CASES-99
Overview of Experimental Design
July 21, 1998
Edited by: Bill Blumen1, Dave Fritts2 and Greg Poulos2
1 University
of Colorado, Boulder, Program in Atmospheric and Oceanic Sciences (PAOS)
2 Colorado
Research Associates (CoRA), Boulder, Colorado
Contributing Authors
Ray Arritt, Ben Balsley, Bob Banta, Dan Cooper, Rich Coulter, Joan Cuxart, Henk deBruin, Rod Frehlich, Nimal Gamage, Reginald Hill, Mike Jensen, Jerry Klazura, Peggy LeMone, Don Lenschow, Larry Mahrt, Torben Mikkelsen, Chin-Hoh Moeng, Jim Moore, Andreas Muschinski, Carmen Nappo, Steve Oncley, Holly Peterson, Antti Piironen, John Prueger, Russ Qualls, Zbignew Sorbjan, Boba Stankov, Jeilun Sun, Marv Wesely
Cover Figure: This panel shows 4 m resolution data collected with the U.S. Army's 2.9 MHz FM-CW radar during routine and continuous data collection at the White Sands Missile Range in New Mexico. The top panel exhibits a height-time cross-section of the clear-air refractive index turbulence structure parameter, Cn2, on 12 September 1995 between 0101 and 0232 UTC, with the height ranging from ground to 2800 m AGL. It shows a boundary layer in transition from daytime to nighttime conditions with a wave like structure at ~2 km height, broken patches underneath, and the onset of the stable boundary layer at the right-hand side that is indicated mostly by the increase in insect population at night. The bottom panel exhibits a height-time cross-section of Cn2 for 12 December 1995 between 0545 and 0716 UTC and the height ranging from the ground to 2000 m AGL. These observations provide evidence of the non-stationary and inhomogeneous nature of small-scale turbulence within the stable boundary layers. Layered structure of the nocturnal boundary layer (NBL) is seen, with a persisting breaking wave or shear instability at ~600 m AGL throughout the entire period. Starting at 0650 UTC another thin breaking event appears at the 700 m AGL level. Both events exhibited crest-to-trough displacement of less than 150 m and oscillations of ~5 min period. Some NBLs, such as shown by Eaton et al. (1995), show much larger amplitude gravity waves and Kelvin-Helmholtz billows. The vertical lines occur when a large target (e.g., a bird) flies through or close to the main beam. The point targets are believed to be insects and birds.
CASES-99 will combine measurements and data analyses with state-of-the-art numerical modeling to investigate 5 areas of scientific interest. The choice of these scientific topics is motivated by both the need to delineate physical processes that characterize the stable boundary layer, which are as yet not clearly understood (see Nappo and Johansson, 1998), and the specific scientific goals of the investigators who are expected to participate in the observational, analysis and numerical modeling activities of CASES-99 (See Appendices A and B). The field program would use the CASES instrumented site within the Walnut River Watershed (WRW, see Figure 1), including the Department of Energy ARM-CART, the Argonne National Laboratory Atmospheric Boundary Layer Experiments (ANL ABLE Department of Energy), and the NOAA Wind Profiler Network in south central Kansas, together with enhanced instrumentation, for one month of extensive measurements during October 1999. More information on the site can be found at the CASES 1997 website at http://www.mmm.ucar.edu/cases/ and from ABLE program information located at http://www.atmos.anl.gov/ABLE. Although the study area is the same as that for CASES-97 as shown in Figure 1, our goals are considerably different. The experimental site is a watershed, but is actually quite flat (relief is approximately 30 meters across 60 km), and is climatologically favored for clear sky and weakly stable to very stable conditions during autumn. Here, `weakly stable' means that turbulence is continuous, Monin-Obukhov theory is valid and 0.25 < Ri < 1.0, whereas `very stable' means that turbulence is intermittent, Monin-Obukhov theory is not valid and generally Ri > 1.0 (Mahrt 1998).
Figure 1 (above):
A large-scale plan view showing the location of the CASES field site or
Walnut River Watershed (hatched) within the Arkansas-Red River Drainage
Basin. Figure 2 shows the WRW in detail and Figure 3 shows a preliminary
instrument deployment for CASES-99 within the WRW.
Figure 2 (above): The Walnut River Watershed and site of CASES-99 field program (within red [grey if in greyscale] boundaries). The potential area of focussed instrumentation for CASES-99 is outlined in the small rectangle just to the right of the Central Site, southeast of Leon, KS along Route 96 (see Figure 3 for details). The filled circles indicate areas of existing, meso-b scale instrumentation to be supplied by ABLE (towers, sodars, wind profilers with RASS).
The nonstationarity associated with gravity wave instabilities, overturning Kelvin-Helmholtz (KH) billows, terrain-generated phenomena, surface heterogeneity and heat and radiative flux divergences contributes to the uncertainties and conceptual difficulties encountered in the various attempts to construct a physical basis for events and concomitant vertical transports that occur under statically stable regimes. Most studies to date have not been able to establish the source(s) of intermittent turbulence that is (are) often observed at ground level. This lack of knowledge inhibits the development of reliable surface layer and subgrid-scale turbulence parameterizations of the very stable nighttime boundary layer regime, and numerical modelers have been forced to tolerate these errors to the detriment of simulation accuracy. This problem is discussed in depth, along with measurement strategies, by Wyngaard and Peltier (1996). A few efforts have attempted to identify the source(s) of errors in surface layer parameterizations for stable flows. It is argued by Poulos (1996) that a numerical oscillation created in the stable surface layer parameterization can induce occasionally unrealistic cooling under low wind conditions. The resulting gradient is enhanced by the turbulence parameterization and, in some cases, "runaway" cooling can occur. This undesirable effect is also discussed by Mahrt (1998) to be the result of radiatively driven heat loss that is not sufficiently compensated for by the heat flux calculated in the stable surface layer parameterization.
The goal of the proposed one month field program is to identify the sources and quantify the physical characteristics of the mixing phenomena that populate the stable boundary layers, using an optimal arrangement of a variety of observational tools. These phenomena include internal gravity waves, Kelvin-Helmholtz billows, inertial oscillations, shallow drainage flows and turbulent bursts that occur intermittently under clear sky and light wind conditions. The program of data analysis and numerical simulations that follow will then focus on establishment of the relative contributions to surface layer heat, moisture and momentum fluxes and flux divergences associated with the presence of these phenomena. The difficulty of making observations of stably stratified atmospheric conditions, including both instrumentation limitations and the spatial and temporal distribution of the sensors themselves is deferred to later sections.
1) provide a time history of internal gravity waves, KH shear instabilities, and turbulence events in the nighttime stable boundary layer, and to evaluate the relative contributions to intermittent heat, moisture and momentum fluxes that can be associated with these phenomena. Sources of turbulence bursts include, but are not restricted to, surface and elevated shear layers and KH instability, internal gravity waves within the stable boundary layer, drainage currents, and surface vortex shedding.
2) measure heat and momentum fluxes and their divergences accompanying the events contributing to turbulence, transports, and mixing throughout the nocturnal boundary layer, and especially within the surface layer (~ 10 to 20 m), to assess the departures from similarity theory under weakly stable and very stable conditions.
3) define the relative importance of surface heterogeneity, particularly under very stable light wind conditions, on the initiation of shallow drainage currents (O [10m]), and the horizontal and vertical transports that accompany such boundary undulations.
4) improve our current understanding of the diffusion, dispersion, meandering and concentration fluctuations of plumes that emanate from ground-based and, possibly, elevated sources, during both clear and cloud-topped nocturnal boundary layer conditions.
5) acquire data during the transition from a convective to a stable boundary layer regime and vice-versa to compare with existing models of this transition, and to assess the role of this transition period in the initiation of inertial oscillations and the enhancement of low-level jets ~ 100-300 m above the surface.
Unstable shear flows occur under stable NBL conditions for several reasons. Perhaps most prevalent are the shears capping the inversion layer (Thorpe and Guymer, 1977), which arise due to frictional decoupling of the boundary from the quasi-geostrophic flow at greater heights accompanying the transition from an unstable (convective) daytime to a stable nocturnal boundary layer. This allows the development of a super-geostrophic jet at lower levels (e.g., the Great Plains low-level jet, Blackadar, 1957; Zhong et al., 1996) and the occurrence of shear instability, including gravity wave excitation, KH instability, and turbulence generation at these altitudes. Examples of these dynamics include the FM-CW radar measurements by Eaton et al. (1995) using the U.S. Army Dugway radar, which is expected to be available for the CASES-99 field program (see also the cover of this document). The occurrence of turbulence bursts then depends on the evolving shear flow and the tendency for KH instability to occur and trigger a transition to turbulence, as suggested in recent laboratory and modeling studies (Thorpe, 1985; Fritts et al., 1996b; Palmer et al., 1996; Werne and Fritts, 1998).
Other sources of instability and mixing include strongly sheared surface flows, vortex generation and separation due to surface friction, the interface between drainage flows and the NBL structure above, and the layering of more and less stable layers within the NBL (Gossard et al., 1985), which imply instability at the density interfaces (Fritts and Rastogi, 1985; Simpson, 1997). In these latter cases, mixing and transport accompanying such events will often be initiated at the lowest levels of, or within, the NBL.
A key goal of the CASES-99 field program will be to define the characteristics and statistics of episodic turbulence bursts due to these various internal gravity wave and shear sources and to quantify their contributions to transport and mixing within the stable nocturnal boundary layer. Knowledge of such instability processes has progressed slowly in the last few decades due to instrumentation and modeling limitations, except as noted above. However, the atmospheric science community is now in a position to merge state-of-the-art observational and numerical capabilities to dramatically advance the understanding of the dynamics, statistics, and consequences of turbulence bursts in the NBL and to describe their influences via more quantitative subgrid-scale and surface layer parameterizations in large-scale models of these flows. Specifically, ground-based and in-situ (see Section 3) instrumentation is now able to define both the large-scale wave and KH (quasi-two-dimensional) and turbulence (three-dimensional) fluxes of momentum and heat experimentally, while recent modeling studies have allowed the evolution of and transport processes accompanying instability dynamics to be assessed for the first time (Werne and Fritts, 1998, see Figure 4).
Monin-Obukhov similarity theory has been extended
to stable conditions by using local turbulence scales (Wyngaard, 1973,
1994; Nieuwstadt, 1984a,b; Lenschow et al., 1988a). The vertical size of
the turbulent eddies in the stable boundary layer is restricted and therefore
the turbulence is minimally influenced by the surface and the turbulent
scaling laws do not depend on the height above the surface. This has been
called z-less stratification (Wyngaard, 1973, 1994). The local scaling
laws of the temperature and velocity structure constants, CT2 and Cv2,
were developed for stable conditions by Wyngaard (1994) and compared with
results from data and LES. Wyngaard noted that the scaling for structure
constants was robust because these statistics measure the small-scale turbulent
processes which have more statistical reliability. Reliable measurements
of
CT2 and Cv2 are possible in the nocturnal jet using kite-borne turbulence
probes (Balsley et al., 1998) and by FM-CW radar for CT2 only. The central
region of the nocturnal jet should be stable and the predictions of these
local scaling laws can be tested.
In the absence of cold air advection, the vertical
divergence of the sensible and radiative heat flux must account for the
observed, oftentimes large, cooling in stratified boundary layers. The
measurements of the vertical divergence of the sensible heat flux must
be sufficiently resolved to contend with the decrease of transport scale
close to the ground surface. Without high-frequency information, a portion
of the flux closer to the ground surface could be eliminated, leading to
underestimation of the surface heat flux (Howell and Sun, 1998). An analysis
of this possible flux loss is required. High frequency response instruments
(approaching 50 Hz) are desirable for measurements within ~ 5 m of the
surface and will hopefully be available for CASES-99. Further, a possible
approach that may be carried out is the measurement of the carbon dioxide
flux, or
another passive tracer flux, to provide a quantitative estimate of
the flux divergence. The measurement of the vertical divergence of the
radiative flux with existing affordable instrumentation is unreliable.
Thus, without high vertical resolution radiative flux divergence measurement
capability, the measurement of the vertical structure of temperature and
moisture, and the computation the radiative flux from numerical simulation,
represents the state-of-art procedure.
The relative importance of shallow gravity or density
currents to turbulent fluxes in the nighttime stable boundary layer is
not yet clear. Their presence is usually not observed unless high frequency/multiple
level tower measurements, radar, lidar or sodar are available for documentation.
Both laboratory and observational studies (Simpson, 1997) indicate that
there can be intense mixing associated with the leading edge or head of
the current. Further, the wind and temperature structure of these currents
is conducive to the formation of waves that may propagate relative to the
current (Ralph et al., 1993), or KH billow activity along the elevated
interface of the current may often lead to short-lived turbulence bursts.
The vertical shear that develops due to the katabatic flow jet may become
strong enough to initiate temporary turbulent activity or gravity
waves. Furthermore, this shear may be enhanced by the existence of
opposing or cross-current ambient flow which may provide, independently,
another instability source at low levels (Fitzjarrald, 1984; Arritt and
Pielke, 1986). With the routine occurrence of a southerly low-level jet
over the CASES experimental site, such a mechanism will be present and
quantifiable.
Tracer releases undertaken during CASES-99 will provide evidence of the stable boundary layer influences on dispersion characteristics under both clear and cloudy conditions. Sulfur hexafluoride and perfluorocarbon tracers will be emitted near the ground from a computer-controlled release system, and concentrations will be measured downwind of the source at a rate of 1 Hz with multiple sensors. Elevated releases, which have seldom been explored, will be completed with use of either a kite system, the 40 m tower, or a tethered balloon, to ascertain the significance of top-down dispersion via intermittent transports to the surface-based measurement array. An important consideration is the effect of convection, generated by radiative cooling at the top of the cloud, on nocturnal dispersion in both the vertical and horizontal directions.
Under clear, very stable atmospheric conditions, the goal of the dispersion releases is to improve current understanding of diffusion, meander, and concentration fluctuations for ground-based plumes. The evaluation of these releases will have enormous implications for the problem of odor transport from animal confinement facilities and pesticide vapor transport during stable NBL conditions. Concentration data from fast-response tracer analyzers located within 1 km of a source will be used to relate measured relative diffusion coefficients, concentration statistics, travel time, and plume meander to on-site wind data from fast-response anemometers and to evaluate the performance of a meandering plume model. The project will produce a detailed database of plume diffusion, and a more complete understanding of plume behavior on short time scales in the nocturnal boundary layer.
The evening transition can be characterized as follows. Under clear and mostly clear skies, the vertical contraction of the convective boundary layer begins in the late afternoon as the positive surface heat flux weakens in response to the decrease in the Sun's elevation. This transition from a convective to a stable boundary layer has been modeled by, for example, Nieuwstadt and Brost (1986) and Sorbjan (1997), but detailed observations have not, in general, been available for model verifications. Sorbjan's extension of the Nieuwstadt and Brost study considered a gradual, rather than an abrupt, heat flux evolution during the transition period. During this decaying period, turbulent eddies persist even when the heat flux at the surface becomes negative and a low-level surface inversion is becoming established. Large-scale updrafts are able to penetrate the stable layer aloft and entrain air. The observations carried out in conjunction with dispersion of tracers during the transition period will be used to provide data to establish whether or not turbulent eddies persist in the presence of negative heat flux, and to examine the characteristics of dispersion during the transition from convective to stable conditions. Measurements should enable the evaluation of whether near-surface contaminants within the stably stratified NBL can be transported into the residual layer, reducing the near-surface concentration and potential hazard.
The primary source of instrumentation is expected to be that supplied by the National Science Foundation's (NSF) instrument deployment request, administered through the National Center for Atmospheric Research's Atmospheric Technology Division (NCAR ATD). Additional in-situ boundary/surface layer instrumentation provided by CASES-99 investigators (see Tables 1 and 2) would define the meso-b- to microscale boundary layer evolution and structure. Specifically, these instruments will be used to define the evolving temperature, wind, and constituent profiles and the wave and turbulence fluxes of heat and momentum. At the lowest elevation AGL (< 5 m), the CASES-99 community of scientists is advocating and seeking the deployment of instruments with above average frequency response (approaching 50 Hz). Existing data sources in and around the Walnut River Watershed (WRW) field site, such as from the ongoing Argonne Boundary Layer Experiments (ABLE), the National Weather Service (NWS) in Wichita and the Atmospheric Radiation Measurements (ARM) CART site, would provide enhanced observation of meso-a- and meso-b-scale features.These observational resources will be utilized optimally, and combined with modeling studies of mesoscale and small-scale processes to achieve the science goals described in Section 2. The assimilation of both standard and state-of-the-art instrumentation will allow researchers to construct the most comprehensive description of the NBL structure, evolution, and instability to date.
| Instrument | Provider | Comments or Capability |
|---|---|---|
| Aircraft (4) | ||
| Wyoming King Air | NCAR Deployment | turbulence, radiation in-flight |
| NOAA LongEZ | Nappo, NOAA/ARL - Univ. Tennessee | turbulence (to 20Hz, u, v, w; 5 Hz in T) |
| NOAA Twin Otter | Wellman, NOAA | turbulence, lowest altitude 150 m |
| HELIPOD | Muschinski, Lenschow, NOAA, NSF, NCAR, Army | turbulence to 50 Hz |
| Archived data | ||
| Ceilometer | ||
| Data Hub | ||
| Energy balance | ||
| Hot-wire Anemometers | ||
| Instrument pad | ||
| Kite | ||
| Lidars | ||
| Microbarographs | ||
| Mini-sodars | ||
| Other | ||
| Radars | ||
| Rain guages | ||
| RASS | ||
| Scintillometers | ||
| Sodars | ||
| Sonic anemometers | ||
| Soil moisture probes | ||
| Soil temperature probes | ||
Supplemental in-situ measurements of the NBL wind,
temperature, and humidity profiles and
of heat, moisture, and momentum fluxes at a number of levels would
employ a variety of instru
ment types. Kite profiling technology (or a substitute balloon during
very light winds) will allow
nearly continuous measurements of high-resolution mean profiles of
temperature, humidity and
winds, in addition to turbulent statistics ( and ) throughout the NBL
(permission for kite
operations near CASES have been previously obtained for ~ 1.2-2.3 km
AGL). The ARM-CART
site in the Towanda sub-basin (NW corner of the WRW), along with other
nearby instruments, is
currently measuring soil temperature and moisture. The soil measurements
will provide a rela
tively rare opportunity to initialize soil conditions in our mesoscale
models and improve simula
tion of moisture flux heterogeneity. Likely non-tower-based instrumentation
includes aircraft-
based winds, temperatures, and turbulence (the Wyoming King Air and
possibly the NOAA Long
EZ), lidar winds, and tracer and remotely-sensed water vapor measurements.
We will have the
Army's FM-CW radar, now located at Dugway Proving Grounds to provide
very high resolution
[O(1 m)] measurements of and visualization of NBL dynamical phenomena
(see cover). The
FM-CW has already obtained clearance for use at the CASES site and
will not interfere with the
NWS Wichita WSR-88D radar. The Army will also provide the Turbulent
Eddy Profiler (TEP),
currently under final testing at the University of Massachusetts (Mead
et al., 1998; Pollard et al.,
1998). The TEP is a radar capable of measuring winds and turbulent
quantities in a 3-dimensional
cone from 150 m AGL to 2000 m AGL with 30 m horizontal and vertical
resolution.
ABLE provides three remote sensing sites where vertical
profiles of wind and temperature are
sampled by 915-MHz radar wind profilers, Doppler acoustic sounders,
and radio acoustic sound
ing systems (RASS). These measurements extend into and through the
NBL and will provide both
surface and upper-level conditions. Each site is also instrumented
with a surface flux station, and
an eddy correlation flux station is located at the ARM-CART site in
the northern portion of the
WRW.
NOAA/ETL plans to make the following measurement
systems available for CASES-99: a
449-MHz wind-profiler, a state-of-the-art high-resolution Doppler sodar,
three mobile microbaro
graphs and a high-resolution (~ 30 m range gates) Doppler lidar. In
addition NOAA/ETL plans to
rent the helicopter-borne, state-of-the-art turbulence instrument HELIPOD
(Muschinski and
Wode 1998) from the University of Hannover, Germany. The 449-MHz wind-profiler
is sensitive
to turbulent fluctuations at length scales around 33 cm (the radar's
Bragg scale) while the Bragg
scales of a 915-MHz radar and of the FM-CW radar are 16 cm and 5 cm,
respectively. Collocated
radar measurements at different wavelengths, in addition to contributing
to the main CASES-99
scientific goals, will allow testing of the existing hypotheses on
the nature of the dominating radar
scatter/reflection mechanisms. With additional signal processing wind
profilers have the potential
to provide automatic high-quality profiles of refractive turbulence
(0th spectral moment), mechan
ical turbulence (2nd spectral moment), 5 min horizontal winds, 10 s
winds (u, v, w) and tempera
ture for heat and momentum flux estimates. From those high-quality
observations, it is also
possible to derive humidity gradient profiles and near-continuous,
high vertical resolution
humidity profiles throughout the troposphere (Gossard et al., 1998,
Stankov, 1998).
CASES-99 is also very fortunate to be receiving instrumentation
committed by European col
leagues. The Risoe National Laboratory (RNL) will be bringing a laser
scintillometer for turbu
lence measurements, a thermocouple chain consisting of 30 units to
be placed at high density on a
tower (or multiple towers), three sonic anemometers, and a backscatter
lidar. Vaisala Oyj, Finland
will provide a standard ceilometer to study nocturnal boundary layer
structure jointly with the
RNL group. The Spanish Meteorological Institute has stated their intention
to bring three
microbarographs, a tethersonde and an eddy flux tower.
The NOAA Long EZ would allow nighttime measurements
down to 50 m AGL with full
wind, temperature, humidity and turbulent flux capabilities, and will
be available for two weeks or
more if resources are found to cover the cost of its use. In well moonlit
conditions, this aircraft
could fly as low as 20 m AGL. Such measurements would complement the
Wyoming King Air
measurements, whose minimum altitude is 150 m, by giving improved spatial
representation of
events between the King Air top of the NCAR-supplied 40 m tower. Since
turbulence events and
gravity waves in the NBL can frequently occur in the 50 - 150 m range
(i. e., the cover figure,
Eaton et al., 1995), at times the Long EZ would be a very valuable
tool with which to supplement
the lidar(s), balloons, sodars and radars (except for the lidars, these
instruments do not have scan
ning capability).
The in-situ sonde HELIPOD has a fast-responding (pre-averaging
sampling rate is 1000 Hz)
and low-noise (a few mK) temperature sensor, a fast and drift-free
three-component (a dewpoint
mirror, a humicap and a Lyman-a hygrometer) humidity sensor, and a
high-resolution five-hole
probe for 3D wind vector measurements. The resulting pressure, temperature,
humidity, and 3D
wind data are provided at a rate of 100 Hz, corresponding to one sample
every 40 cm if the HELI
POD is flown with the typical forward velocity of 40 m s-1. Therefore,
the HELIPOD's spatial res
olution is close to the radars' Bragg scales, allowing a direct comparison
between in-situ
turbulence spectra and radar reflectivity to be made. The HELIPOD is
designed to provide reliable
data of turbulent fluxes of momentum, heat and humidity in the stably
stratified boundary layer.
Muschinski and Wode (1998) reported on the first in-situ evidence of
coexisting sub-meter tem
perature and humidity sheets in the lower free troposphere above the
Arctic sea ice using HELI
POD.
Sensitivity to boundary layer structure, evolution,
and heterogeneity on the meso-b-scale will
be provided on ~ 15-min time scales by ANL ABLE Wind Profilers with
RASS (3), mini-sodars
(3), and 10-m towers (3) spaced at ~ 50 km and spanning the CASES Central
Site and the likely
location of CASES-99. These measurements will be supplemented at the
meso-a-scale by addi
tional measurements using the Wind Profiler Network (an additional
6 systems) and aircraft (the
Wyoming King Air and possibly the Long EZ) to define spatial variability
of the boundary layer
and the embedded smaller-scale dynamics. Together with the more comprehensive
measurements
to be performed at the CASES-99 core instrument cluster (including
additional instrumentation
described below), these systems will define the meso-b-scale variations
in boundary-layer struc
ture relevant to CASES-99 Science Goals 1, 3, and 5. Specifically,
these measurements will define
the spatial and temporal variability of the evolving surface- and boundary-layer
flows, the spatial
extent of shear flows capping the stable boundary layer, and the structure
and evolution of large-
scale drainage currents under various forcing conditions.
More focussed measurements are anticipated in a region
~3 to 5 km on a side surrounding the
CASES-99 core instrument cluster (see Figure 3). Instrumentation here
will include microbaro
graphs (~16), 10-m tower and/or flux stations (~20) yielding wind,
temperature, and relative
humidity measurements, CLASS balloon sounding systems (3) providing
~hourly high-resolution
profiling of the entire boundary layer, and aircraft measurements along
smaller baselines to
increase temporal sensitivity to small-scale dynamics around the CASES-99
core instrument clus
ter. Instrumentation that may be contributed by other participants
includes additional microbaro
graphs, balloons, tethersondes, and meteorological and surface flux
towers. This instrument
cluster will address CASES-99 Science Goals 1, 2, 3, and 5 on the micro-g-scale
by providing
sensitivity to 1) the structure and dynamical variability of the evolving
boundary layer, 2) the
presence of gravity wave perturbations of the boundary layer on scales
of a few to tens of km, 3)
the spatial and temporal variability of surface-layer structure and
its fluxes of heat, moisture, and
momentum, and 4) the horizontal extent and spatial structure of the
processes accounting for
instability, mixing, and transports.
Figure 3 (above): A preliminary sketch of meso-g- and microscale instrumentation array for CASES-99. Mb = Microbarographs (3 mobile), T = standard tower or flux station, Te = tethersonde, Li = lidar, K = Kite, T40 = 40 meter tower, S = sodar, G = G:ASS sounding system, R = rawinsonde, F = FM-CW radar, TEP = Turbulent Eddy Profiler, 4 = 449 MHz profiler, La = Laser wind profiler, Sc = Laser scintillometer This potential site is located southeast of Leon, Kansas (see Figure 2) in the meso-b-scale CASES site/Walnut River Watershed.
Finally, CASES-99 core instrumentation will include a special 40 m tower, the Dugway FM- CW radar, sodars (2), the turbulent eddy profiler (TEP), tethered balloons (2 or 3), a CLASS bal loon sounding system, a 449 MHz profiler, and aerosol Doppler lidars (1 or 2). Other instrumenta tion that may also be available for use at the core instrument cluster includes a kite-borne sounding system (currently proposed to NSF), additional Raman and Doppler lidars, and the NOAA/Germany HELIPOD for in-situ turbulence measurements. Together with the Wyoming King Air and hopefully NOAA Long EZ instrumentation, this suite of instruments offers the opportunity for correlative measurements and nocturnal boundary-layer characterization of unprecedented scope and detail. To optimize these measurement capabilities, the core cluster will be configured so as to benefit as fully as possible from instrument synergism. For example, measurements of turbulence structures and intensities using the FM-CW radar will be used in real time to identify regions having an elevated level of turbulence (or dynamics that appears likely to initiate turbulence bursts) in order to guide in-situ measurements using the kite instrument pay load, the TEP, the HELIPOD, and the King Air and/or Long EZ. Likewise, real-time radar, sodar, lidar, and balloon wind measurements will enable us to align the lidars for optimal sensitivity to momentum fluxes due to the larger-scale (2D) wave and instability processes occurring within, and dependent on, the evolving boundary-layer structure. These more focussed measurements will then enable a characterization of the vertical and horizontal structure, the dynamics, the horizontal extent, and the temporal variability of these flows and the turbulence bursts, transports, and mixing that accompanies them. Simultaneously, the spatially distributed array of towers and microbarographs will capture the surface signature of propagating or spatially limited events.
The Doppler lidars will likely be employed for two
purposes. Periodically, one will be used to
map the spatial variations of the large-scale wind field in the plane
of the upper-level, boundary-
layer shear flow. A second, and possibly more important, task will
be to define the vertical and
temporal variations of the quasi-two-dimensional (2D) momentum fluxes
accompanying the grav
ity wave and shear flow instability processes accounting for momentum
transports at larger scales
of motion. This requires a co-planar, dual-beam configuration for each
component measurement,
as has often been employed with various atmospheric radars (Vincent
and Reid, 1983; Fritts and
Vincent, 1987; Fritts and Yuan, 1989). Such measurements are key to
identifying and understand
ing the wave and instability dynamics accounting for turbulence bursts
and their associated mix
ing, fluxes, and transports. The Doppler lidars can be configured in
any of several ways for these
purposes, but at least two systems are necessary to define vector momentum
fluxes, and occa
sional horizontal (PPI) scans of the 2D motion field, in order to insure
the smallest statistical
uncertainties in radial wind and momentum flux estimates. Either individual
systems can make
dual-beam measurements, or two systems could be configured to make
measurements in overlap
ping volumes so as to minimize the degree of spatial or temporal averaging
needed to provide
valid momentum flux estimates.
Other instrumentation (the sodars, tethered balloons,
CLASS system, 449 MHz profiler, and
kite profiling instrumentation) will further define the environment
in which these dynamics are
occurring, the spatial homogeneity or heterogeneity of the boundary
layer, and the dynamics
accounting for turbulence bursts and their effects. In-situ measurements,
in particular, will be
highly beneficial in defining the intermittency and strength of the
turbulence bursts, the character
istics of the resulting temperature, velocity, and constituent structures
following mixing, and in
providing the most quantitative comparisons with simulations of these
dynamics.
For more than two decades, both FM-CW radars and
pulsed Doppler radars operating at UHF
frequencies have proven to be efficient tools to monitor structure
and dynamics of the clear-air
turbulence within the planetary boundary layer (Gossard, 1990; Wilczak
et al., 1996). Due to their
ability to probe vertical profiles of radar reflectivity, Doppler shift
and spectral width quasi-con
tinuously with time, the clear-air radars' potential to monitor profiles
of the wind, of the refrac
tive-index turbulence structure parameter Cn2 (see cover figure), of
the turbulent energy
dissipation rate and of vertical gradients of mean quantities, is tremendous.
However, although
there have been many studies on how to use clear-air radars to monitor
variables other than the
wind (e.g., Gossard et al., 1998; Muschinski, 1997; Stankov, 1998),
generally, there is still a lack
of convincing data from well-designed multi-sensor field experiments
which make full use of the
synergism between state-of-the-art remote sensing, state-of-the-art
in-situ measurements and
state-of-the-art modeling. CASES-99 will help significantly to fill
this gap. In addition, CASES-
99 will test and establish new radar-data acquisition techniques, new
data-processing strategies,
and new theories describing the complicated mechanisms of electromagnetic
reflection and scatter
from turbulent refractive-index fluctuations in the clear air.
For example, FM-CW radars provide very high resolution
measurements (down to about 1 m
in range and several seconds in time) of Cn2. Hence, an FM-CW radar
can play the role of a
"microscope" which monitors the fine-structure and small-scale dynamics
of turbulence, waves
and instabilities within a vertical column of the nocturnal boundary
layer above the radar site.
Collocated in-situ measurements (towers, kites, balloons, aircraft,
HELIPOD) and lower-resolu
tion remote sensing measurements made with pulsed Doppler radars and
lidars, together with
computer simulations of the nocturnal boundary layer will benefit from
that "microscopic" moni
toring of the atmosphere. Since FM-CW radars, in contrast to pulsed
Doppler radars (typical
range resolution 100 m), have a very large bandwidth and because of
frequency-allocation and
radio interference problems, they cannot be deployed operationally
but can only serve as research
tools. However, during the past several years, substantial progress
has been made in the develop
ment of interferometric techniques using pulsed Doppler radars. While
Spatial Interferometry (SI)
enables one to get more detailed information about the transverse distribution
of scattering cen
ters within the radar's resolution volume (the new "Turbulent Eddy
Profiler" is a multi-receiver
spatially interferometric pulsed Doppler radar operating at 915 MHz),
Frequency-Domain Inter
ferometry (FDI) allows one to observe details about the radial distribution
of scattering centers
within the resolution volume. In contrast to FM-CW radars, those interferometric
pulsed radars do
have the potential to become important parts of future wind-profiler
networks. Chilson et al.
(1997) have used FDI at VHF frequencies for the first time to monitor
small-amplitude KH bil
lows in the upper troposphere, and the implementation of FDI at existing
boundary-layer Doppler
radars will contribute to a better monitoring of waves and instabilities
in the nocturnal boundary
layer.
Recently, the existing theory of the zero and first
moments of the variance- and cross-spectra
of signals observed with pulsed Doppler radars operating in the conventional
mode, in the Fre
quency-Domain Interferometry (FDI) mode and in the Spatial Interferometry
(SI) mode has been
extended by Muschinski (1998). The new theory provides a detailed understanding
of the advan
tages of the interferometric techniques over the standard techniques.
The same theory can help
identify and interpret biases resulting from a naive interpretation
of clear-air radar data, that does
not take into account the intermittent and anisotropic nature of atmospheric
turbulence.
Used together, CASES-99 measurements and the three
classes of numerical simulations antic
ipated above would help quantify the dynamics accounting for the structure,
fluxes, and evolution
of the stable NBL. Additionally, CASES-99 measurements of NBL structure
and fluxes would
enable testing of the surface flux and turbulent mixing parameterizations
presently incorporated in
the mesoscale and LES models. Given the current belief that these parameterizations
are often
most deficient under conditions of stable stratification (Poulos, 1996;
Mahrt, 1998), the data/
model comparisons enabled by our proposed CASES-99 program would also
provide a means of
improving and further quantifying these parameterizations. Since most
atmospheric models use
similar parameterizations for stable flows, the benefits of improving
surface layer and turbulence
parameterizations would be very large indeed.
For case study reconstruction and testing of the
stable surface layer and subgrid-scale parame
terizations, emphasis will be placed on comparison with 3-d flux, intermittency,
and turbulence
measurements from the CASES-99 field experiment. Also, 1-d column model
tests will be com
pleted for rapid evaluation of parameterizations from the various mesoscale
models used. This
will provide the foundation for analysis of simulation quality and
the evaluation of the parameter
izations. At sufficient resolution, mesoscale models have been shown
capable of reproducing
Kelvin-Helmholtz billows, small-scale gravity waves, shallow gravity
currents, very stable layers
(these, typically produced by strong, near-surface, radiative flux
divergence, have been simulated
as strong as 100oC km-1 by Poulos, 1996) and vertical jet structure,
although subject to parameter
ized turbulence (or SGS) and surface layer effects. These phenomena
may be less influenced by
the turbulence parameterization in the LES modeling (see discussion
below) and not at all
affected in DNSs. Thus, codes such as these will only be relied upon
to evaluate the finest details
of turbulent processes. Inaccuracies in the simulation and parameterization
of the very stable sur
face layer will be evaluated to pose new solutions to that problem,
as hypothesized by various
CASES-99 participants, and will hopefully lead to the improvement or
development of entirely
new parameterization schemes.
Because the optimal simulation scale of the LES lies
between that of the mesoscale model and
that of the DNS, simulated stable NBLs using LES are an important part
of determining the flaws
in SGS parameterization. Although LESs are unable to simulate the details
of turbulence shown in
the DNS technique, we expect these simulations to capture the smallest
scales of turbulence possi
ble at high Re (107-109), thereby implicating certain aspects of the
mesoscale model as deficient
in capturing realistic NBL evolution, particularly as regards intermittent
turbulent behaviour. Of
course, such simulations will be most valuable if the smallest resolvable
scale reaches the upper
end of the inertial subrange, which may be quite difficult since the
inertial subrange is typically
shifted to considerably smaller scales in the stable NBL. Only then
will the SGS parameterization
provide a realistic energy cascade for the LES. Similarly, comparisons
with low Re (O[104])DNS
will show where, perhaps, the unrealistic Re (via viscosity much larger
than that in the natural
atmosphere) in the DNS fails to allow the development of realistic
atmospheric features. Such
comparisons will ensure the proper interpretation of the dynamical
evolution in all three modeling
techniques. In general, idealized profiles based on CASES-99 measurements
would be used to ini
tialize the LES, due to the cyclical boundary conditions inherent in
the LES technique.
Focussed simulations using at least one highly-optimized
spectral code will be performed to
examine flow instability within and above the NBL, the occurrence,
intensity, and intermittency of
turbulence, and its penetration into the stable NBL, and the structure
of, and the turbulence fluxes
accompanying, such penetration events. Applications of one such code
in current shear flow insta
bility studies use resolution of 5123 or greater, yield a buoyancy/inertial
sub-range of turbulence
spanning a decade or more, and enable assessment of turbulent fluxes
and transport accompany
ing such events (see Figure 4). Indeed, such simulations now access
turbulence Reynolds numbers
(Re) of 4 x 104 and as such can describe the implications for mixing
and transport in real flows.As
such, these simulations will become an integral part of improving the
SGS turbulence and surface
layer models in both LES and mesoscale models. Indeed, such simulations
may provide our only
hope for high spectral frequency comparisons to CASES-99 data.
Incompressible and/or anelastic studies will be performed
to supplement the mesoscale model
and LES code simulations and will focus on the dynamical processes
leading to instability, turbu
lence, and the associated fluxes and transports. Because these studies
will 1) focus on dynamical
instability processes rather than the NBL structure and evolution and
2) achieve greater numerical
efficiencies and resolution than are possible with the high Re models
(which must rely on an SGS
parameterization that is known to be flawed in stable conditions),
they will be able to address the
statistical structure, intensities, and implications for turbulence
fluxes and transports accompany
ing the major sources of turbulence within the evolving NBL.
It is also essential to connect the DNS with standard
micrometeorological theory. The basic
micrometeorological equations are the budgets of momentum, energy,
and scalars (moisture, con
stituents). In standard micrometeorological practice, these budgets
are averaged in the horizontal
and terms are deleted on the assumed basis of horizontal homogeneity
the motivation being to
eliminate terms that are very difficult to measure and are usually
not quantified by measurements.
DNS provides the opportunity to quantify those difficult terms. More
fundamentally, the basic
hydrodynamic equations of the DNS can be used to derive the micrometeorological
budgets that
are averaged over a horizontal region, but without deletion of any
terms: in particular, the assump
tion of horizontal homogeneity will not be necessary. These fundamental
budgets must balance to
within numerical round-off errors, which can be verified. The DNS has
the unique ability to quan
tify all terms in the fundamental budgets, even those that cannot be
measured. This is especially
important with the intermittent turbulence in stratified flows, where
locally strong turbulence and
sharp thermal or constituent gradients may lead to strong, spatially
and temporally localized, mix
ing and transport (for example, at the edges to he mixing layer in
the central panels of Figure 3),
which may differ greatly from mean turbulence statistics. By performing
such simulations, and
assessments of turbulence mixing and transport, for flows that are
observed during CASES-99, it
will also be possible to make them directly comparable with measurements
of such mixing events.
As such, they should contribute significantly to the evaluation of
existing stable layer parameter
izations and the formulation of improved versions.
Figure 4 (above): An incompressible direct numerical simulation at Re = 2000, Pr = 1 and Ri = 0.05 of the turbulent breakdown of a Kelvin-Helmholtz billow in stratified shear flow. The left panels show vorticity and the right panels show potential temperature. The interaction of secondary instabilities in the form of streamwise-aligned vortex tubes interact to initiate the turbulent cascade to smaller and smaller scales of motion. In this case, the spectrum spans ~ 1.5 decades and the simulation contained 384 x 128 x768 grid points. See also Werne and Fritts (1998).
where tij is the subgrid-scale flux of momentum (for uj components).
The quantities in the above
equation are interpreted as averages over a small volume, which is
usually just the grid volume.
To parameterize the SGS eddy viscosity and diffusivity, KM, in (4.1),
all of the above volume-
averaged statistics are required, i.e., 3D low-pass filtered fields.
At a minimum, 2D flow fields are
needed to obtain estimates of the 2D low-pass filtered statistics.
KM can be expressed as
where e is the SGS turbulent kinetic energy (TKE), and l is a characteristic
length scale.This
length scale is often assumed to be (or proportional to) the effective
grid spacing
, where Ds is the grid length scale. Muschinski (1996) has discussed
in
some detail the relationship between the grid scale and the "characteristic"
length scale. But when
the local stratification is stable, one expects the SGS mixing scale
to be suppressed somewhat.
Thus, Deardorff (1980) proposed a stability correction to l that when
if .
This `stability correction' for l in (4.3) gives smaller SGS KM in
(4.2), and thus smaller mix
ing, within (locally) stably stratified regions in all three directions,
not just in the vertical. Without
a force in the horizontal component analogous to the increased restoring
force with increasing sta
bility in the vertical component, such a formulation would appear to
be oversimplified. (4.3) was
given by Deardorff without any observational or theoretical basis,
though l is intended to repre
sent the free path a particle can travel before being stopped by buoyancy
effects when displaced
from its level with an energy equal to its TKE. The coefficient 0.76
is purely empirical.
The measurements from a well-strategized CASES-99 program can address
the deficiencies
and will allow us to be more certain about the treatment of the SGS
length scale. The data
required include the volume-averaged quantities of wind (ui), potential
temperature (q), momen
tum fluxes (tij), heat fluxes (tqi), and SGS TKE (e) (see Wyngaard
and Peltier, 1996). These 3D
low-pass filtered quantities can be reasonably approximated by 2D low-pass
filtering. Adequate
data at the Re of the natural atmosphere can only from comprehensive
field measurements, since
it is difficult to generate small-scale inertial subrange eddies in
laboratory experiments. CASES-
99 promises to provide such information. These data can also now be
generated by state-of-the-art
DNS (see Section 4.3) at lower Re, and will be used where relevant,
in conjunction with measure
ments.
One potential source of observational data relevant
to the SGS turbulence parameterization
problem, are wind profiling radars, which measure the radial velocity
spectrum of target move
ment in the pulse volume. From the zeroth, first and second moments
(i.e., the raw measurements
of the radar moments) of the spectrum, Gossard et al. (1998) and Stankov
(1998) have shown that
measurements of the turbulent eddy diffusivity, Kh, are possible. It
should be kept in mind, how
ever, that those techniques, although very promising, rely on the assumptions
of homogeneous,
volume-filling and isotropic turbulence, and it is not known how sensitively
the results are biased
if those assumptions are not fulfilled (e.g. Figure 4), as it is to
be expected in the stable layers. The
multi-sensor observations to be made during CASES-99 will help to estimate
the degree of reli
ability of wind-profiler measurements of quantities other than the
wind, and therefore the poten
tial for the use of radar moments in evaluating fundamental outputs
from SGS parameterizations
(eddy diffusivity).
Develop and distribute a PI questionnaire on CASES-99 data issues to
compile data needs.
Develop and formalize data documentation and format procedures.
Implement CASES-99 on-line catalog for the field phase.
Data available via UCAR/JOSS CODIAC data management system.
The CASES-99 field phase will build on the data management strategy
implemented for the
CASES-97 study. The following data protocols are specified in the CASES-97
Science Plan, and
form the basis of the data management strategies that will be used
in CASES-99.
- Ensure open access to all CASES-99 datasets. This requires
a data management strategy that
facilities data exchange and investigators accepting responsibility
for making data available.
- CASES-99 will take advantage of several existing data centers
to house a variety of datasets
to be collected. These include the UCAR/Joint Office for Science Support
(JOSS), NCAR,
NOAA, NASA/LaRC/DAAC and ARM/ORNL. Cooperative agreements for unrestricted
exchange and access of CASES-99 data from all these locations will
be established.
- All investigators participating in CASES-99 must agree to promptly
(within one year of the
end of the project) submit data to the appropriate data archive center
to facilitate the data process
ing, archival activities and distribution. Datasets must be submitted
to the archive in a usable for
mat and with sufficient documentation to allow easy access and understanding
by others. CASES-
99 will utilize a distributed archive strategy whereby datasets will
be housed at the most conve
nient location to permit easy access. A single access point with appropriate
links to the distributed
data centers will be provided through UCAR/JOSS CODIAC.
- Ensure that the CASES-99 dataset is comprehensive by acquiring
appropriate data from
NOAA, NASA, DOE-ARM, local agencies, etc. that might be conducting
programs in conjunc
tion with, or of interest to, CASES-99 investigators.
The CASES-99 Science team will coordinate the development and implementation
of a data
management plan. The final objective is a high quality data archive
that has easy and timely
access by a large community of investigators. This is a large task
given the diversity of participa
tion and instrumentation planned for CASES-99. It is proposed that
the UCAR/JOSS collaborate
with the Science Team in this effort.
To that end, the following tasks are viewed as essential to providing
an acceptable level of data
management support for CASES-99.
Provision of an on-line catalog at the CASES-99 field site during the
field season.
Access to preliminary datasets and selected operational data during
the analysis phase.
Suggest standardized format(s) and guidelines for dataset documentation,
status and sum
mary reporting and other important data management procedures as necessary
to assure
complete documentation of project activities.
Assist CASES-99 management team with planning and coordination of data
management
activities among other agencies, projects and groups to meet investigator
needs.
Details of the CASES-99 experiment design and dataset requirements
continue to be refined
by the CASES-99 Science Team. Therefore, the proposed data management
activities represent a
cohesive plan based on the information available at this time.
In recognition that previous attempts at nocturnal/stable
layer observation have been ham
pered by insufficient correlative measurements, the scientists participating
in CASES-99 planning
have chosen to focus a large variety and number of instruments in a
relatively small (~ 5 x 5 km,
meso-g- [2 - 20 km] scale) region, rather than undersampling a larger
area. The instrumentation
currently probable for CASES-99 (Table 1, Figures 2 and 3) is truly
exceptional in its scope;
should some portion of the other possible instrumentation come to bear
on the problem, this
would only improve. The emphasis will be on identifying and quantifying
the evolution of various
nocturnal boundary layer phenomena, with regards to fluxes in particular,
both with height and in
3 dimensions. Atmospheric evolution on the meso-b (20 - 200 km) and
meso-a (200 - 2000 km)
scales will be addressed by outer arrays of wind profilers and sampling
stations provided by
ARM-CART, ANL ABLE and the Wind Profiler Network.
The modeling effort will consist of three parts.
First, mesoscale modeling and intercompari
son with multiple case nights will be pursued to define the NBL evolution
and structure. Through
comparison with CASES-99 observations, quantifiable deficiencies in
or faults in physical basis
of the surface layer and SGS parameterizations will be generated. LES
modeling will assess the
NBL evolution on finer scales, will further quantify departures from
observed flows due to SGS
parameterization deficiencies and will be employed to test improved
SGS parameterizations of
stable layer dynamics. Finally, DNS modeling at moderate Re will be
performed using both
mesoscale model and observation al data as inputs to define the dynamics
and fluxes of individual
instability phenomena. Comparison of these DNS results with CASES-99
measurements are
expected to both validate the DNS descriptions of NBL dynamics and
enable the formulation of
new descriptions of such events based on correlations with the larger-scale
flows.
While any field experiment is subject to the limitations
of data collection methods, and the sta
ble NBL and very stable NBL pushes those capabilities to their limits
(Wyngaard and Peltier,
1996), the combination of instruments brought together for CASES-99
to address joint NBL-
related scientific goals is unprecedented. The CASES-99 community of
scientists is aware of the
difficulties and previous efforts a variety of researchers have had
investigating these problems, but
believe that we now have the computational and measurement resources
to make dramatic further
improvements in understanding and quantifying stable layer dynamics
and transport.
| Name | Eddress | Affiliation |
|---|---|---|
| Name | Research Interest | Source |
|---|---|---|