The Lee and Heghinian’s bayesian approach was applied to CRU TS 3.1 grid precipitation data to detect change points at the Congo watershed scale. The locations that were sensitive to change point have been widely detected during 1969 and have been grouped in two zones that are located mainly in (a) the sub-basins of Bangui and (b) the Kasai and in the Cuvette Centrale. The signal of the persistence over two zones has been estimated at 8 years and 15 years covering respectively the period 1966-1973 (78% of the years on the total area) and 19661-1975 (68% of the years on the total area). Moreover, the change points over mentioned zones are respectively associated with 85% and 77% of the negative values of the shift magnitude. However, about 20.0% and 10.6% of the total area of the Congo watershed were sensitive to change points and the base of precipitation, respectively.

The local or regional approach is often used in studies to detect the point of change of the mean in the hydro-climatic time series. These approaches have been applied by previous studies on a limited part of the Congo basin, in particular on the right bank of the Congo river. This study is a comprehensive spatio-temporal analysis over the Congo Basin that deals with the detection of the point of change of the mean in the precipitation time series using the Lee et Heghinian approach. The latter has been applied on 9021 precipitation time series over the Congo watershed which resulted in a spatio-temporal representation of the probabilities of the point of change of the mean. A clear cut of the study area unto distinctive zones with high or low probabilities has been obtained. The point of change of the mean in the precipitation time series occurred in 1969 or 1970 with high probabilities over two areas of the basin which correspond to the sub-basins of Oubangi and Kasai. These sub-basins are seemingly more vulnerable to climate change than other sub-basins such as the Batéké Plateau sub-basin as well as regions not yet studied.