We present a strategy for calculating uplift rates as a function of space and time from large sets oflongitudinal river profiles. This strategy assumes that the shape of a river profile is controlled by thehistory of uplift rate and moderated by the erosional process. We assume that upstream drainage areais invariant. The algorithm was tested on a set of 100 river profiles which were extracted from a digitalelevation model of Madagascar. This set of profiles was simultaneously inverted to obtain uplift rate as a smooth function of space and time. The fit between observed and calculated profiles is excellent and suggests that Madagascar was uplifted by 1–2 km at rates of 0.2–0.4 mm/yr during the last 15 Myrs.The location of Madagascar suggests that its topographic elevation is maintained by convective circulation of the sub-lithospheric mantle. Residual depth anomalies of oceanic fragments encompassing the island show that the island straddles a dynamic topographic gradient which generates asymmetric Neogene uplift. Volcanism, warped peneplains and uplifted marine terraces corroborate the existence of youthful uplift.We suggest that sets of longitudinal river profiles contain useful information about the history of regional uplift which can be extracted by inverse modeling and calibrated by independent geologic observations.