High-quality bottom boundary layer measurements and bottom photographs were obtained over a sand substrate during a 10-day deployment of the GEOPROBE tripod at an inner shelf (35-m water depth) location off northern California. The seafloor surrounding the tripod was composed of well-sorted medium-grained (mean diameter, 0.25 mm) sand which was formed into symmetrical wave ripples with heights of 3--4 cm and wavelengths of 22--30 cm. Mean velocity profiles in the region from 23 cm to 102 cm above the rippled bed were highly logarithmic (R>0.95) approximately 30% of the time. Nineteen profiles exhibiting R>0.997 were analyzed to obtain the shear velocity (U*c) and roughness length (Z0c) for the mean current. The near-bottom flow field was composed of quasi-steady currents (up to 12 cm s-1 at z=102 cm) and wave-induced, oscillatory currents (up to 14 cm s-1). The data-derived estimates of U*c and Z0c were 0.3--0.93 cm s-1 and 0.82--1.5 cm, respectively. The mean shear estimates are 50--100% larger than those predicted using a drag coefficient (CD) of 3¿10-3 that is typical for rough boundaries, and the roughness lengths are up to an order of magnitude larger than the maximum expected values based on the observed wave-rippled bottom. These results indicate the importance of the combined flow turbulent interaction in producing a large apparent Z0c.

However, comparison of the shear and roughness estimates derived from the velocity profile analysis to predictions made by the combined flow model of Grant and Madsen (1979) show that direct application of the wave-ripple roughness equation of Grant and Madsen (1982) yields large overestimates of z0c and U*c. Selecting the physical roughness length kb(=30z0) that produced the best agreement with the data resulted in z0 values ranging from 0.03 to 0.43 cm. Moreover, a direct correlation exists between these physical roughness estimates and the angle (&thgr;cr) formed by the mean current and the trend of the wave ripple crests. A simple linear relationship between kb and &thgr;cr is suggested by out limited data set.