In the 1990s, AfricaRice developed RIDEV crop model to assess irrigated rice phenology and extreme temperature impacts on yields, and defined sowing windows for irrigated rice in Senegal River valley (SRV) and Niger River valley (NRV) that minimized climate risks. Based on recent observations, extension agents comment that "more and more farmers sow out of the recommended sowing windows and get good yield. It seems that the climate changed". Those observations raise the following questions: Did farmers' practices change? If so, why? Is there any relationship with climate or other factors? Did the climate really change? Are recommended sowing windows still valid? To address these issues, AfricaRice, SAED, IER, Hohenheim University, Office du Niger and CIRAD launched an integrative project funded by Agence National de la Recherche (ANR) Changement Environnementaux et Socio en Afrique: Passé, Présent et Futur (ESCAPE) and Climate Change, Agriculture and Food Security (CCAFS, CGIAR Research Program) projects, the final objective of which is to define the most adapted sowing windows for recommended irrigated rice varieties in SRV and NRV in Mali in order to reduce climate risks and maximize production. We will further assess the `residual' risks related to those sowing windows in order to develop insurance tools, and to evaluate future situations using climatic scenarios. To achieve this, we started different research activities: (a) - Focus groups and surveys to study farmers' practices, their evolution, constraints and driving factors, and farmers' perception of climate and its evolution in last years. (b) - Agronomic trials, grain quality and biotechnological analysis to verify whether genetic drift has occurred on the main cropped variety, Sahel 108, in the last 20 years that could explain changes in crop phenology and behavior. Thirty Sahel 108 lines/sources have been compared: 2 original Sahel 108 sources from IRRI, 2 sets of breeder seed from ISRA, 19 sets of foun