Abstract
A Bayesian decision framework is used to determine the value of information from water sampling with respect to alternative pollution abatement strategies. The critical low stream flow months of August 1966-77 for a New England river basin are examined in a restricted ordinary least-squares regression analysis using daily river flows as hydrologic variables, daily precipitation, and policy variables representing hydro-power generating practices. Posterior probabilities for discrete states of water flow/quality are calculated from forecast confidence intervals.
