We propose to use a regional orthogonal Slepian Functions (SF) decomposition to represent and analyse the surface water mass variations sensed by GRACE satellite over the Niger basin. As band-limited Spherical Harmonics (SH) remain not adapted for describing hydrological signals in small geographical areas, Slepian functions can solve the signal energy loss by searching the concentration of information in both spatial and spectral domains of various size and shape. Another advantage of using such regionallydefined basis of orthogonal functions is the drastic reduction of the problem dimensions, so that much faster computation of only a few coefficients, instead of thousands of SH. We propose to use SF to adjust the corresponding 10-day SF coefficients for the semi-arid Niger basin from real geopotential data of the GRACE gravimetry mission. Multiyear time series of the Niger coefficients retrieve the seasonal cycle including a short-wet season every end of summer.