Bayesian methods for analysing climate change and water resource uncertainties

Abstract

The purpose of this paper is to outline the advantages of the Bayesian approach for analysing uncertainties involving climate change, emphasizing the study of the risks such changes pose to water resources systems. Bayesian analysis has the advantage of basing inference and decisions on a coherent and normatively appealing theoretical framework. Furthermore, it can incorporate diverse sources of information, including subjective opinions, historical observations and model outputs. The paper summarizes the basic assumptions and procedures of Bayesian analysis. Summaries of applications to detection of climate change, estimation of climate model parameters, and wetlands management under climatic uncertainty illustrate the potential of the Bayesian methodology. Criticisms of the approach are summarized. It is concluded that in comparison with alternative paradigms for analysing uncertainty, such as fuzzy sets and Dempster-Shafer reasoning, Bayesian analysis is practical, theoretically sound, and relatively easy to understand.