Flow structure modification and drag reduction induced by sediment stratification in coastal tidal bottom boundary layers

The Karman-Prandtl logarithm formula of velocity distribution has been widely used to depict flow structures and estimate parameters including bed roughness length and friction velocity in coastal tidal bottom boundary layers. However, the basic assumptions are of homogeneous flow, but the effects of stratification should be taken into account. Here, field observations of hydrodynamics and sediment dynamics were carried out in the well-mixed area outside of the Changjiang estuary, China, during May 31-June 2, 2011. Velocity profiles within the ~2 m bottom layer suggest that the flow structure is prominently influenced when the gradient of suspended sediment concentration reaches O (0.1 kg m−4). The direct application of the log-fit method resulted in a pronounced deviation from the actual boundary parameters. The extended log-fit method was used to represent the tidally averaged vertical flow structure, indicating that the modification of the flow structure diminishes down to the bed, and the calculated Z0 based on the lowest layer close to the bed (0.22–0.38 m above the bed in the observation) approaches the real value. The observed data were also used to examine the modified logarithmic velocity distribution model, in which the flux Richardson number is modified with an empirical constant A to parameterize the stratification effects. It was found that the constant A is approximately 1, which is significantly different from previous studies. Results of a vertical one-dimensional model supported the above findings. The sediment stratification effects should be taken into account, and the logarithmic velocity distribution model tends to be applicable in the water layers very close to the bed. •The stratification effects on flow structure are significant when vertical SSC gradient reaches O (0.1 kg m-4) in the BBL.•The stratification effects diminish to the bed.•The modified logarithmic velocity distribution model is applied with an different empirical constant.