Observations of turbulent mixing are quantified in the near-field region (first several kilometers) of a controlled river flow entering Doubtful Sound, New Zealand. Measurement techniques include direct turbulence measurements using microstructure profilers in both Eulerian and quasi-Lagrangian perspectives and turbulent overturn analysis. Over 500 profiles of vertical microstructure were collected within the inner fjord. With high flow speeds of the surface plume over 2 m/s and some of the strongest stratification (N 2 ∼ 10 −1 s −2 ) recorded in the coastal ocean, turbulent overturning (Thorpe, L T ) scales are used to indirectly infer turbulence dissipation rates ( ) and compared with direct measurements in a quasi-idealized laboratory setting. There exists an inverse relationship between directly measured estimates of and distance from the plume discharge point. Peak values of >10−2 W/kg are observed within 500 m from the discharge point, among the highest turbulence dissipation rates recorded in oceanic shear flows. Thorpe scales range from below 1 cm in the pycnocline to a few meters in the weakly stratified ambient water below the plume. Stratification within the surface layer constrains the vertical dimension, limiting L T . Therefore, L T is an unreliable estimator of L 0 , where L O is the Ozmidov scale and is used to infer . Care must thus be taken when applying this method of overturn analysis in an environment where stratification or physical boundaries limit the vertical extent of overturns.Direct turbulence measurements in estuaries and river plumes have been used to examine stratified-shear flow instabilities as a mechanism of mixing (Geyer et al.