A Bayesian Method for Sequential Sampling and Forecasting in Agricultural Pest Management*

We present a method for sequential estimation of the state of infestation of an agricultural crop by a pest arthropod. The method, which is based on Bayesian statistical theory, establishes stopping curves by requiring that a confidence interval be placed entirely on one side or the other of an economic threshold. The results of a given sequential sample are used to modify the prior of the subsequent sample by assuming that the infestation follows a particular growth law. The parameters of this growth law are themselves uncertain and are described by subjective distributions that are modified by the sampling. A major advantage of our method is that it readily admits the incorporation of a forecasting scheme. After describing the development of the method in general terms, we give a specific example of the application of the method to the sampling of spider mites on cotton in the San Joaquin Valley, California.