A Comparison of Methods for Estimating Reynolds Stress from ADCP Measurements in Wavy Environments

Abstract

urbulent Reynolds stresses are now routinely estimated from acoustic Doppler current profiler (ADCP) measurements in estuaries and tidal channels using the variance method, yet biases due to surface gravity waves limit its use in the coastal ocean. Recent modifications to this method, including spatially filtering velocities to isolate the turbulence from wave velocities and fitting a cospectral model to the below-wave band cospectra, have been used to remove this bias. Individually, each modification performed well for the published test datasets, but a comparative analysis over the range of conditions in the coastal ocean has not yet been performed. This work uses ADCP velocity measurements from five previously published coastal ocean and estuarine datasets, which span a range of wave and current conditions as well as instrument configurations, to directly compare methods for estimating stresses in the presence of waves. The computed stresses from each were compared to bottom stress estimates from a quadratic drag law and, where available, estimates of wind stress. These comparisons, along with an analysis of the cospectra, indicated that spectral fitting performs well when the wave climate is wide-banded and/or multidirectional as well as when instrument noise is high. In contrast, spatial filtering performs better when waves are narrow-banded, low frequency, and when wave orbital velocities are strong relative to currents. However, as spatial filtering uses vertically separated velocity bins to remove the wave bias, spectral fitting is able to resolve stresses over a larger fraction of the water column.