Development of a Bayesian network based hybrid-decision support process for potable water management in the context of the water framework directive

Abstract

Uncertain and complex environmental legislation governing the management of water resources has presented significant challenges to those responsible for identifying investment options to manage potable water supplies. This study aimed to develop a decision support process to enable a UK water company to understand and characterise the complex and uncertain implications of the Water Framework Directive (WFD) on the management of potable water supply. A flexible, exploratory and participatory approach was adopted, and included a central reference group comprised of managers representing different departments within the water company. Semi-structured interviews, informal discussions, focus groups, field visits, water company data, academic and legislative documentation, as well as UK water sector literature and observations by the researcher provided data which informed the criteria for and the population of a new Bayesian Network (BN) based Hybrid-Decision Support Process (Hybrid-DSP). Using BNs as a basis for decision support allowed the integration of diverse variables, as well as identifying and representing the relationships between them. The visual representation that BNs provided of the interrelationships between the variables, facilitated organisational learning in relation to the implications of the WFD for potable water management, which led to clearer identification of potential organisational responses. This study demonstrates the practical implications for the use of BNs within a water company in the UK. Furthermore a new BN based Hybrid-DSP has been developed through this study, which offers a systematic and holistic template to identify and analyse water company responses to the implementation of environmental legislation.