n

CONCLUSION. This study provides the first continent-wide position/velocity solution for Africa, expressed in ITRF2008, and is therefore a first contribution to the upcoming African Reference Frame (AFREF). Except for a few regions, the Afri-can continent remains largely under-sampled by continuous space geodetic data. Efforts are in order to augment the geo-detic infrastructure and openly share existing data sets so that the objectives of AFREF can be fully reached. In ad-dition to the geophysical implications of this work, we argue that using a static frame for territories lo-cated on stable Nubia is acceptable because of its low level of internal deformation. However, we also argue that deforming territories in northern and eastern Africa should consider using a semi-dynamic (or prefer-ably dynamic) datum in order to avoid network distor-tions that would be easily detected given the deforma-tion rates involved and the level of preci-sion currently achievable by well-equipped surveyors. The kinematic model proposed here for East Africa can serve to define such a datum. Close up on the East African Rift and the Somalian plate. A. Re-sidual velocities with respect to each plate (only showing sites whose velocity uncertainties are less than 1.5~mm/yr). Residuals are small (less than 1~mm/yr) and show no systematic trend. Grey arrow at site ABPO in Madagascar shows ve-locity with re-spect to So-malia, white arrow shows velocity with respect to Nubia. Site ABPO does not fit either plate. B = Pre-dicted ve-locities with respect to Nubia for the Somalia, Victoria, and Rovuma plates. 0