Measuring and understanding spatial variation of pests is a fundamental component of population dynamics. The resulting maps can drive spatially variable pest management, which we define as precision integrated pest management (IPM). Precision IPM has the potential to reduce insecticide use and slow the rate of resistance development because of the creation of temporally dynamic refuges. This approach to IPM requires sampling in which the objective is to measure spatial variation and map pest density or pressure. Interpolation of spatially referenced data is reviewed, and the influence of sampling design is suggested to be critical to the mapped visualization. Spatial sampling created problems with poor precision and small sample sizes that were partially alleviated with choosing sampling units based on their geostatistical properties, adopting global positioning system technology, and mapping local means. Mapping the probability of exceeding a threshold with indicator kriging is discussed as a decision-making tool for precision IPM. The different types of sampling patterns to deploy are discussed relative to the pest mapping objective.

No abstract

No abstract

No abstract

Most yield estimation practices for commercial vineyards are based on longstanding but individually variable industry protocols that rely on hand-sampling fruit on one or a small number of dates during the growing season. Limitations associated with the static nature of yield estimation may be overcome by deployment of trellis tension monitors (TTMs), systems that provide dynamic measurement of changes in the tension of the main trellis support wire. In 10 commercial vineyards from which two commercial juice processors annually collect data to derive yield estimates, TTMs were installed. Processor and TTM data were subjected to three permutations of the basic linear computational approach to estimating yield and their accuracies evaluated given known harvested yield at various spatial scales. On average, TTM data produced more accurate estimates of actual yield than did the computational protocols of the juice processors. There was high vineyard-to-vineyard variability in the accuracy of the estimate under all approaches, even from those permutations designed to match the spatial scale of the data collected for yield estimation with the spatial scale of the actual harvested yield. The processor protocols appear to be more sensitive than the TTM approach to the selection of the antecedent years used for comparison with the current year's data. Trellis tension monitoring may be useful to supplant traditional, labor-intensive yield estimation practices or to supplement longstanding practices with real-time information that can be applied to dynamic revision of static yield estimates.