No abstract

A method is proposed to reduce the noise contribution to mean turbulence parameters obtained by 3D acoustic Doppler velocity profiler measurements. It is based on a noise spectrum reconstruction from cross-spectra evaluations of two independent and simultaneous measurements of the same vertical velocity component over the whole water depth. The noise spectra and the noise variances are calculated and removed for the three fluctuating velocity components measured in turbulent, open-channel flow. The corrected turbulence spectra show a Ϫ5/3 slope over the whole inertial subrange delimited by the frequency band of the device, while the uncorrected turbulence spectra have flat high-frequency regions typical for noise effects. This method does not require any hypothesis on the flow characteristics nor does it depend on device-dependent parameters. The corrected profiles of turbulence intensities, turbulent kinetic energy, shear stress, and turbulent energy balance equation terms, such as production, transport, and dissipation, are in better agreement with different semitheoretical formulas and other measurements from the literature than those from the uncorrected data. Combined with the use of a phase array emitter, the proposed correction method allows measurements with a relative error under 10% in the outer flow region. The corrected inner flow region measurements are still affected by errors that may originate from spatial averaging effects within the sample volume due to the high local velocity gradient or the lack of validity of the universal laws concerning turbulence quantities over a rough bed.

No abstract

[1] During the winters of 1998 and 1999, observations were made of the cascading of cold water from the nearshore, shallow ''shelf'' zones and down the sloping sides of Lake Geneva. Cascading starts on the average 10 hours after the onset of surface cooling. The draining cold water descends like a gravity current, and the downslope speed of the head of these slugs of cold water, U, has a mean value of 5.2 cm s À1 , with slugs persisting, on the average, for 8 hours. When the Monin-Obukov length scale at the water surface, L, is negative, implying convection occurs, and d/|L|> 1, where d is the mean shelf depth, the nondimensionalized speed of the front of ''slugs,'' U/b 1/3 is found to be 1.3 ± 0.4, where b is the surface buoyancy flux integrated over the time period from one slug to the next. Each slug is unsteady, the head being followed by several fronts in which the temperature of the current decreases and its thickness increases. These fronts travel faster than the mean flow by a factor of r = 1.38 ± 0.3. Dynamical similarities are found with roll waves observed in turbulent open channel flows. The circulation induced by the cascade is found to give a positive skewness to the time derivatives of near-surface temperature in shallow waters, in contrast with negative values close to the slope. The volume of cold water carried by a slug increases with downslope distance as a consequence of turbulent entrainment and the contribution of convectively unstable plumes from the surface. The average volume carried by the slug across the 21 m depth contour is about 1.9 times the volume of water in shallower water (i.e., that on the shelf between shore and a depth of 21 m), implying that cascading is an efficient means of flushing shelf water. Integrated around the lake the mean total volume flux amounts to 11.5 the average winter river inflow.

We analyzed season-long water level records at 12 stations around the Lake of Geneva (local name Léman) for evidence of internal seiches modified by Coriolis force and compared the results with predictions from a two-layer numerical model with real bottom topography for typical wind situations. Results are also compared with those obtained from current and temperature measurements in the lake. Agreement was satisfactory in all cases. Model predictions and measurements both indicated that only three internal seiche modes are excited: the 1st mode and the 3rd mode, which are Kelvin-seiche oscillations, and the 12th mode, which is a Poincaré seiche. The model, driven by winds from different directions, demonstrates that the wind field, constrained by the local topography, determines which of the modes is generated.

No abstract

No abstract

No abstract