This thesis reports measurements of turbulent parameters based on the use of the variance method applied to the along beam velocities measured by an acoustic Doppler current profiler (ADCP). Estimates of Reynolds stress ( τ), turbulent kinetic energy production rate (P) and eddy viscosity (N_z) are used to analyse the cycle of turbulence and its relationship to water column stratification in the partially stratified York River estuary, Virginia.
The estimates of Reynolds stress are validated by comparison with direct measurements from an acoustic Doppler velocimeter (ADV). The uncertainties in turbulence estimates using the variance method are analysed, as are improvements
resulting from recent advances in ADCP technology which allow higher ping rates. For weak flows, the uncertainty in the measurements in τ and P arise mainly from instrument noise. For stronger flows, the principal determining parameter is the number of individual independent velocity measurements over which the variance is calculated. These results are validated by detailed statistical analyses of two data sets from a RDI 1200kHz Workhorse ADCP using different ping rates. While increasing ping rate generally reduces the effects of instrument noise, it does not alleviate the influence of flow related noise once the sampling interval is less than the autocovariance time scale of the turbulence.
The evolution of τ  , P and N_z was observed over a spring-neap cycle, and the effect of tidal straining and stratification on these parameters is analysed for two shorter periods. At neaps, τ , P and N_z are significantly affected by the density structure: the highest values are restricted to a thin layer near the bed when the water column is stratified during the ebb, but high values propagate through the water column on the flood. At springs, increased stirring reduces stratification and results in a cycle of τ  and P which resembles that observed in homogeneous conditions. Tidal straining has a marked influence on N_z: higher values occur on the flood with a mid-water column maximum; on the ebb, the maximum is much smaller and occurs nearer the bed.
In a test of the simplified dynamical balance of the York River estuary, τ  is observed to covary with the other terms to give a first order balance, but there is a strong peak in τ  at times of high flow speeds indicating the effect of higher order harmonics. The drag coefficient is estimated using a quadratic drag law, and the drag coefficient is found to increase significantly at higher current speeds.