ments. As a result, many of the models that link data to concepts in a way that might be useful to managers are not effective. Technology that takes advantage of modem advances in geographic information systems and decision support systems are poorly developed at present. The collection of basic environmental data describing the changing conditions of watersheds is in jeopardy as agencies react to shrinking budgets by eliminating monitoring sites for the hydrologic system. Risk and uncertainty must be adequately accounted for in planning and predictive models. Watershed science in general has yet to develop an effective interface between what we know and how we use that knowledge. Good science is not enough; we need useful science. Watershed management without significant input of new scientific understanding, especially understanding of watershed processes and of the human dimensions, is doomed to inefficiency and eventual loss of credibility; research without input from involved stakeholders and those with real management acumen will always prove less than useful. In the end, watershed management is both institutionally and scientifically complex, and thus inherently difficult to implement.
During most of the mid- to late-20th century, watershed management has been a top-down process, but this approach has led to numerous barriers to effective citizen involvement and to use of locally developed knowledge. A truly effective watershed management effort is most likely to be a bottom-up process, driven largely by citizen concerns about local or regional problems and guided by sound data and information. Successful collaborative planning requires broad participation by those likely to be affected by the outcome. Sometimes these stakeholders are beyond the physical boundaries of the watershed or river basin in question, so that a "problemshed" must be accounted for. In a successful process, 'scientific analysis is married with public participation, ensuring that decisions based on cultural values are informed decisions with respect to likely consequences and a clear understanding of who benefits and who pays.
Organizations for watershed management are most likely to be effective if their structure matches the scale of the problem. Individual local issues related to site planning, for example, should be the purview of local self-organized watershed councils, while larger organizations should deal with broader issues. These larger organizations, however, must include the nested smaller watershed groups within their areas of interest, and must account for downstream interests. A major barrier to effective watershed management for large basins in the past has been limitations on the transfer of powers. The various levels of government in the United States developed historically with specific authorities and powers, and most governmental entities are unlikely to give up those powers to some larger all-encompassing organization. In addition, large federal agencies defined by their topical missions (flood control for the Army Corps of Engineers and water management and delivery for the Bureau of Reclamation) are antithetical to over-arching regional organization. Partnerships among levels of government and various agencies are required for effective watershed management. The era of a