The severe drought that affects the Sahel since the late 1960s has been very closely studied and monitored during the last three decades. Recently, after several wet years, it was questioned from a statistical point of view whether the drought was over. The conclusions of a recent study were that the rainfall deficit was not over at the end of 2000 and that the drought continues. The analysis of the change points in the station rainfall time series suggests differentiating these findings. There is now growing evidence that there is a potential shift towards a more humid state. However, the present analysis shows that the assumption that a significant increase in rainfall may have occurred around the early 1990s could only be verified at the customary confidence level in about 10 years from now. Résumé La sécheresse qui touche la région sahélienne depuis la fin des années soixante a été extrêmement bien étudiée et suivie au cours des trois dernières décennies. Récemment, à la suite de quelques années fortement pluvieuses, certaines recherches ont été menées pour tenter de voir si cette sécheresse était statistiquement terminée. Les conclusions d'une récente étude montrent que ce déficit pluviométrique n'est pas terminé en fin 2000 et que la sécheresse continue. Sur la base de l'analyse de la première rupture pluviométrique, nous pensons qu'il est nécessaire de nuancer ces propos. Plusieurs signes suggèrent qu'une tendance vers une période plus humide pourrait s'être amorcée aux alentours des années 1990. Cependant, la continuité de la période de sécheresse ou l'identification d'une rupture pluviométrique vers des conditions plus humides ne pourra se vérifier statistiquement que dans une dizaine d'années.

After about two decades of dramatic rainfall deficits that started in the late 1960s, the Sahel of West Africa has experienced increasing precipitation since the early 1990s. The implementation of the United Nations Convention to Combat Desertification (UNCCD) needs the identification of areas that record declining vegetation productivity over long‐time periods. In this scope, we analyse the state of the vegetation productivity using long‐term time series of NOAA AVHRR NDVI data and compare it to rainfall data. For this, 128 rain gauge data (RR) were compared with the integrated NDVI during the growing period (iNDVI) values in Burkina Faso from 1982 to 1999. During the analysed period, most of the studied stations in the country were stable for the iNDVI/RR (57.8%). However, 39.8% showed a weak to strong negative trend in the iNDVI/RR while only 2.4% showed a weak positive trend. These negative trends may reflect ongoing desertification processes in Burkina Faso and could be a starting point for the identification of hot‐spots areas to determine where to take action to combat desertification.