BAYESIAN-ANALYSIS OF AGRICULTURAL DRAINAGE PROBLEMS IN CALIFORNIA SAN-JOAQUIN VALLEY

Bayesian network analysis was found to be an effective way of avoiding the problems of inconsistent and attitudinal deduction rules that often characterize traditional expert systems. Key experts and decision makers on the management of saline and toxic irrigation drainage in California's San Joaquin Valley have widely different and strongly attitudinal causal explanations of the drainage problem. It is therefore impossible to design a deduction method for solving the problem on the basis of individual expert rules. Bayesian analysis of the network of aggregated expert rules redefines the drainage problem. The inconsistencies and attitudes found in individual experts' problem definitions are in the Bayesian approach direct indicators of potential remedies. According to the analysis, a systemic decision-making dilemma persuades drainage experts to avoid action toward a long-term, Valley-wide drainage solution for the sake of short-term organizational and political benefits. Since the organizational and political context is outside the immediate domain of knowledge of drainage experts, the dilemma has consistently escaped effective remedies. Any techno-economic intervention in San Joaquin Valley's drainage management should therefore be facilitated by simultaneous administrative restructuring of California's drainage authority. © 1993 Academic Press. All rights reserved.