The NCAR Geophysical Turbulence Program (GTP) within MMM is comprised of Jackson Herring, Robert Kerr, Donald Lenschow, Chin-Hoh Moeng, and Peter Sullivan. Of these, the research of Herring, Kerr, and collaborators is primarily devoted to basic research in turbulence theory and direct numerical simulation (DNS). Kerr's research has been instrumental in developing DNS codes and using them for investigating fundamental issues in turbulence flows and in basic fluid dynamics.
One important component of GTP consists of international workshops on various timely issues in the area of geophysical and astrophysical turbulence and fluid dynamics. In recent years, these have occurred every summer. In the summer of 1995, GTP held a workshop on convection, examining both its geophysical and astrophysical realizations; in 1996 a GTP workshop was held on stably-stratified turbulence. Both will result in reviewed publications of a selected set of the papers presented. The organizing committee of the 1995 workshop was Peter Fox [NCAR's High Altitude Observatory (HAO)]), Kerr, and Moeng; for 1996 it was Paul Charbonneau (HAO), Herring, and Ralph Milliff [NCAR's Climate and Dynamics Division (CGD)].
In addition to permanent staff members, Yoshi Kimura (Nagoya University, Japan) was a collaborator with Herring on stratified turbulence studies. Significant collaborations occurred with Fabian Waleffe [Massachusetts Institute of Technology (MIT)], Leslie Smith (Yale University), and Jeffrey Chasnov (University of Hong Kong) on two-dimensional turbulence.
Kerr investigated small-scale turbulence, both on a fundamental level and for its treatment in atmospheric applications. The fundamental work consisted of large, DNS and analysis of vortex dynamics and thermal convection. In vortex dynamics, new analysis of earlier Euler simulations was consistent with the existence of two dynamical length scales and velocity blowup, in addition to the earlier evidence for vorticity blowup and a finite-time singularity in this simulation. These vortex interactions are similar to vortex interactions in windstorms discussed under Front Range Windstorms in the section on Flow Over Complex Terrain.
Kerr and Herring continued their many high Rayleigh number convection simulations reported this year in the Journal of Fluid Mechanics. The Prandtl numbers (for the ratio of viscosity to thermal diffusivity) used in these simulations are 0.7 for gases, 7.0 for water, and 0.07 for liquid metals. Interesting changes in the convective structure near Rayleigh number of 1 million for the Prandtl number 7 case need to be investigated further. Work continued with Alex Brandenburg (Nordita University, Denmark) on small aspect ratios designed to help interpret the experiment of Belmonte, Tilgner, and Libchaber (1994) and their proposal that a boundary layer length scale for velocity can be backed out of how the frequency spectrum cutoff scales with Rayleigh number and height.
Kimura and Herring published their DNS results on stably-stratified turbulence (Journal of Fluid Dynamics, 1996); their study of stable stratification with rotation is well underway. Included in this study is an investigation of the approach of the system to the quasi-geostrophic limit. The first panel of Figure 1 shows a comparison of enstrophy (mean squared vorticity) contours for the full Navier-Stokes equations with the diagnostic evaluation of enstrophy. This flow is strongly stratified, and rapidly rotating (Froude number of 0.051, and Rossby number of 0.013). We note a similarity between the vorticity of the Navier-Stokes equations with that of the quasi-geostrophic approximation, but the Navier-Stokes system is much more disorganized. On the other hand, for a Froude number of 0.007 and the same Rossby number, the flow structures are much more similar, as shown in the second panel of Figure 1. The study also stresses the effects of stratification and rotation on the dispersal of particles released in the fluid. Here, it is known that the (first order) quasi-geostrophic approximation is unable to give satisfactory results.
Two-dimensional turbulence evolves into compact vortex structures, which become increasingly sparse with time (Basdevant et al. (1981); McWilliams (1985)). On the other hand, Batchelor (1969) formulated laws for the decay of total enstrophy and scale-size distributions based on the paradigm of universal self-similarity. Batchelor's results are at variance with the DNS cited above. The reasons for the discrepancy clearly lie in the strong intermittency of the decay process. Nonetheless, recent DNS of Chasnov (1996) indicate that the decay process is such that the energy spectra evolve self-similarly. Using this fact, we seek a modification of the statistical theory (an extension of Batchelor's formalism), which enforces a self-similarity not otherwise present. The resulting phenomenological theory has the correct scaling laws, and we are collaborating with Jeffrey Chasnov (University of Hong Kong) in comparing the spectra with DNS.
Joseph Prusa (Iowa State University), Piotr Smolarkiewicz, and Rolando Garcia [NCAR's Atmospheric Chemistry Division (ACD)] continued their collaborative study of gravity wave activity in the upper mesosphere and lower thermosphere. A major thrust in this year's effort involved full 3D simulations of breaking waves and spectral analysis of the resulting evolution of the wavenumber field. The most outstanding result was that given a 2D wave forcing source, the primary wavefield remained largely 2D even during vigorous wavebreaking for up to 10 Brunt-Vaisala periods after the onset of breaking. The spectral analysis clearly revealed that the primary wavefield was monochromatic before overturning. As the wavefield became saturated, the first nonlinear effect to appear was a second harmonic in the zonal wavefield due to wave self-interaction. As wave activity increased, other higher harmonics as well as subharmonics appeared. Ultimately breaking caused wave activity to be spread over an extended range of wavenumbers that range from zero to well beyond the evanescent limit. Power spectra for the spanwise wavenumbers showed the classical -5/3 Kolmogorov slope for turbulence in the inertial subrange. Over an extended wavenumber range, the zonal and vertical power spectra showed a slope of -3, which was in strong agreement with their 2D computations (Prusa et al., 1996). This -3 slope was indicative of a buoyancy subrange of turbulence. For wave numbers greater than 2 kmsup-1, a slope of -5/3 appeared, indicating a transition to the inertial subrange. Finally, for the zonal power spectra, a slope of -5/3 occurred for wavenumbers less than that of the fundamental mode of the primary wavefield. This corresponded to a reverse energy cascade that transferred energy from the fundamental mode to larger scales.
This investigation (also noted last year) consisted of analyzing a DNS thermal convection data set by the confidence interval technique to estimate the statistical properties of near-boundary fluctuations in temperature and velocity fields. The strategy was to first investigate the possible benefits of the method by applying it to a known data set, as generated by Kerr's DNS, and then possibly to extend it to meteorological data. This was a collaborative effort among Alex Praskovskaya [NCAR' Research Applications Program (RAP)], Herring, Robert Grossman (University of Colorado, Boulder), and Kerr. A paper on this work is nearing completion.
Ilga Paluch, Lenschow, and Qing Wang (former visitor; currently with U.S. Naval Postgraduate School) analyzed data collected by the NCAR C-130 aircraft on research flights during the Beaufort Arctic Storms Experiment (BASE), which was the atmospheric component of a Canadian GEWEX (Global Energy and Water-cycle Experiment) program focused upon the hydrologic balance of the Mackenzie River. The experiment was conducted over the Beaufort Sea during the fall season, when strong radiative cooling occurs. Data from two flights were examined in detail: one over open water, the other over frozen sea surface with open leads. The stratus-topped boundary layer over open water was found to be quite turbulent, with large temperature and moisture fluxes from the sea surface. These were associated with large temperature and moisture differences at the sea-air interface, produced by the advection of cold air from eastern Alaska over the warmer surface of the unfrozen part of the Beaufort Sea. Such conditions are favorable for the formation of well-mixed, stratus-topped boundary layers. Over the frozen sea the situation was the opposite. Strong radiative cooling of the ice surface produced a very stable temperature stratification that suppressed turbulence and surface fluxes, causing the well-mixed layer to descend below the lowest observation level (40 m). The very stable temperature stratification prevented plumes from open leads from rising to the lowest observation level. Above the areas with many open leads, they found gravity waves, along with an increase in the low-level heat and moisture fluxes. The coexistence of turbulence and gravity waves introduced uncertainties in flux measurements, because the waves generate intermittent turbulence and "apparent fluxes" that are very sensitive to wavelength at the long-wavelength end of the energy spectrum.
Entrainment is important both as a fundamental issue in turbulence physics and because it is poorly represented in simple 1D planetary boundary layer (PBL) models. Sullivan and Moeng analyzed Large Eddy Simulation (LES) results with a focus on understanding entrainment through high-resolution nested-grid LES solutions. Flow visualization and statistical measures were used to identify entrainment and entrainment events for PBLs with varying stratification and shear at the interface. For the case of a buoyancy-driven PBL capped by weak stratification, they identified active entraining events and tracked the time evolution of these events which occured at the edges or valleys of thermal plumes as a result of larger-scale engulfment or overturning processes. Sullivan and Moeng plan to refine and extend their analysis to PBLs driven by shear and radiation with different stratification. They also intend to compare the simulated interfacial structure with lidar measurements taken during Boreal Ecosystem-Atmosphere Experiment (BOREAS), in collaboration with Lenschow, Kenneth Davis (former visitor; currently with University of Minnesota), and Christoph Kiemle and Gerhard Ehret [both of German Aerospace Research Establishment (DLR)].
James McWilliams (University of California, Los Angeles), Sullivan, and Moeng recently completed a study on Langmuir turbulence in the oceanic PBL (to appear in the Journal of Fluid Mechanics, 1996). This near-surface phenomenon occurred when surface water waves were sufficiently strong to produce a Stokes drift in the oceanic PBL. Their work showed that Langmuir cells can substantially alter the mean current and flux profiles in an oceanic PBL. They also produced a video to show the time evolution of 10,000 randomly released surface particles. This animation clearly illustrates the migration into preferred lines as a consequence of the Langmuir cells near the water surface. From the video they were also able to locate and track a merger event when two Langmuir cells interact. Langmuir cell merger events usually occur when cells collide at oblique angles with a cancellation of the interior opposite-sign vortices. The remnant from a collision is a single plus-minus vortex pair (a new Langmuir cell) that moves roughly in the direction of the wind and wave propagation.
Ching-Long Lin (former visitor; currently with NOAA), collaborating with Moeng, McWilliams, and Sullivan, investigated the influence of surface roughness on the dynamics of coherent structures in a neutrally stratified PBL using LES. They found that the temporal evolution of the mean profiles in LES after a sudden change in surface roughness was similar to that observed experimentally downstream of an abrupt roughness change. Quadrant analysis of the vertical momentum flux showed that during the transient phase, ejections and sweeps are significantly altered. Three-dimensional flow visualization revealed that the distribution and strength of coherent vortical structures increase (decrease) in the smooth-to-rough (rough-to-smooth) transition. Also, the correlation between the drag coefficient and flow structures was studied.
In collaboration with Moeng and Sullivan, Jeffrey Weil [visitor, Cooperative Institute for Research in the Environmental Sciences (CIRES)] used velocity fields from large-eddy simulations to model dispersion of passive "particles" in a Lagrangian framework. Calculations of the mean concentration field and displacement statistics were made for a weakly-convective boundary layer (CBL) in which the stability index is -h/L = 0.8 and compared to earlier results for stronger convection, -h/L = 14 and 110; h is the CBL height and L is the Monin-Obukhov length. The new results showed that the concentration contours near elevated sources tilt downwards or descend in much the same manner as for stronger convection, the tilt being due to the vertical velocity skewness. However, once material from an elevated source reaches the surface, it appeared to become temporarily trapped in a near-surface layer and rises from the surface more slowly than with stronger convection. This slow rise was attributed to the greater surface shear and dissipation rate in weak convection and the correspondingly smaller turbulence scales near the surface.
In collaboration with scientists at the Environmental Protection Agency (EPA), Fluid Modeling Facility, Weil designed experiments to model scalar dispersion in a laboratory convection tank. The experiments, conducted by EPA personnel, focused on the mean and fluctuating concentration fields from nonbuoyant or buoyant (point) sources in a simulated convective boundary layer (CBL). The fields were obtained from an ensemble of plume cross sections using video imaging of a fluorescent dye tracer. Initial analyses show that (1) nonbuoyant plumes are well-mixed to a height about 15% greater than the CBL depth (h) found from the heat flux profiles; (2) plume penetration of the elevated inversion increases systematically with the buoyancy flux and is followed by slow reentrainment into the mixed layer; and (3) the scaled lateral dispersion matches field observations, a result unattained in previous experiments. The results will be used to develop and test dispersion models for the CBL. In particular, the results with nonbuoyant tracers will assist in the evaluation of a Lagrangian particle model driven by LES velocity fields.
Margaret LeMone, Moeng, Lenschow, and L. Jay Miller in collaboration with Ming-Yu Zhou (visitor from the National Research Center for Marine Environmental Forecasts, Beijing, China), and Grossman studied the baroclinic PBL, with a focus on the vertical profile of the horizontal wind in the convective PBL. Radiosonde boundary-layer wind profiles for the two days showed good agreement with Doppler radar and aircraft data. Miller found that the radar data provide reliable histories of wind profiles and boundary layer depth, but could not be used to estimate turbulent fluxes or variances because of the contribution of ground clutter and other contaminants to the very weak radar signal. Large-scale horizontal temperature gradients were determined from analyses of surface data, as well as surface and upper-air data produced by the ETA and NGM models, and compared with aircraft-measured gradients. They analyzed two days with similar baroclinity and stability from Stormscale Operational and Research Meteorology-Fronts Experiment Systems Test (STORMFEST) in 1992: one (27 February) with the thermal wind roughly perpendicular to the mean wind direction and significant vertical wind shear, and the other (10 March) with the thermal wind roughly parallel to the mean wind and little vertical wind shear. There also were significant differences in the horizontal transport of horizontal momentum. These cases will provide the basis for an LES study to further quantify the role that the relative orientation of the thermal wind with respect to the mean wind direction plays in the turning of the wind with height.
Jielun Sun (visitor, University of Colorado, Boulder), Larry Mahrt (Affiliate Scientist, Oregon State University), William Blumen and Robert Grossman (both of University of Colorado, Boulder), Nimal Gamage (UCAR), Steven Oncley [NCAR's Atmospheric Technology Division (ATD)], and Lenschow analyzed data collected from microfronts, which took place in Kansas in March 1995. The NCAR/ATD Atmosphere-Surface Turbulent Exchange Research (ASTER) facility was used to study turbulence dissipation in a region of frontogenesis, as well as the application of surface radiation temperature to bulk parameterization formulations of the surface layer. The dataset was processed using a package from Oregon State University for removal of bad or questionable data points and will be made available to the meteorological community.
The Boreal Ecosystem-Atmosphere Study (BOREAS) took place over the boreal forest of Canada during the summer of 1994. The NCAR Electra aircraft played a central role in BOREAS by carrying out flux measurements across the boreal forest from the prairie grassland region at the southwestern edge in central Saskatchewan to the arctic tundra in the Northwest Territories. Fluxes of momentum, latent and sensible heat, carbon dioxide and ozone were measured during three intensive measurement periods spanning the growing season from late May to mid-September. Participants in the Electra field program and subsequent data analyses included Lenschow, Oncley, Kenneth Davis, Teresa Campos (ATD), Sun, Jakob Mann (Risoe National Laboratory, Denmark), Wang, and Shane Mayor (former visitor; currently with University of Wisconsin). An additional collaborative study involved the DLR differential absorption lidar (DIAL) for measuring water vapor concentration and aerosol backscatter which was deployed on the Electra by Gerhard Ehret, Andreas Giez, Christoph Kiemle, and Hans-Georg Schreiber (all of DLR).
Results of the Electra participation are contained in five papers submitted to a special BOREAS issue of the Journal of Geophysical Research. A smooth transition was found in mid-summer photosynthetic activity from the primarily deciduous forests in the southern area through the mainly coniferous regions in the bulk of the boreal forest to rather low values in the subarctic tundra. Spring and late summer values were correspondingly lower, with negligible or slightly negative values in the subarctic tundra. Lakes have a significant effect on the fluxes because of their relatively large areal coverage and their cool daytime temperature compared to the surrounding forest. This modifies boundary-layer development. For example, horizontal pressure differences can be generated that induce mesoscale circulations (lake breezes). The lake effect on area-averaged fluxes sometimes leads to a negative heat transfer coefficient for an averaging scale of several times the lake width.
The DLR lidar was used for measuring vertical cross-sections of both aerosol backscatter and water vapor concentration throughout most of the PBL. The mean and turbulent water vapor structure observed throughout the upper part of the convective PBL compared well with LES, except that mesoscale variability, which was not present in the LES, sometimes contributed to the variance of the lidar-measured depth of the PBL. The depth showed a height variability that was less than expected from simple theory, but was consistent with previous lidar observations. The negative of the ratio of virtual temperature flux at the top of the PBL (the entrainment flux) to that at the surface agreed with the commonly-assumed value of 0.2, which was considerably less than what was estimated in the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE). Humidity flux divergence measurements showed a net drying of the PBL by entrained air during the day. Combined with net warming by sensible heat flux, this meant that the relative humidity decreased considerably through the day. There was some evidence for positive feedback between this drying and stomatal control of transpiration. No significant impact was found of fair-weather cumulus development on surface fluxes.
The DLR lidar was also operated concurrently on the ground with the NCAR 10.6 micron Doppler lidar during two days in late spring of 1994. As reported in a paper submitted to the Journal of Geophysical Research, Giez, Ehret, Ronald Schwiesow (formerly ATD; currently with Ball Brothers), Davis, and Lenschow showed that the data can be used to obtain vertical profiles of water vapor flux by eddy correlation with a vertical resolution of about 100 m. The data showed the diurnal variation in the fluxes in the upper part of the PBL, and were used to calculate both horizontal and vertical length scales for turbulent fluctuations in both water vapor and vertical air velocity. The flux data also resolved the net upward flux of water vapor from the PBL into developing cumulus clouds, and its decrease to negligible values as the clouds matured and dissipated.
During November and December of 1995, Lenschow participated in the Aerosol Characterization Experiment (ACE-1), which is the first in a series of experiments that will quantify the chemical and physical processes controlling the evolution and properties of the atmospheric aerosol relevant to radiative forcing and climate. Lenschow's primary interest was to carry out tests of a strategy he developed to obtain measurements of entrainment velocity at the top of the marine PBL by two independent techniques. The NCAR C-130 airplane was deployed over the remote Pacific Ocean, with most of the flights occurring south of Australia. Preliminary analyses, carried out in collaboration with Steven Siems (Monash University, Australia), Paul Krummel [Commonwealth Scientific and Industrial Research Organization (CSIRO), Australia], and Wang, indicate that entrainment estimates of useful accuracy will be feasible. This is an essential variable for estimating budgets of many trace species in the PBL.
During August and September of 1996, Lenschow participated in the NASA-funded Pacific Exploratory Mission--Tropics (PEM-Tropics), using the NASA P-3B airplane mainly to study the evolution of odd nitrogen and sulfur species in the PBL and lower troposphere over the remote Pacific Ocean. He directed several flights with the goal of estimating budgets of several trace species in the marine PBL.