River basin management is highly complicated as it addresses a complex social–ecological system. It consists of a large area that crosses many administrative borders, it involves different stakeholder’s views on the problem and its solutions, and knowledge about the river basin system is uncertain and fragmented. Many problems in the river basin are of the unstructured type, in which policy objectives are contested and high uncertainty exists. Collaborative knowledge production is especially suited to establish the connection between science and policy for this type of problem. From boundary spanning theory, it is argued that collaborative knowledge production requires (a) sound process management (boundary spanning process), (b) people in both worlds that are willing to cross the boundaries (‘boundary spanners’) and (c) production of joint knowledge objects for instance a model, a map or joint paper (boundary objects).
In river basin management roughly four groups with different roles and dynamics can be discerned: (1) scientists from various scientific backgrounds, (2) stakeholders with different interests, (3) policymakers from different policy sectors and (4) politicians from different political parties. This implicates multiple boundaries not only between these groups but also within these groups. In two case studies (both complex research projects), the ‘science–science’ boundary and the ‘science–policy’ boundary are explored. From these cases it is recommended that specific boundary spanning processes should be designed and facilitated by professionals who have preferably an education or training in mediation or process management. ‘Boundary spanners’ are the people who can act at both sides of the boundary, either as a scientist or as a policymaker. They should know both ‘worlds’ very well.