The precipitation regime, associated to the West African Monsoon (WAM), shows short, medium and long-term trends, with a marked intraseasonal variability, as well as interannual and interdecadal. On an intraseasonal timescale, the analysis of the characteristic frequencies of rainfall has showed quasi-peridocities at 5, 15 and 45 days (Janicot and Sultan, 2001).

The knowledge of the complex phenomenology of WAM is, to date, far from a full understanding of the multiple mechanisms driving it, and consequently the quantitative precipitation forecasting (QPF) remains a great challenge to realize. From this point of view, satellite observations can offer a deep insight into the time-space domain of these convective mechanisms.

The last generation of geostationary satellites (METEOSAT Second Generation, MSG) has opened up new possibilities to study and analyze the complexity of the monsoon dynamics, due either to the new spectral channels, or the higher temporal sampling (every 15 minutes) with respect to the old generation.

A four-years climatology (2004-2007) of 15-minutes rainfall intensities, as derived from a multi-sensor precipitation estimation approach, involving geostationary (MSG) and polar (SSMI) satellite data, is used to investigate the dynamical mechanisms driving the African warm-season precipitation episodes.

The seasonal and intraseasonal variability of the African Monsoon for the core (JJA) of the rainy seasons, are firstly inspected by means of suitable Hovmöller analysis on precipitation patterns, to investigate both the longitudinal distribution of rainfall and the zonal component of motion. To quantify the coherence and phase speed of rain streaks, a two-dimensional autocorrelation analysis has been performed, investigating both the zonal-span and duration properties.

The periodicity and phase of precipitation within and beyond the diurnal cycle have been also analyzed, also performing 1-D Discrete Fourier Transform along the time dimension of rainfall intensities occurrences diagrams, to attempt to isolate signals in the power spectra that are associated with diurnal, shorter and longer periods of oscillation.

The proposed method has clearly revealed its capability in detecting firstly and then following the convective precipitation activity, allowing both its quantitative and time-space characterization, accordingly with past studies in literature.

Finally, the intriguing feature of the African diurnal cycle has been completely evidenced by satellite observations, in both time-space properties and characteristic frequencies of the rainfall episodes maxima.