Past attempts to manage the Senegal River to improve the national economies of West African nations have generally not been successful. Analysis of the problem requires an understanding of a large number of intertwining variables that are difficult to quantify. The authors demonstrate the use of a Bayesian network to investigate the effects of three policy scenarios: emphasis on improving macro-economies, emphasis on improving rural livelihoods, and "integrated water resources management." The authors also discuss the variables involved in their modeling approach.