Abstract
The economies of the sub-Saharan countries are based mainly their on rainfed agriculture. Medium- and long-term demographic projections show that population pressure will increase dramatically. In the future, greater emphasis will need to be given to the development of agricultural strategies, which, in turn, will require a better understanding of future water resources. To this end, climate and hydrological scenarios need to be developed involving different actors skilled in a range of activities, not only scientists in various fields (climatologists, hydrologists, sociologists,.), but also financial, political and social actors and decision makers. The process of developing scenarios is accompanied by many uncertainties, more or less understood, to which project managers and regional development planners must adapt. Illustrated by the example of the Bani River basin, a tributary of the Niger River, this study discusses the setting of "hydrological scenarios" and the predetermination of parameters of a hydrological model in a non-stationary global context. The choice of parameters has many impacts on the results of different scenarios, for instance for water resource assessment and the variability of water resources. These two factors are fundamental in determining who should lead the decision-making process in response to a request for a management or a development strategy. A methodology is proposed for determining parameter sets based on mixing several sets of parameters derived by moving calibration, and is in line with recommendations of an increasing number of managers: not to try to predict the future, but to envisage several possible futures that force impact models, thus providing managers with decision elements either to adapt to, or to use for adapting new infrastructure solutions to future conditions.