There are few empirical measurements of velocity, shear velocity, sand transport, morphological change on the windward slopes of dunes.This thesis compares field measurements on a barchan dune in Oman with calculations using a mathematical model (FLOWSTAR) and measurements in a wind tunnel. All three techniques demonstrate similar patterns of velocity, confirming the acceleration of flow up the windward slope, deceleration between the crest and brink and significant flow deceleration upwind of and at the toe of the dune. The FLOWSTAR model is unable accurately to predict airflow at the brink and its predictions near the surface are highly susceptible to small-scale terrain irregularities. The measurements of shear velocity in the field and those predicted by the FLOWSTAR model reflect observations of previous studies including the widely reported upwind reduction in shear velocity. Such a reduction in shear velocity upwind of the dune should result in a reduction in sand transport and sand deposition. This is not observed in the field. Wind tunnel modelling using a near-surface pulse-wire probe suggests that the field and FLOWSTAR methods of shear velocity derivation are inadequate. The wind tunnel results exhibit no reduction in shear velocity upwind of or at the toe of the dune. This maintenance of upwind shear stress may be caused by concave (unstable) streamline curvature, which is not taken into account by the field and FLOWSTAR techniques. From this hypothesis, a new model of dune dynamics is developed relying on the establishment of an equilibrium between windward slope morphology, streamline curvature and streamwise acceleration.