From 22 April through 22 May, the Cooperative Atmosphere Surface Exchange Study's 1997 field program (CASES-97) took place in the Walnut River watershed east of Wichita, Kansas. The objectives of the dry part of the experiment included the documentation of the effects of soil-moisture distribution on the structure and evolution of the atmospheric PBL over the 24-hour cycle, the development of a dataset suitable for testing of linked surface-layer boundary-layer parameterization schemes, and the dissemination of the data to other investigators for this purpose. The Principal Investigators were Robert Grossman (visitor, University of Colorado, Boulder), Russell Qualls (University of Colorado, Boulder), Richard Cuenca (Oregon State University), Bruce Hicks (NOAA/ARL), Robert McMillen (NOAA/ATDD), and Margaret LeMone. Julie Lundquist (University of Colorado, Boulder) helped LeMone coordinate the surface-based component of the experiment from Augusta, Kansas. Grossman coordinated the aircraft portion of the experiment from Ponca City, Oklahoma, where the aircraft were based.

Observations (see figure) were designed to integrate surface and PBL processes with free-atmosphere and synoptic-scale processes. Surface heat, moisture, and momentum fluxes and soil conditions were sampled at 12 stations: 10 temporary stations (8 from NCAR/ATD, 1 from NOAA/ATDD, one from CU), and two semipermanent stations (one each from ARM and ABLE). The stations were sited to get as representative a sample of the surface as possible (different crops, etc). Ozone fluxes were sampled at the NOAA site and one of the NCAR sites, where CO₂ fluxes were also sampled. After the experiment, CO₂ fluxes were also sampled at the Argonne site, where data are still being collected as part of ABLE. Periodic soil-moisture profiles were obtained at the two NCAR/ATD sites with NCAR's Atmosphere-Surface Turbulence Exchange Research (ASTER) stations, and soil-moisture data were collected using Cuenca’s time-domain reflectometry (TDR) sensors during the latter part of the experiment from the three Argonne profiler sites. Vertical profiles of wind and virtual temperature were determined continuously at three locations by ABLE, using 915 MHz wind profilers with RASS (200 m to 3+ km), sodars (20-200 m), and surface data (10 m). Cross-Chain LORAN Sounding System (CLASS) radiosonde systems were collocated with the profilers. Two sondes were released, at 1100 and 1400 CST, on most non-IOP days. On the six IOP days, radiosondes were released at 90-minute intervals during most of the daylight hours (one IOP) or around the clock (four IOPs) from all three profiler sites; and during the daylight hours from a single site for one IOP. During IOPs, the NOAA Twin Otter and Wyoming King Air flew coordinated patterns to obtain flux and mean profiles, and horizontal gradients. The S-POL radar, in the vicinity for testing and development of precipitation-estimations algorithms, provided both RHI (Range-Height Indicator) scans and 360-degree scans at several shallow angles to help define PBL structure and windfields. This experiment involved James Wilson (ATD/RAP), Edward Brandes (RAP), and Jothiram Vivekanandan (CAP). Close coordination was also maintained with the JETEX experiment, conducted to study the low-level jet by Ray Arritt and Moti Segal (Iowa State University). Satellite and synoptic data were collected to provide the larger-scale context.

The six IOPs captured a variety of situations, with varying boundary-layer depths, growth rates, and horizontal variability. It is already evident that horizontal variability in surface characteristics is reflected in surface temperature variability and flux statistics. However, other influences on PBL growth make it difficult to isolate the effects of surface characteristics at this early stage. Tony Delany (ATD), who was testing a new instrument, and Jennifer Price, a SOARS student who also helped in the field, did a preliminary analysis of the ozone data. David Yates (ASP) is using data from the CU site to test a land-surface model, and was helped this last summer by Shirley Murillo of SOARS.

LeMone, Chin-Hoh Moeng, L. Jay Miller, and Donald Lenschow (joint appointment with ATD), along with Mingyu Zhou (Center for Marine Environmental Forecasts, Beijing) and Grossman, completed a study to isolate the mechanisms that determine the vertical shear of the horizontal wind in the convective mixed layer. They did so by comparing two days: one with substantial vertical shear through the mixed layer (27 February 1992), and one with almost no shear (10 March 1992). Data are from the 1992 STORM-Fronts Experiments Systems Test (FEST). The approach involved evaluating the terms in the budget equation for the vertical shear of the horizontal wind, namely: time rate of change, differential advection, the Coriolis terms (a thermal wind term and a shear term), and the second height derivative of the vertical transport of horizontal momentum. This evaluation was based on wind profiles from 915 MHz profilers, aircraft, and an NCAR 5-cm radar (in the PBL); wind profiles from radiosondes and aircraft (above the PBL); fluxes from aircraft; horizontal gradients from aircraft; and synoptic data (surface data analyzed at NCAR; analyses above the surface provided through NWP model output). The radar, aircraft, and profilers were all located within the FEST boundary-layer array in northeastern Kansas.

They found that the shear on 27 February was related to the rapid growth of the PBL. Computing the shear budget over a fixed depth (the final depth of the PBL), showed that the time-tendency term dominates, reflecting ingestion of high-shear air from above the PBL. Apparently, this term was sufficiently large that the shear-reduction terms-namely the vertical flux-divergence and differential advection terms-cannot compensate. In contrast, the tendency term was small on 10 March, resulting in the classic balance between the Coriolis terms and the flux-divergence term. While there is considerable uncertainty in the flux-divergence term, the result appears robust. However, the behavior of the wind in rapidly growing PBLs should be examined with other data sets, including that from CASES97, and numerical simulations run to confirm and refine these results.

The Geophysical Turbulence Program held a workshop, Numerical Simulations and Physical Reality, 3-5 September 1997. It focused on the issue: to what extent can numerical modeling overlap with real systems in the sense of predictability and as a description of the phenomenon? Twelve experts were invited to lecture on this topic, drawn from the disciplines of astrophysics, atmospheric chemistry, atmospheric dynamics (both global scale and mesoscale), and turbulence (both atmospheric-oceanographic, and engineering). About thirty-five international experts from the above disciplines participated. The organizing committee for the workshop consisted of Jackson Herring, L. Mark Berliner (GSP, Ohio State University), Thomas Bogdan (HAO), B. C. Low (HAO), Sasha Madronich (ACD), Steven Oncley (ATD), Peter Sullivan, and Joseph Tribbia (CGD).

Herring and Yoshi Kimura (Nagoya University, Japan) continued their research on stratified and rotating turbulence, with special focus on the effects of turbulent structures on the dispersion of particles, and eddy diffusivity. They explored the efficacy of various traditional theories (such as Rapid Distortion Theory) in reproducing quantitatively the results of their Direct Numerical Simulations. They also studied the use of Corrsin's statistical independence principle to obtain necessary information about Lagrangian covariances in the flow.

Recent numerical simulations of decaying two-dimensional flows indicate that they decay in a self-similar manner. Such decay is not predicted by statistical theories of turbulence. Herring, in collaboration with Kimura and Jeffrey Chasnov (Hong Kong University of Science and Technology), are investigating how to justify modifications of the statistical theory so as to obtain self-similar decay. The important physics here is the apparent statistical dependence of dissipation range scales upon inertial range scales. The corollary question is whether the decay rate of enstrophy is sensitive to the dissipative mechanisms at small scales.

Robert Kerr and Herring continued their calculations of high Rayleigh number laboratory convection. A series of calculations at Prandtl numbers .07, .3, .7, 2, 4 and 7 were nearly completed at a Rayleigh number of 10 million in order to determine the Prandtl number dependence of the normalized heat flux, or Rayleigh number. Preliminary results support experimental work showing a strong dependence for Pr<1 and a weak dependence for Pr>1. These simulations bear upon a recent theory of thermal convection by Shiarman and Siggia, which predicts a trend of Nusselt number with Prandtl number, which is contrary to the DNS reported by Kerr and Herring.

Taking advantage of the new Gigaword J9 (Ouray) in SCD, Kerr and Herring extended their earlier work using 256 cubed grid points by carrying out a new series of fundamental turbulence calculations using 512 cubed grid points. The new simulations are primarily designed to follow reconnection events from the time of a near singularity to fully developed turbulence, and to investigate the singular trends of the Euler equations for general initial conditions. So far, the calculations indicate that the previous concerns about resolution at 256³ were addressed.

Sullivan, Moeng, Bjorn Stevens (ASP), Lenschow, and Shane Mayor (University of Wisconsin) analyzed entrainment processes in clear convective PBLs using high-resolution LES. Flow visualization of the simulated flow fields showed that coherent structures in the convective PBL (i.e., thermal plumes) are the primary agents for entrainment. The interaction of plumes with the stable overlying inversion thins the interface locally and induces strong rotational motions inside the plume. At low bulk Richardson number, Ri (i.e., weaker inversion), the rotational motions are strong enough to fold the interface and draw in warmer inversion fluid at the plume's edge, thus causing entrainment. At large Ri, the strong stability of the inversion prevents large-scale folding of the inversion interface and instead strong horizontal and downward motions near the plume's edge pull down pockets of warm air below the nominal inversion height. The structure of the inversion interface from LES was in good agreement with atmospheric measurements obtained from the LIDARs in Flat Terrain (LIFT) field experiment. The normalized entrainment rate from the clear convective PBL simulations agreed well with convection tank measurements. They will next examine entrainment in clear neutral PBLs driven by wind shear.

Lenschow, Moeng, Sullivan, Andreas Muschinski (visitor, Universitat Hannover, Germany) and Steve Cohn (ATD) began a joint effort to explore the feasibility of using four-dimensional (x, y, z, t) data from LES to help calibrate atmospheric radars. In the past, many theoretical, experimental and numerical studies were dedicated to a physical understanding of Doppler-radar returns from clear air. Although volume scatter from turbulent refractive-index fluctuations in clear air is formally well understood, it is difficult to define realistic models for the atmospheric refractive-index fluctuations at the Bragg scales of UHF/VHF radars. Such models, however, are the basis of interpreting radar echoes. The LES technique can simulate space-time distributions of wind, temperature and humidity down to scale lengths that are small compared to the size of the radar's resolution volume but large compared to the radar wavelength. For this study, an extensive database of more than 100 gigabytes of data over a two-hour period from a LES of a free convective PBL is being used. Muschinski used inertial-subrange arguments to parameterize the space-time distribution of the refractive-index turbulence structure parameter within the radar's resolution volume. The time series of the moments of the spectra of the simulated radar returns were calculated and compared to the time series of the parameters characterizing the "known atmosphere" generated by the LES.

Recent field experiments off the coast of California using sophisticated measuring platforms (FLIPs) found strong interactions between the atmospheric PBL and the underlying ocean waves. These observations suggest possible departures from traditional Monin-Obukhov (MO) scaling in the atmospheric surface layer due to the presence of surface waves. Sullivan, Moeng, and James McWilliams (UCLA) examined the ocean-wave effect on boundary layer structure using LES. Their first LES experiment treated the ocean waves as a simple modification to the lower boundary conditions, and the results also showed deviations from MO theory in the situation when the phase speed of the waves is near the wind speed at 10 meters above the waves. Encouraged by these results, they are now undertaking an effort to develop a surface fitted grid capability in the MMM LES code to carry out a more detailed study of wind-wave interaction on both sides of the air-sea interface.

Stevens (ASP) worked with Moeng and Sullivan to understand the sensitivity of entrainment to details in the numerical formulation and sub-grid-scale physics used in LES. Simulations of the classic dry convective boundary layer are known to be rather insensitive to grid resolution, numerics and sub-grid formulations. However, controversy exists on the sensitivity of simulating the cloud-top radiatively driven PBL; among different LES groups, some report sensitivity while others do not. This sensitivity shows up predominantly in the entrainment-rate prediction. Work is currently underway to understand why, and what the implications of this sensitivity are for the role of small scales in entrainment processes.

In collaboration with Moeng and Sullivan, Jeffrey Weil (visitor, University of Colorado) analyzed turbulence profiles from large-eddy simulations (LES) in order to explain earlier dispersion model results in the convective boundary layer (CBL). The results, obtained using LES fields to drive a Lagrangian "particle" model, showed a noticeably slower dispersion rate in moderate than in strong convection, contrary to previous ideas on convective scaling of dispersion. The difference in stability as measured by the ratio of the CBL depth to the Monin-Obukhov length was an order of magnitude. Despite this, the vertical velocity and skewness profiles for the two CBLs exhibited small differences and also differed only slightly from those for free convection (no mean wind). The key difference among the three CBLs was in the dissipation rates, which in the surface layer was 2-3 times larger for moderate than for free convection. This in turn led to much shorter turbulence time scales for moderate convection in the lowest 20 percent of the CBL and hence to the slower dispersion rate.

Recently an effort began between MMM scientists, Mary Barth (joint appointment with ACD), Moeng, Kenneth Davis (University of Minnesota), and Edward Patton (visitor, University of Minnesota) to investigate how boundary layer processes such as chemical reactivity and forest canopy venting, influence ozone production, biogenic hydrocarbon oxidation rates, and the distribution of isoprene and its oxidation products. This study is being performed using LES coupled with chemical reactions. The first simulation coupled a clear convective boundary layer with simple decay chemistry to determine characteristics of ozone-like and isoprene-like tracers. Analysis of this simulation is underway. Future simulations include coupling the clear convective boundary layer with realistic chemical reactions that include non-methane hydrocarbon chemistry and coupling a homogeneous forest canopy with the simple chemical decay mechanism.

Eileen Saiki (ASP), Moeng, and Sullivan investigated the stable boundary layer (SBL) using LES. The SBL is not as well defined as its daytime counterpart, the CBL, and is characterized by intermittent turbulence combined with other phenomena such as gravity waves and low-level jets. One of the objectives of this investigation was to take advantage of nested grids in order to examine fine-scale features of the very stable boundary layer in greater detail. Early work revealed that the current subgrid-scale model does not work well for this particular problem: thus other alternatives are currently being examined.

Ilga Paluch and Lenschow, in collaboration with Christoph Kiemle, Gerhard Ehret, and Andreas Giez (all from the Institute of Atmospheric Physics, German Aerospace Research Establishment [DLR]), and Kenneth Davis (University of Minnesota) examined data collected from the NCAR Electra research aircraft during the Boreal Ecosystem-Atmosphere Study (BOREAS) over the boreal forest of Canada. Typically the height of the clear, convective boundary layer showed fluctuations up to about 13 percent of the PBL depth, as determined from aerosol backscatter from the DLR LIDAR. They examined aircraft soundings and boundary-layer turbulence statistics for cases where the variations form regular wave-like patterns, and cases where the variations are quite irregular. The wave patterns tended to be associated with gravity waves above the inversion. Occasionally, there is evidence of organization in the boundary layer consistent with horizontal roll vortices. This organization is sometimes destroyed by updrafts from irregular hot spots at the surface. The most regular oscillations of the boundary layer height observed in BOREAS were the result of a combination of roll vortices and standing gravity waves with matching wavelength. For this case, the winds above the inversion were nearly perpendicular to the boundary layer winds, and there was a close match between the wavelength of gravity waves, calculated from the temperature profile above the inversion, and the typical spacing and orientation of boundary layer roll vortices.

Jielun Sun (visitor, University of Colorado) collaborated with Lenschow on a study of mesoscale fluxes observed from the NCAR Electra aircraft. For mesoscale models, the desired grid-averaged fluxes include not only turbulent fluxes, but also mesoscale fluxes with horizontal scales less than the grid size but greater than what are normally considered turbulent scales. Well-instrumented aircraft can now resolve these mesoscale contributions to the fluxes. Sun and Lenschow analyzed over 30 flight legs between 228 km and 666 km in length at 100 m altitude over the boreal forest of Canada (BOREAS), and over 20 flight legs between 60 km and 120 km in length at 30 m altitude over the tropical Pacific (TOGA COARE). Preliminary results indicate that the NCAR Electra air motion system has sufficient accuracy to measure the mesoscale flux if the accuracy of the aircraft height measurement is within 3 m over the required 100-1000 s time periods.

Sun also worked with James Howell (ASP) on finding an objective way to calculate fluxes in a stably stratified boundary layer using data from MICROFRONTS. They found that there is significant heat flux divergence even in the layer between 3m and 10m height above the surface. This divergence is partly responsible for the decrease in air temperature at night. The stability functions for MO similarity theory thus depend on the level at which the heat flux is measured. The momentum flux, however, was found to be independent of height, similar to the heat and momentum fluxes in the unstable boundary layer.

Dimethyl sulfide (DMS), produced by phytoplankton, is the dominant sulfur species emitted from the oceans. Its oxidation product, particulate sulfate, is believed to be the primary source of cloud condensation nuclei (CCN) in the remote marine atmosphere. Since global warming has the potential to affect DMS flux and hence CCN and cloud formation and evolution, DMS flux has been targeted as a possible feedback mechanism for global change. Yet despite its importance, simple relationships between DMS flux and variables thought to control it have not been forthcoming. Lenschow and Paluch in collaboration with Alan Bandy and Donald Thornton (Drexel University) used a mixed-layer similarity technique, which relates mixed-layer gradients and variances to surface and entrainment fluxes, to estimate DMS emissions from mean DMS measurements obtained from the NASA P3-B aircraft near Christmas Island during PEM-Tropics. The values they obtained were relatively high, but not incompatible with previous estimates based on less indirect budget estimation techniques.

Analysis of boundary-layer measurements from the NCAR C-130 aircraft during the Aerosol Characterization Experiment (ACE-1) was carried out by Lenschow, Lynn Russell (ASP), Qing Wang (U.S. Naval Postgraduate School) and Krista Laursen (ATD). They analyzed two Lagrangian cases in the marine boundary layer south of Australia during December 1995, where the C-130 was used to probe an air mass that was labeled with constant level balloons (released from a ship) as it advected southeasterly in the region southwest of Tasmania, Australia. Each Lagrangian included three C-130 flights within a one-day period (see figure). A primary goal of the study was to investigate the formation and evolution of aerosols in a remote marine environment far removed from sources of pollution. As part of this effort, it is necessary to characterize the boundary-layer structure in order to better understand the sources and sinks of precursor gas species and particulates in the marine boundary layer (MBL). At the same time, these trace species can also be used as tracers of boundary-layer processes in a Lagrangian experiment. In this way, they were able to document the bidirectional exchange of air between the turbulent boundary layer and an intermittently turbulent layer immediately above, which they labeled the buffer layer. They found that this process has important implications in understanding the evolution of trace constituents in the MBL.

Lenschow, Davis (University of Minnesota), and Sun (University of Colorado, Boulder) participated in the Southern Great Plains Experiment (SGP97), which took place over central Oklahoma during June and July of 1997. The primary objective of SGP97 was to characterize the hydrological cycle and to establish that the retrieval algorithms for surface soil moisture developed at higher spatial resolution using truck- and aircraft-based sensors can be extended to the coarser resolutions expected from satellite platforms. The core of the 1997 experiment involved the deployment of the L-band Electronically Scanned Thinned Array Radiometer (ESTAR) on a NASA P-3 aircraft for daily mapping of surface soil moisture over an area greater than 10,000 km² and a period on the order of a month. In addition, the LIDAR Atmospheric Sensing Experiment (LASE) was also flown on the P-3 to provide continuous water vapor and backscatter soundings along the flight track. In support of these measurements, the National Research Council Canada (NRCC) Twin-Otter and the NOAA/ATDD Long-EZ aircraft made in situ flux measurements. During the coming year, these data sets will be analyzed with the goal of looking at water-vapor budgets, the relation of mesoscale atmospheric variability to soil moisture variability, and the application of the LASE to study boundary-layer structure in a region of varying soil moisture and ground cover.

Alan Hills (joint appointment with ACD), Lenschow, and John Birks (University of Colorado) developed a prototype DMS sensor with sufficiently rapid time response and sensitivity to measure DMS flux in the marine boundary layer. The sensor utilizes the chemiluminescent reaction of fluorine with DMS and senses the resulting radiation with a photomultiplier tube. It appears to be sufficiently selective that it can be used in the remote marine environment without significant interference from other species. Laboratory tests showed that an instrument can be built with sufficient portability and ruggedness that it can be deployed on aircraft or towers.

Results from simulations performed by Jordan Powers using a coupled mesoscale atmosphere-ocean model demonstrated the impacts of different sea state roughness parameterizations and coupling approaches on mesoscale atmospheric forecasts. The coupled model used consists of the MM5, the Princeton Ocean Model, and the GLERL-Donelan Wave Model. In a case involving cold frontal passage across Lake Erie, the results show that wave age and wave motion are significant factors in the atmospheric simulations produced by the coupled model. Forecast accuracy is found to vary with model assumptions of wave age and their relation to the windsea regime. Also found to be important in coupled simulations is the handling of roughness element mobility in estimations of marine surface stresses. Collaborations with scientists at NCEP and Ohio State University in this research are ongoing.

James Bresch (visitor, University of Washington) examined the impact that MM5's different PBL parameterizations have on a simulation of an oceanic cyclone that occurred during ERICA. It was recently shown that MM5 simulations of the ERICA IOP5 storm were sensitive to the type of convective parameterization, but because of these storms' typical insensitivity to surface fluxes and the fact that most PBL schemes produce comparable surface fluxes, it was assumed that mesoscale simulations of such storms were not sensitive to the PBL scheme employed. However, simulations of the ERICA IOP2 storm with MM5 using the Blackadar, MRF, or Burk-Thompson PBL parameterizations gave surprisingly different tracks and central pressures. The Burk-Thompson scheme yielded the best simulation while the Blackadar scheme generated surface fluxes that were too strong and, thus, produced a cyclone that was too strong. In collaboration with Jimy Dudhia, experiments with various roughness lengths were conducted (simulating the effects of a coupled wave model) which showed that with increased roughness, surface fluxes increased resulting in a stronger cyclone unless roughness lengths were greater than 5 cm, in which case surface friction dampened the storm's intensification. Further work is needed to determine how the PBL scheme interacts with the model's precipitation parameterizations to influence cyclone development.

NCAR | UCAR | NSF | NCAR FY96 ASR | NCAR FY97 ASR