Observations of Langmuir Circulation From FLIP


Jerome A. Smith

Scripps Institution of Oceanography, La Jolla, CA 92093, USA

Email: jasmith@ucsd.edu, web: http://jerry.ucsd.edu/


Langmuir circulation (figure 1) has significance across the marine disciplines. Enhanced deepening and inhibited re-stratification can alter the surface temperature and hence net air-sea exchanges. Organization of bubbles into windrows introduces dramatic sound speed variability and also affects air/sea gas fluxes. Organization of seaweed and plankton affects marine life, including pelagic fisheries. Dispersal by Langmuir circulation is a major component in models for oil-spill tracking and for search-and-rescue operations. To get an adequate picture of the forcing and response of Langmuir circulation (and the wind-mixed layer in general), the observations needed include windstress, directional waves, wave breaking, heat and moisture fluxes, stratification (temperature and salinity profiles), velocity profiles across the mixed layer and thermocline, spacing and orientation of windrows, and a measure of the strength of the circulation (e.g., surface rms velocities). These measurements span both the air/sea interface and the thermocline, and must be maintained continuously for many days to span storms and daily, tidal, and inertial cycles. In addition, the total power requirements exceed that comfortably supplied by batteries or local generation by wind or solar energy. It appears that FLIP is uniquely qualified as a platform from which the required range of measurements may all be made. Some findings concerning the evolution and dynamics of Langmuir circulation that were facilitated by FLIP are reviewed and summarized, focusing on major experiments that took place in 1983, 1986, 1990, 1995, and 2002.

Figure 1. Cartoon of Langmuir circulation, first described by Langmuir [1938].


In the 1980’s, attempts were made to obtain horizontal "transects" of velocity through the middle of the oceanic mixed layer. Although the initial motivation was to help define the direction of propagation of internal waves, it was quickly realized that this provides a useful tool for studying mixed layer motion and Langmuir circulation (LC) [Smith et al., 1987], and also surface waves [Pinkel and Smith, 1987]. Indeed, an "inverted side-scan" system had already been employed to study LC in Loch Ness and the North Sea [Thorpe and Hall, 1983], and the backscatter-intensity-only information provided valuable insight into the kinematics of the circulation and some characteristics of the advection and diffusion of surface-active materials (such as bubbles or, perhaps, oil).

Further development of "surface-skimming" Doppler sonar systems was motivated largely by investigations into the dynamics of the oceanic surface mixed layer. In general, one-dimensional "slab-models" of the mixed layer have performed remarkably well [Pollard et al., 1973; Price et al., 1986; O'Brien et al., 1991; Large et al., 1994; Li et al., 1995]. In these, the erosion rate is driven mainly by inertial velocity shear across the pycnocline [Pollard et al., 1973]. Surface stirring by wind and waves can cause continued slower erosion [Niiler and Krauss, 1977], and inhibits restratification. Surface stirring has been parameterized in the energy budget as proportional to the third power of wind friction velocity: (u*)3. However, the constant of proportionality appears to vary with location, and even at a single location the results are quite scattered. In the late 1970’s, a mechanism for the generation of Langmuir circulation was identified a combination of waves and wind-induced shear [Craik and Leibovich, 1976; Craik, 1977; Leibovich, 1977]. This suggests that where Langmuir circulation is involved, deepening should depend on a combination of wave Stokes’ drift and wind stress [Li et al., 1995; Plueddemann et al., 1996; Smith, 1996; Li and Garrett, 1997]. With great optimism, it was ventured that wave climate variations would explain the remaining variability in mixed layer evolution.

It has also been recognized that the existence of large coherent structures in the mixed layer should have profound impacts on life at the ocean surface [Woodcock, 1944; Stommel, 1949; Woodcock, 1950]. However, the degree of organization (e.g., how long the "streaks" remain coherent in the alongwind direction), and even the direction or orientation of the surface features relative to the wind, remained difficult to estimate objectively, particularly in stormy, cloudy, or foggy conditions. This motivated the development in the mid-1990’s of a phased-array system, permitting acoustic backscatter strength (roughly the bubble density) and Doppler shift (radial velocity) to be mapped every second or so over a pie-shaped area some 450 m radius by 90° in bearing [Smith, 1998].

LC Studies and Horizontal-Looking Doppler Sonars

To investigate wind-mixing at the surface of the ocean, observations of wind stress, waves, stratification, velocity profiles, and surface fields of radial velocity and acoustic backscatter intensity have been made over several month-long experiments over the past couple of decades.

Figure 2. Experiment sites. The ocean depths at the MILDEX, PATCHEX, and SWAPP sites are about 4000 m; depths at the HOME and MBL-1 sites about 1000 m.

The following is a chronological accounting of experiments conducted from the "FLoating Instrument Platform" (FLIP) by our group at SIO (figure 2), together with a few of the more important aspects and findings from each:

MILDEX (1983)

Figure 3. Surface-skimming sonar data. Color contours of (upper left) high-passed Ix; (upper right) high-passed Vx; (lower left) high-passed Iy; and (lower right) high-passed dVy/dy. Contour intervals are 0.25 dB for intensity, 0.5 cm/s for velocity, and 0.0003 s-1 for convergence. [Smith et al., 1987]

Figure 4. Spectra of crosswind convergence. Two-hour mean power spectra of dVy/dy at 1-hour intervals, 8 to 10 Nov., 1983 (times PST). Note the appearance of a peak between 1800 (7) and 2200 (6) on the 9th, progressing to lower wavenumbers with time. Peak wavelengths designated are (4) 180 m; (5) 144 m; (6) 120 m; and (7) 103 m. [From Smith et al., 1987]

Figure 5. Mixed layer depth (MLD) spanning the same time segment as above, showing also some near-tidal frequency oscillations occurring after the 8-hour time segment chosen for close examination of LC activity. [From Smith et al. 1987]

PATCHEX (1986)

Figure 6. An example of crosswind velocities from PATCHEX: an eight hour period during the night of Oct. 3 and 4, 1986. The left column shows near-surface velocities (~1 m scale-depth), and the right shows subsurface measurements, at a nominal depth of ~40 m below the surface. The surface data show coherent features, which "swirl" first toward FLIP (to the left), and then away, as they are carried by the inertial and tidal currents. The subsurface data reveal little structure, consistent with the conceptual picture of Langmuir circulation: there is little crosswind-velocity at mid-depth in a Langmuir "cell."

Figure 7. Acoustic backscatter intensity from the same time-interval as figure 6. The surface intensity (left) shows finer detail than the surface velocity, but with good overall correlation. The subsurface intensity fluctuations (right) also correlate with the surface velocity field (compare the right hand side of this figure with the left hand side of the figure above). These observations are consistent with the interpretation that bubble clouds (or sheets) are carried down from the surface convergence zones by the vertical velocities associated with Langmuir circulation, to mid-depth or more in the mixed layer.

Figure 8. An example of a high-frequency internal wave "packet," from Oct. 6, 1986 (time in PST). Both the velocity (upper) and intensity (lower) fields show a remarkably monochromatic wavetrain, with a period of about 15 minutes. The sloping lines correspond to a phase-speed of about 22 cm/sec relative to FLIP. The mean flow past FLIP at the time was 5 to 10 cm/sec, yielding a net phase speed of about 15 cm/sec relative to the water.

SWAPP (1990)

Figure 9. Density profiles over 0700-1000 PST, March 4. A narrow range of densities are contoured: from 24.78 to 24.855 kg/m3. The contour interval is 0.005 sT units (kg/m3). The superimposed straight line corresponds to a depth increase at a rate of 20 m/h, beginning with 0 m at 0721. Note the "blobs" of denser water appearing within the mixed layer; e.g, at 0820 between 10 and 16 m depth, and at 0850 near 17 m depth. There also appear to be blobs of lighter water below the thermocline, for example just before 0900 at 33 m depth. [from Smith 1992].

Figure 10. Mixed layer depth versus time, over the beginning of a wind event (March 4, 1990). This is estimated from a fixed increase in density over the value at 6 m (the shallowest dependable data point). The results using two increments are shown: 0.02 and 0.015 sT units. Also shown is a reference line, 20 m/h. From 0740 to 0830 the two choices of density jump both yield MLD estimates paralleling this reference. [from Smith 1992].

Figure 11. Weighted mean spacing vs. time, over the same time-span as figure 10. A noise level of 12.5 (cm/s)2 per cpm was subtracted from the spectra of the "beam 2" velocity (most nearly crosswind). The squared spectral densities were used to determine a weighted wavenumber. The same procedure was applied to data from beam 3, and the results combined to estimate a vector wavenumber (true "crosswind" lay between beams 2 and 3); the difference is not significant. This shows the results in terms of the spacing (inverse magnitude of the wavenumber). For reference, a line corresponding to 40 m/h increase is shown. [from Smith 1992].

MBLEX (1995)

Figure 12. Langmuir circulation vacillation: alternating irregular, regular features. In each frame, the left image is the acoustic backscatter intensity (mapping bubble density) and the right image is the Doppler shift (radial velocity component). Four frames 15 min apart during strong forcing conditions are shown. Note how the stripes alternate between well-defined and irregular. The arrows indicate the wind; a 50 m vector (on the image’s scale) represents a 10 m/s wind. North is up. [From Smith 1998.]

Figure 13. RMS radial velocity (black symbols) and intensity (red symbols) associated with LC versus time. Each symbol represents a half-hour average; crosses represent dubious estimates, circles more reliable ones. For scaling and comparison, 0.25Us (blue line) and 0.002W (green line) are also shown. [After Smith 1998.]

Figure 14. The surface velocity scale (Vrms), wind (u*), and surface-wave Stokes' drift (S) vary over several orders of magnitude, and all three are well correlated. To help determine what combination of u* and S is the best predictor for Vrms, log(Vrms/u*) is plotted against log(S/u*). The slope on the log-log plot is about 1.0 for each event, indicating that S alone is best, once LC has formed. The "constant" relating Vrms to S varies from event to event, indicating another parameter is needed. After [Smith, 1999]

HOME (2002)

Figure 15. E. Slater (Sr. Eng., MPL/SIO) and the new 50 kHz array on FLIP.

Figure 16. Surface waves observed with the new sonar. A single frame (ping) of backscatter intensity (left) and radial velocity (right) from the new 50 kHz system. Wind is from lower right to upper left. Roughly 10 s seas going downwind are seen in both fields. Intensity (left) also shows hints of smaller scale streaks parallel to the wind, visible from 200 to 1000 m.

Figure 17. Langmuir circulation. Data from roughly the same time period as in figure 16, but with the surface waves time-averaged out (~40 second smoothing, or 16 pings). Elongated features parallel to the wind are now seen in both the intensity (left) and velocity (right), indicating active Langmuir circulation. Wind is about 10 m/s, from lower right to upper left.

Figure 18. Data similar to figure 17 (40 second smoothing), but for a calm period near the peak tidal flows. High-frequency internal waves are seen to propagate across the field from left (South) to right (North). The signal is seen in both intensity and velocity.

4. Conclusions

The rms crosswind velocity V associated with the surface expression of Langmuir circulation, as measured by the surface-skimming Doppler sonars, provides an objective, quantitative measure of the strength of the flow. Previously, the strength of Langmuir circulation has often been gauged by a subjective judgement of how fast flotsam (e.g. computer cards) distributed on the surface lined up into "windrows," and how regular those rows were. The correspondence between this former subjective and new objective measures was assessed, and a remarkable level of agreement found [Plueddemann et al., 1996].

Scaling of the surface motion was sought as a function of both wind and waves. Theoretical considerations suggested that the near-surface velocity V would scale as (u*Us)1/2, where u* is friction velocity and Us the surface Stokes’ drift due to the waves. However, the observations indicate that V scales with just Us within each storm event, once Langmuir circulation is established. The relation is nonlinear in that a threshold must be set for the existence of Langmuir circulation before it holds. Further, the constant of proportionality varies significantly from one event to another, so averaging over several events destroys the correlation. These two observations suggest that (1) fully developed Langmuir circulation does not scale the same way as in current theories of initial growth, and (2) some additional variable is needed to parameterize this motion. It is suggested that this is related to the ratio of wavelength to mixed layer depth, as parameterized by the "final" or maximal values. Dynamic effects of the near-surface bubble layer could also play a role, as could interactions with the internal wave field. Investigations into these effects continue.

This communication focuses on work from FLIP, and is far from a balanced review of studies into Langmuir circulation. A more comprehensive (but still far from balanced) review may be found in [Smith, 2001].


Craik, A.D.D., The generation of Langmuir circulation by an instability mechanism, J. Fluid Mech., 81, 209-223, 1977.

Craik, A.D.D., and S. Leibovich, A rational model for Langmuir circulations, J. Fluid Mech., 73, 401-426, 1976.

Langmuir, I., Surface motion of water induced by wind, Science, 87 (2250), 119-123, 1938.

Large, W.G., J.C. McWilliams, and S.C. Doney, Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization, Rev. Geophys., 32 (4), 363-403, 1994.

Leibovich, S., Convective instability of stably stratified water in the ocean, J. Fluid Mech., 82, 561-581, 1977.

Li, M., and C. Garrett, Mixed layer deepening due to Langmuir circulation, J. Phys. Oceanogr., 27 (1), 121-132, 1997.

Li, M., K. Zahariev, and C. Garrett, Role of Langmuir circulation in the deepening of the ocean surface mixed layer, Science, 270, 1955-1957, 1995.

Niiler, P.P., and E.B. Krauss, One-dimensional models of the upper ocean, in Modeling and Prediction of the Upper Layers of the Ocean, edited by E.B. Krauss, pp. 143-172, Pergammon, Tarrytown, N. Y., 1977.

O'Brien, M.M., A. Plueddemann, and R.A. Weller, The response of oceanic mixed layer depth to physical forcing - modelled vs observed, Biol. Bull., 181 (2), 360-361, 1991.

Pinkel, R., and J.A. Smith, Open Ocean Surface Wave Measurement Using Doppler Sonar, Journal of Geophysical Research, 92, 12967-12973, 1987.

Pinkel, R., and J.A. Smith, Repeat-sequence coding for improved precision of Doppler sonar and sodar, J. Atmos. Oceanic Technol., 9 (2), 149-163, 1992.

Plueddemann, A.J., J.A. Smith, D.M. Farmer, R.A. Weller, W.R. Crawford, R. Pinkel, S. Vagle, and A. Gnanadesikan, Structure and variability of Langmuir circulation during the surface waves processes program, J. Geophys. Res., 101 (C2), 3525-3543, 1996.

Pollard, R.T., P.B. Rhines, and R. Thompson, The deepening of the wind-mixed layer, Geophys. Fluid Dyn., 4, 381-404, 1973.

Price, J.F., R.A. Weller, and R. Pinkel, Diurnal cycling: Observations and models of the upper ocean response to diurnal heating, cooling, and wind mixing, J. Geophys. Res., 91 (C7), 8411-8427, 1986.

Smith, J.A., Observed growth of Langmuir circulation, J. Geophys. Res., 97, 5651-5664, 1992.

Smith, J.A., Observations of Langmuir circulation, waves, and the mixed layer, in The Air Sea Interface: Radio and Acoustic Sensing, Turbulence, and Wave Dynamics, edited by M.A. Donelan, W.H. Hui, and W.J. Plant, pp. 613-622, Univ. of Toronto Press, Toronto, Ont., 1996.

Smith, J.A., Evolution of Langmuir circulation during a storm, Journal of Geophysical Research, 103 (C6), 12,649-12,668, 1998.

Smith, J.A., Observations of wind, waves, and the mixed layer: the scaling of surface motion, in The Wind-Driven Air-Sea Interface, edited by M.L. Banner, pp. 231-238, University of New South Wales, Sydney, Australia, 1999.

Smith, J.A., Observations and theories of Langmuir circulation: a story of mixing, in Fluid Mechanics and the Environment: Dynamical Approaches, edited by J.L. Lumley, pp. 295-314, Springer, New York, 2001.

Smith, J.A., and R. Pinkel, Improvement of Doppler estimation through repeat sequence coding, in Oceans '91/Proceedings, October 1-3, 1991, Honolulu, Hawaii, USA/Ocean Technologies and Opportunities in the Pacific for the 90's, pp. 977-984, IEEE, Piscataway, N.J., 1991.

Smith, J.A., R. Pinkel, and R. Weller, Velocity structure in the mixed layer during MILDEX, J. Phys. Oceanogr., 17 (4), 425-439, 1987.

Stommel, H., Trajectories of small bodies sinding slowly through convection cells, J. Mar. Res., 8, 24-29, 1949.

Thorpe, S., and A.J. Hall, The characteristics of breaking waves, bubble clouds, and near-surface currents observed using side-scan sonar, Cont. Shelf Res., 1, 353-384, 1983.

Woodcock, A.H., A theory of surface water motion deduced from the wind-induced motion of Physalia, J. Mar. Res., 5 (3), 196-205, 1944.

Woodcock, A.H., Subsurface pelagic sargassum, J. Mar. Res., 9 (2), 77-92, 1950.