This paper presents the first phase of an interdisciplinary study for the management of Kongulai catchment, a primary water source providing up to 60% of the water supply for Honiara, the capital of the Solomon Islands. There are complex economic, social and ecological factors involved, including multiple uses for both water and catchment and many potential threats to the water supply, including logging, agriculture, pathogens, unstable geology and the institutional arrangements. In such cases a risk-based approach is widely recommended (Rothstein et al. 2006, Slater and Jones 1999). The approach used here is based on catchment-based Ecological Risk Assessment (e. g. Hart et al. 2003, Hart et al. 2006), and was used to guide development of a model to quantitatively assess threats and hazards to water in the catchment. There are widespread and ingenious applications of modelling to a variety of real-world situations, but transfer of modelling technologies to managers and real-world application of these models to improve management remains a challenge (e. g. McIntosh et al. 2005, Matthews et al. 2006). Increased involvement of stakeholders, both the managers and the communities who are part of or live in the system being modelled, has been widely suggested as a key requirement for closing the gap between model development and use (Jakeman et al. 2003). For this study, participatory approaches were used from the initial problem formulation phase and involved local water managers, a number of government departments and other stakeholders including customary landowners and non-governmental organisations. This process has been emphasized in constructing models intended for use in management decision-making (van Delden and Engelen 2006). There is a need for a quantitative tool to integrate the complex factors involved and prioritise management actions. Because of the sparse data, high uncertainty and incomplete understanding of the system, and because of its intuitive graphical basis (Bromley et al. 2005, Welp et al. 2006), Bayesian techniques were considered appropriate and a Bayesian network model is being developed. An additional consideration is that the development context of the Pacific region requires care to ensure any model-based tool is appropriate. There is a lack of data and restricted human and financial resources. Conceptual maps of the primary factors affecting the key water value were elicited from stakeholders and used to discuss water resource issues. This is also being used as the basis for a quantitative model. Ongoing consultation with the managers aims to display underlying assumptions and keep model construction transparent. Although the managers are receptive to the potential benefits of the model, the extent to which it will be applied is not yet clear. At minimum, outcomes from the model will enable application to the government or aid agencies for resources to quantify important risks to the system. Additional benefits would derive from ongoing model use by managers to indicate where limited resources would best be applied. The primary barriers to uptake are resource scarcity for training in the use and updating of the model, and institutional inertia. Additionally, building a relationship with the stakeholders as well as facilitating the relationship between stakeholder groups will be required for successful management.
