Structure of turbulence in the marine atmospheric surface layer

TR Number
Date
1995-12-09
Journal Title
Journal ISSN
Volume Title
Publisher
Virginia Tech
Abstract

Turbulence research in the laboratory has confirmed the existence of quasi-coherent structures amidst the chaos of a turbulent boundary layer. It has been observed that a quasi-periodic phenomena called “bursting” accounts for a major contribution to the turbulent Reynolds stress and the production of turbulent kinetic energy. Bursting is the term used for a sequence of events, where a low-speed streak of fluid from the near wall region lifts away from the wall, slowly at first, and then rapidly moves away from the wall as it convects downstream where it becomes unstable and breaks up violently upon interaction with the outer flow. This ejection of low speed fluid into the mean flow is responsible for locally high values of turbulent kinetic energy. Although a great deal is known about these structures in laboratory flows, little has been done to investigate if such structures are universal in turbulent flows, i.e., their existence in large Reynolds number flows such as the turbulent air flow over the ocean. It would seem, intuitively, that such structures, if present in the marine atmospheric boundary layer, would play a major role in the transfer of momentum, mass and heat across the air-sea interface. It is speculated that these motions may also be associated with large scale organized motions in wall bounded turbulent shear flows. The effort aimed at elucidating the physics underlying such structures would be invaluable in contributing to our understanding of the air-sea flux mechanism.

In this dissertation, standard ejection detection schemes like the quadrant, the VITA and the modified u-level techniques have been applied to turbulent wind data measured over the ocean to confirm the existence of burst like structures. The proportions of contributions to the Reynolds stress from the four quadrants of the u’w’ plane are in close agreement with the corresponding contributions for a laboratory flow. Ejection detection followed by the grouping of ejections into bursts yielded a mean burst period of 47 s, at a height of 8.2 m above the water surface, where the mean wind velocity was 6.74 m/s. This burst period corresponds well with the peaks obtained from the autocorrelation of the streamwise velocity signal and the first moment of the stress spectrum, confirming the quasi-periodic nature of this phenomena. Furthermore, phase averages of these events show a structure which is similar to the structure of events detected in laboratory flows.

The ejection periods are seen to decrease with increasing wind speed. The burst periods decrease at first with increasing wind speed and then appear to attain a constant value after a wind speed of 6-7 m/s. This has been attributed to the breakdown of the grouping algorithm at higher wind speeds. Ejection and burst frequencies exhibit no discernible dependence on the surface wave field.

Ejection and sweep motions have been studied at various length scales. The original velocity signal is bandpass filtered for various frequency bands. For each band, the percentage contributions to the Reynolds stress from the quadrants of the u'w’ plane are close to the corresponding quadrant contributions of the other bands. This indicates similar turbulence structure at different scales. The velocity signals for each band have been normalized by their root mean square (RMS) value. Visualizing the signals on nondimensional time shows the signals from each band to be very similar. These results can also be interpreted as evidence for the ejection and sweep motions existing simultaneously at different scales, indicating the fractal nature of these events.

Large scale motions, which appear to be associated with ejection and sweep motions, have been identified in the marine atmospheric surface layer using velocity probe measurements at multiple heights. Visualizing these velocity signals suggests that the organized features extend across the depth of the surface layer. Converting the temporal signals to spatial fluctuations suggests that these structures are inclined at an angle while convecting downstream. The inclination angle near the surface (z < 18 m) is approximately 15° and it increases with increasing height to about 45° when z = 45 m. The spatial velocity fluctuations also indicate that these organized features may be large transverse vortical arches.

Description
Keywords
ejection motions, velocity signals, sweep motions
Citation