A Bayesian framework for the use of regional information in hydrology

Abstract

Water resource designs are perfect examples of decision making under uncertainty. In fact, three types of uncertainties may exist in any design problem: natural, parameter, and model uncertainties. The last two may be considered as informational uncertainties that are due to the lack of perfect information about the streamflow processes. The use of regional information has been suggested as a technique for reducing parameter uncertainties. The use of Bayesian methodology provides a framework for combining regional information with at‐site historical records. Moreover, Bayesian methods allow the hydrologist to consider the parameter uncertainties as well as the natural uncertainties within the decision‐making process. Because of these two advantages the Bayesian approach is a more complete and realistic approach to problems of uncertainty in hydrology and water resource planning than presently used methodologies.