It is argued in this commentary that, in order to understand better the physical mechanisms that generate boundary shear stress over water-worked gravel beds, fl ow velocity data should be re-evaluated by spatial averaging the Reynolds equations to produce time-and space-averaged (double-averaged) momentum equations. A series of laboratory experiments were conducted in which the fl ow velocities were measured using a PIV system over two water-worked gravel deposits. Combined with detailed data on the bed surface topography and vertical porosity, the physical components of shear stress were obtained. This enabled the various momentum transfer mechanisms present above, within and at the interface of a porous, fl uvial deposit, to be quantifi ed. This included the examination of the relevant contributions of temporal and spatial fl uctuations in velocity and surface drag to the overall momentum transfer. It is demonstrated that double-averaging represents a logical framework for assessing the fl uid forces responsible for sediment entrainment and for investigating intragravel fl ow and sediment-water interface exchange mechanisms within the roughness layer in water-worked gravel deposits. By considering the physical components of shear stress and their relative sizes it was possible to provide a physically based explanation for existing observations of enhanced mobility of gravel-sand mixtures and the transfer of solutes into porous, gravel deposits. This analysis reveals the importance of obtaining co-located, high quality spatial data on the fl ow fi eld and bed surface topography in order to gain a physical understanding of the mechanisms which generate boundary shear stress.