Successful inverse modeling of observed longitudinal river profiles suggests that fluvial landscapes are responsive to continent-wide tectonic forcing. However, inversion algorithms make simplifying assumptions about landscape erodibility and drainage planform stability that require careful justification. For example, precipitation rate and drainage catchment area are usually assumed to be invariant. Here, we exploit a closed-loop modeling strategy by inverting drainage networks generated by dynamic landscape simulations in order to investigate the validity of these assumptions. First, we invert 4,018 African river profiles to determine an uplift history that is independently calibrated, and subsequently validated, using separate suites of geologic observations. Second, we use this tectonic forcing to drive landscape simulations that permit divide migration, interfluvial erosion and changes in catchment size. These simulations reproduce large-scale features of the African landscape, including growth of deltaic deposits. Third, the influence of variable precipitation is investigated by carrying out a series of increasingly severe tests. Inverse modeling of drainage inventories extracted from simulated landscapes can largely recover tectonic forcing. Our closed-loop modeling strategy suggests that large-scale tectonic forcing plays the primary role in landscape evolution. One corollary of the integrative solution of the stream-power equation is that precipitation rate becomes influential only if it varies on time scales longer than ∼1 Ma. We conclude that calibrated inverse modeling of river profiles is a fruitful method for investigating landscape evolution and for testing source-to-sink models.