“…The low, 1 Hz sampling frequency represented an important instrumental limitation (Soulsby, 1980), and we were unable to infer specific characteristics of turbulence (i.e., higherorder moments, autocorrelation functions, or power spectra). The 180 s record length allowed for averaging over the passage of several flow structures (Babaeyan-Koopaei et al, 2002), however, and, rather than performing detailed, time-domain analyses of individual measurement locations as in previous studies (e.g., Roy et al, 2004), we used the resulting summary statistics to characterize reach-scale spatial patterns of velocity and turbulence intensity. For the cross-sectional deployment, we approximated the depth-averaged velocity by assuming a logarithmic velocity profile and placing the ADV at 0.6 of the flow depth h where h < 45 cm and at 0.2h and 0.8h where h > 45 cm (Whiting, 2003); summary statistics computed for the two depths were then averaged to provide a single data point for the plan view location.…”