Participatory development of a Bayesian network model for catchment‐based water resource management

Abstract

[1] A participatory approach was used to develop a Bayesian network model for assisting integration of water resource management in the Kongulai catchment in the Solomon Islands. This catchment provides 40-60% of the water for Honiara, the capital, and management is complex, with sparse data and many competing uses including drinking, domestic, agricultural, and industrial uses for both water and land in the catchment, as well as a range of threats from pollution, increasing population, changing land use, and variable hydrogeology. There are socioeconomic considerations including customary landownership and overlapping institutional responsibilities. A participatory process involving representative stakeholders of three main groups, the customary landowners, the government, and nongovernmental organizations, was central to analyzing the system and building trust in the model development process and model outcomes, and additionally facilitated relationship building between the different groups affecting, and affected by, the catchment. A conceptual model of the Kongulai system with respect to water was developed with all stakeholders. Further elicitation of quantitative aspects took place with a subset of water management professionals for development into a working Bayesian network model. Stakeholder representatives were then presented with the model, some analysis, and scenarios for discussion and feedback. The model provided a number of recommendations that support local management decision making, which were accepted by the wider stakeholder group. The process demonstrates the worth of a well-designed participatory approach to enhance stakeholder contributions and confirms the appropriateness of Bayesian networks for use in developing country contexts where capacity and data may be scarce.

Development of a participatory Bayesian network model for integrating ecosystem services into catchment-scale water resources management

“…Public understanding of the environmental system can help to provide an integrated and 30 qualitative representation of the catchment system, as well as for quantitative modelling (Chan et al, 2010). However, public perspective takes a long time to achieve and carries a large cost, despite detailed documentation available in the relevant literature and as part of local studies and reports (Cain, 2001).…”

“…Many studies have highlighted the importance of participation in system modelling and decision-making, especially in the IWRM setting (Henriksen et al, 2007;Chan et al, 2010;5 Zorrilla et al, 2010;Carmona et al, 2011). While a diverse set of participatory modelling tools exist that can be applied to implement the participatory process, BNs provide a potentially more effective alternative to achieve the goal of the decision support system because of the robustness of Bayes' theory and the visual nature of the software, which facilitates interaction and public participation (Cain, 2001).…”

“…The model assessment tools included qualitative evaluation (e.g., the participatory feedback from stakeholders and experts (Zorrilla et al, 2010)) and quantitative validation (such as the evaluation of predictive accuracy by comparison with observed data or results from other models (Poppenborg and Koellner, 2014), and sensitivity analysis (Kragt, 2009;Chan et al, 2010)). Of these two types of model evaluation tools, sensitivity analysis is widely 20 regarded as the more effective method to assess model performance (Cain, 2001;Bromley et al, 2005;Marcot et al, 2006;Pollino and Henderson, 2010).…”

“…CC-BY 3.0 License. possible decision-making options by using experimental data, model output, or expert knowledge (Cain, 2001;Chan et al, 2010). While a great many of the multidisciplinary approaches available -influence diagrams, decision trees, multi-criteria decision analysis, and so on -are broadly applicable, their employment is harder when it comes to dealing with specific environmental modelling problems because of the complexity and uncertainty involved, and the added 5 complication of stakeholder involvement in the IWRM (Bromley, 2005;Pollino and Henderson, 2010;.…”

A Bayesian belief network approach for assessing uncertainty in conceptual site models at contaminated sites

“…BBNs have also been applied to a wide range of other problems within the field of hydrology and water management. For example, Chan et al [2010] used a BBN for assisting catchment-based water resources management, Wang et al [2009] developed a BBN for assessing and managing farm irrigation systems, while Fienen et al [2013] used a BBN with a numerical groundwater model to study the response of groundwater to sea level rise.…”