Abstract
An experimental design procedure is proposed for use as a guide in optimal sensor placement and sampling during meteorological data gathering expeditions. State-of-the-art knowledge about the structure of the phenomenon to be observed is transformed into a null-hypothesis in the form of a space-time covariance function. This function forms the main component required to evaluate the benefits (accuracy) to accrue from contending sampling systems. Contenders are eliminated through the use of a non-linear-programming search technique and the optimal configuration selected. Constraints on the sampling arrays are permitted. Several examples are presented.
