops. The OC95ASN50 ranges of the density-classification sampling plans were within 57-75% of the target action threshold. Simulated sampling of 20 independent data sets showed that for most data sets the correct classification rate was at least 98% and the matching average sample size was <50 plants.