• Bertin, F., , Barat J. , , and Wilson R. , 1997: Energy dissipation rates, eddy diffusivity and the Prandtl number: An in situ experimental approach and its consequences on radar estimate of turbulent parameters. Radio Sci., 32 , 791804.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cadet, D., 1977: Energy dissipation within intermittent clear air turbulence patches. J. Atmos. Sci., 34 , 137142.

  • Caldwell, D. R., 1983: Oceanic turbulence: Big bangs or continuous creation? J. Geophys. Res., 88 , 74537550.

  • Carter, G. S., , and Gregg M. C. , 2002: Intense and variable mixing near the head of the Monterey submarine canyon. J. Phys. Oceanogr., 32 , 31453165.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cohn, S. A., 1995: Radar measurements of turbulent eddy dissipation rate in the troposphere: A comparison of techniques. J. Atmos. Oceanic Technol., 12 , 8595.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dillon, T. M., 1982: Vertical overturns: A comparison of Thorpe and Ozmidov length scales. J. Geophys. Res., 87 , 96019613.

  • Ellsaesser, H. W., 1969: A climatology of epsilon (atmospheric dissipation). Mon. Wea. Rev., 97 , 415423.

  • Fer, L., , Skogseth R. , , and Haugan P. M. , 2004: Mixing of the Storfjorden overflow (Svalbard Archipelago) inferred from density overturns. J. Geophys. Res., 109 .C01005, doi:10.1029/2003JC001968.

    • Search Google Scholar
    • Export Citation
  • Ferron, B., , Mercier H. , , Speer K. , , Gargett A. , , and Polzin K. , 1998: Mixing in the Romanche fracture zone. J. Phys. Oceanogr., 28 , 19291945.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fukao, S., and Coauthors, 1994: Seasonal variability of vertical eddy diffusivity in the middle atmosphere, 1. Three-year observations by the middle and upper atmosphere radar. J. Geophys. Res., 99 , 1897318987.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gage, K. S., 1990: Radar observations of the free atmosphere: Structure and dynamics. Radar in Meteorology: Battan Memorial and 40th Anniversary Radar Meteorology Conference, D. Atlas, Ed., Amer. Meteor. Soc., 534–565.

    • Search Google Scholar
    • Export Citation
  • Gage, K. S., , and Gossard E. E. , 2003: Recent developments in observation, modeling, and understanding atmospheric turbulence and waves. Radar and Atmospheric Science: A Collection of Essays in Honor of David Atlas, Meteor. Monogr., No. 30, Amer. Meteor. Soc., 139–150.

    • Search Google Scholar
    • Export Citation
  • Gage, K. S., , Green J. L. , , and VanZandt T. E. , 1980: Use of Doppler radar for the measurement of atmospheric turbulence parameters from the intensity of clear-air echoes. Radio Sci., 15 , 407416.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Galbraith, P. S., , and Kelly D. E. , 1996: Identifying overturns in CTD profiles. J. Atmos. Oceanic Technol., 13 , 688702.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garett, C., , and Munk W. , 1975: Space-time scales of internal waves. Progress report. J. Geophys. Res., 80 , 291297.

  • Gargett, A. E., 1999: Velcro measurements of turbulent kinetic energy dissipation rate. J. Atmos. Oceanic Technol., 16 , 19731993.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gavrilov, N. M., , Luce H. , , Crochet M. , , Dalaudier F. , , and Fukao S. , 2005: Turbulence parameter estimations from high-resolution balloon temperature measurements of the MUTSI-2000 campaign. Ann. Geophys., 23 , 24012413.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ghosh, A. K., , Jain A. R. , , and Sivakumar V. , 2003: Simultaneous MST radar and radiosonde measurements at Gadanki (13.5°N, 79.2°E) 2. Determination of various atmospheric turbulence parameters. Radio Sci., 38 .1014, doi:10.1029/2000RS002528.

    • Search Google Scholar
    • Export Citation
  • Hocking, W. K., 1985: Measurement of turbulent eddy dissipation rates in the middle atmosphere by radar techniques: A review. Radio Sci., 20 , 14031422.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S-Y., , and Pan H-L. , 1996: Nonlocal boundary layer vertical diffusion in a medium-range forecast model. Mon. Wea. Rev., 124 , 23222339.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hooper, D. A., , and Thomas L. , 1998: Complementary criteria for identifying regions of intense atmospheric turbulence using lower VHF radar. J. Atmos. Sol. Terr. Phys., 60 , 4961.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hooper, D. A., , McDonald A. J. , , Pavelin E. , , Carey-Smith T. K. , , and Pascoe C. L. , 2005: The signature of mid-latitude convection observed by VHF wind-profiling radar. Geophys. Res. Lett., 32 .L04808, doi:10.1029/2004GL020401.

    • Search Google Scholar
    • Export Citation
  • Johansson, C., , Hennemuth B. , , Bosenberg J. , , Linne H. , , and Smedman A. , 2005: Double-layer structure in the boundary layer over the Baltic Sea. Bound.-Layer Meteor., 114 , 389412.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kantha, L. H., 2003: On an improved model for the turbulent PBL. J. Atmos. Sci., 60 , 22392246.

  • Kantha, L. H., , and Clayson C. A. , 1994: An improved mixed layer model for geophysical applications. J. Geophys. Res., 99 , 2523525266.

  • Kantha, L. H., , and Clayson C. A. , 2000: Small Scale Processes in Geophysical Fluid Flows. Academic Press, 888 pp.

  • Kantha, L., , and Clayson C. A. , 2004: On the effect of surface gravity waves on mixing in an oceanic mixed layer. Ocean Modell., 6 , 101124.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kellogg, W. W., 1956: Diffusion of smoke in the stratosphere. J. Meteor., 13 , 241250.

  • Kim, J., 1991: Turbulent and gravity wave transport in the free atmosphere. Ph.D. dissertation, Oregon State University, 103 pp.

  • Kung, E. C., 1966: Large-scale balance of kinetic energy in the atmosphere. Mon. Wea. Rev., 94 , 627640.

  • Kurosaki, S., , Yamanaka M. D. , , Hashiguchi H. , , Sato T. , , and Fukao S. , 1996: Vertical eddy diffusivity in the lower and middle atmosphere: A climatology based on the MU radar observations during 1986–1992. J. Atmos. Terr. Phys., 58 , 727734.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lilly, D. K., , Waco D. E. , , and Adelfang S. I. , 1974: Stratosphere mixing estimated from high altitude turbulence measurements. J. Appl. Meteor., 13 , 488493.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Losleben, M., , and Pepin N. , 2000: Thermal wind frequency on the east slope of the Colorado Front Range. Eos, Trans. Amer. Geophys. Union, 82 .(Fall Meeting Suppl.), Abstract U22A-03.

    • Search Google Scholar
    • Export Citation
  • Luce, H., , Dalaudier F. , , Crochet M. , , and Sidi S. , 1997: An improved interpretation of VHF oblique radar echoes by a direct balloon Cn 2 estimation using a horizontal pair of sensors. Radio Sci., 32 , 12611268.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Luce, H., , Fuka S. , , Dalandier F. , , and Crochet M. , 2002: Strong mixing events observed near the tropopause with the MU radar and high-resolution balloon techniques. J. Atmos. Sci., 59 , 28852896.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McDonald, A. J., , Carey-Smith T. K. , , Hooper D. A. , , Fraser G. J. , , and Lublow B. P. , 2004: The effect of precipitation on wind-profiler clear air returns. Ann. Geophys., 22 , 39593970.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mellor, G. L., , and Yamada T. , 1982: Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. Space Phys., 20 , 851875.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moeng, C-H., , and Sullivan P. P. , 1994: Comparison of shear- and buoyancy-driven planetary boundary layer flows. J. Atmos. Sci., 51 , 9991022.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moum, J. N., 1996: Energy-containing scales of turbulence in the ocean thermocline. J. Geophys. Res., 101 , 1409514101.

  • Nastrom, G. D., , and Eaton F. D. , 1997: Turbulence eddy dissipation rates from radar observations at 5–20 km at White Sands Missile Range, New Mexico. J. Geophys. Res., 102 , 1949519505.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nastrom, G. D., , and Eaton F. D. , 2005: Seasonal variability of turbulence parameters at 2 to 21 km from MST radar measurements at Vandenberg Air Force Base, California. J. Geophys. Res., 110 .D19110, doi:10.1029/2005JD005782.

    • Search Google Scholar
    • Export Citation
  • Osborn, T. R., 1980: Estimates of the local rate of vertical diffusion from dissipation measurements. J. Phys. Oceanogr., 10 , 8389.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peters, H., , Gregg M. C. , , and Toole J. M. , 1988: On the parameterization of equatorial turbulence. J. Geophys. Res., 93 , 11991218.

  • Phillips, O. M., 1977: The Dynamics of the Upper Ocean. Cambridge University Press, 336 pp.

  • Piera, J., , Roget E. , , and Catalan J. , 2002: Turbulent patch identification in microstructure profiles: A method based on wavelet denoising and Thorpe displacement analysis. J. Atmos. Oceanic Technol., 19 , 13901402.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rao, N. D., , Kishore P. , , Rao T. N. , , Rao S. V. B. , , Reddy K. K. , , Yarraiah M. , , and Hareesh M. , 1997: Studies on refractivity constant, eddy dissipation rate and momentum flux at a tropical latitude. Radio Sci., 32 , 13751389.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rao, N. D., , Rao T. N. , , Venkataratnam M. , , Tulasiraman S. , , and Rao S. V. B. , 2001: Diurnal and seasonal variability of turbulence parameters observed with Indian MST radar. Radio Sci., 36 , 14391457.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Smyth, W. D., , and Moum J. N. , 2000: Length scales of turbulence in stably stratified mixing layers. Phys. Fluids, 12 , 13271342.

  • Thorpe, S. A., 1977: Turbulence and mixing in a Scottish Loch. Philos. Trans. Roy. Soc. London, 286A , 125181.

  • Thorpe, S. A., 2005: The Turbulent Ocean. Cambridge University Press, 439 pp.

  • Trout, D., , and Panofsky H. A. , 1969: Energy dissipation near the tropopause. Tellus, 21 , 355358.

  • Wesson, J. C., , and Gregg M. C. , 1994: Mixing in the Camarinal Sill in the Strait of Gibraltar. J. Geophys. Res., 99 , 98479878.

  • Whiteway, J. A., , Pavelin E. G. , , Busen R. , , Hacker J. , , and Vosper S. , 2003: Airborne measurements of gravity wave breaking at the tropopause. Geophys. Res. Lett., 30 .2070, doi:10.1029/2003GL018207.

    • Search Google Scholar
    • Export Citation
  • Wilkins, E. M., 1958: Effective coefficients of diffusivity for atom bomb clouds at one thousand to two thousand miles. Eos, Trans. Amer. Geophys. Union, 39 , 5859.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilson, R., 2004: Turbulent diffusivity in the free atmosphere inferred from MST radar measurements: A review. Ann. Geophys., 22 , 38693887.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilson, R., , Dalaudier F. , , and Bertin F. , 2005: Estimate of the turbulent fraction in the free atmosphere from MST radar measurements. J. Atmos. Oceanic Technol., 22 , 13261339.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Woods, J. D., 1968: Wave-induced shear instability in the summer thermocline. J. Fluid Mech., 32 , 791800.

  • Worthington, R. M., 1998: Tropopausal turbulence caused by the breaking of mountain waves. J. Atmos. Sol. Terr. Phys., 60 , 15431547.

  • Zink, F., , Vincent R. A. , , Murphy E. , , and Cote O. , 2004: Comparison of radar and in situ measurements of atmospheric turbulence. J. Geophys. Res., 109 .D11108, doi:10.1029/2003JD003991.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Variation of the ɛ (x axis) with depth below the surface (y axis). The unit of ɛ, W kg−1 is the same as m2 s−3. Careful microstructure profiler-measured dissipation rates (thin lines) at station B90 in the southern Adriatic Sea at three different periods during the Dynamics of the Adriatic in Real Time campaign in August 2006 (CAR) compared with theoretical values. The thick green line denotes the overall (ensemble) mean of the multiple observations made at the same site. The thick black line denotes the mean theoretical values. Below the active mixed layer near the surface (characterized by nearly uniform dissipation rate profile), LT has been used along with Eq. (3) to obtain the theoretical values. Within the active mixed layer, indicated by the straight portion of the black curve near the surface, the inferred surface momentum and buoyancy fluxes from shipboard meteorological data have been used to deduce theoretical values (see CAR, see also Kantha and Clayson 2000). The dotted line denotes extrapolation of such theoretical values to regions below the mixed layer (figure adapted from CAR).

  • View in gallery

    Dissipation rate inferred by various studies are adapted from Gage et al. (1980). Circles refer to mean values. Solid curves refer to minimum detectable ɛ for the Sunset radar. Clear air turbulence categories are from Trout and Panofsky (1969).

  • View in gallery

    Atmospheric column properties inferred from the 2-s FASTEX sounding on 26 Feb 1997: (a) potential temperature, (b) wind velocity components U (blue) and V (red), (c) N, (d) mean shear, (e) Rig, (f) LT (m) inferred from the potential temperature profile, (g) ɛ inferred from the LT and buoyancy frequency N, and (h) K derived from ɛ (blue), and inferred from the Hong and Pan (1996) formulation (red). Note the high CAT region around 8- and 10-km altitudes that corresponds to high LT and Ri < 1/4, which is shown by a thick red line. Note that (c) shows the N deduced from the sorted profile; therefore, it shows no regions of negative stability. However Ri is computed from the unsorted profile and hence shows regions with Ri < 0.

  • View in gallery

    As in Fig. 3, but on 24 Jan 1997. Note the high CAT region just below the tropopause, which corresponds to large LT, and Ri < 1/4, which is indicated by a thick red line.

  • View in gallery

    Cumulative atmospheric column properties inferred from the 209 two-second FASTEX soundings taken between 6 Jan and 28 Feb 1997, reaching higher than 20-km altitude: (a) potential temperature, (b) N, (c) wind velocity components (m s−1) U (blue) and V (red), (d) Rig, (e) LT (m) inferred from the potential temperature, (f) ɛ (m3 s−2) inferred from the LT and N, (g) K (m2 s−1) derived from ɛ, and (h) K (m2 s−1) inferred from the Hong and Pan (1996) formulation. Thick green lines denote the overall means.

  • View in gallery

    Time series of atmospheric column properties inferred from 285 two-second FASTEX soundings taken between 6 Jan and 28 Feb 1997: (a) Rig, (b) LT (m) inferred from the potential temperature, (c) log10 ɛ (m3 s−2) inferred from the LT and N, and (d) log10 K (m2 s−1) derived from ɛ. Note intermittent but high turbulence levels in the vicinity of the tropopause when the Ri value falls below 1/4 and less often in the troposphere. Turbulence reaches high levels in the stratosphere also (but significantly less than in the tropopause).

  • View in gallery

    As in Fig. 3, but for the 6-s sounding at Tallahassee at 1200 UTC 7 Aug 2005.

  • View in gallery

    As in Fig. 5, but from 609 six-second soundings taken in 2005 at Tallahassee.

  • View in gallery

    As in Fig. 6, but from 690 six-second soundings taken twice a day at Tallahassee in 2005.

  • View in gallery

    As in Fig. 5, but from 717 six-second soundings taken in 2005 at Denver.

  • View in gallery

    As in Fig. 6, but from 717 six-second soundings taken twice a day at Denver in 2005.

  • View in gallery

    The ɛ using the (a), (c), (e) power and (b), (d), (f) spectral broadening methods for (a), (b) 2010–2057 UTC 24 Feb, (c), (d) 0222–0301 UTC 25 Feb, and (e), (f) 1719–1813 UTC 25 Feb (from Cohn 1995).

  • View in gallery

    Vertical profiles of the (a) Rig, (b) displacements, and (c) retrieved ɛ. Note the high correlation between the high displacements and the low Ri regions. The ɛ levels below the tropopause are consistent with radar measurements. No radar data are available in the ABL for comparison. Note the high stability of the atmospheric column from the top of the ABL to about 5 km.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 105 105 4
PDF Downloads 97 97 5

On Turbulence and Mixing in the Free Atmosphere Inferred from High-Resolution Soundings

View More View Less
  • 1 Department of Meteorology, and Geophysical Fluid Dynamics Institute, The Florida State University, Tallahassee, Florida
  • | 2 Department of Aerospace Engineering, University of Colorado, Boulder, Colorado
© Get Permissions
Full access

Abstract

Mixing in the free atmosphere above the planetary boundary layer is of great importance to the fate of trace gases and pollutants. However, direct measurements of the turbulent dissipation rate by in situ probes are very scarce and radar measurements are fraught with uncertainties. In this paper, turbulence scaling concepts, developed over the past decades for application to oceanic mixing, are used to suggest an alternative technique for retrieving turbulence properties in the free atmosphere from high-resolution soundings. This technique enables high-resolution radiosondes, which have become quite standard in the past few years, to be used not only to monitor turbulence in the free atmosphere in near–real time, but also to study its spatiotemporal characteristics from the abundant archives of high-resolution soundings from around the world. Examples from several locations are shown, as well as comparisons with radar-based estimations and a typical Richardson number–based parameterization.

Corresponding author address: Carol Anne Clayson, Department of Meteorology, and Geophysical Fluid Dynamics Institute, 404 Love Building, The Florida State University, Tallahassee, FL 32306. Email: clayson@met.fsu.edu

Abstract

Mixing in the free atmosphere above the planetary boundary layer is of great importance to the fate of trace gases and pollutants. However, direct measurements of the turbulent dissipation rate by in situ probes are very scarce and radar measurements are fraught with uncertainties. In this paper, turbulence scaling concepts, developed over the past decades for application to oceanic mixing, are used to suggest an alternative technique for retrieving turbulence properties in the free atmosphere from high-resolution soundings. This technique enables high-resolution radiosondes, which have become quite standard in the past few years, to be used not only to monitor turbulence in the free atmosphere in near–real time, but also to study its spatiotemporal characteristics from the abundant archives of high-resolution soundings from around the world. Examples from several locations are shown, as well as comparisons with radar-based estimations and a typical Richardson number–based parameterization.

Corresponding author address: Carol Anne Clayson, Department of Meteorology, and Geophysical Fluid Dynamics Institute, 404 Love Building, The Florida State University, Tallahassee, FL 32306. Email: clayson@met.fsu.edu

1. Introduction

Enormous effort has been expended over the past few decades to understand and model mixing within the active geophysical boundary layers, the atmospheric boundary layer (ABL) over both land and sea, and the oceanic mixed layer (OML). Mixing in these planetary boundary layers (PBLs) is invariably turbulent, and large eddy simulation (LES; e.g., Moeng and Sullivan 1994) and second-moment turbulence closure models (e.g., Mellor and Yamada 1982; Kantha 2003; Kantha and Clayson 1994, 2004) have been employed to model mixing in the PBL. While these models do have some deficiencies and can be improved further, overall they have been reasonably successful in modeling the mixing within the active PBL.

Mixing in the free atmosphere above the ABL and the deep ocean below the OML has been a tougher challenge, because it is episodic and hard to model using conventional methods, such as second-moment closure. The generation mechanisms in these generally stably stratified regions are not as well understood. The current consensus is that the mixing in the deep ocean is mainly internal wave driven, although in many places double diffusion is also important. In the free atmosphere, the mean shear and associated shear instabilities, as well as internal wave breaking, drive turbulence.

Mixing caused by the jet stream at the tropopause and internal waves breaking in the vicinity of the tropopause are potentially important in the exchange of trace gases such as ozone and greenhouse gases between the troposphere and the stratosphere. Clear air turbulence (CAT) is also of importance from the point of view of the safety and comfort of air travelers. A dearth of measurements contributes to our lack of knowledge about the characteristics of regions of intense small-scale mixing, which gives rise to questions regarding the role of this mixing in climate and weather prediction (e.g., Wilson 2004). There is significant temporal and spatial variability in the eddy dissipation rate aloft, precluding the use of arbitrary constant values in numerical models. A routine way of deducing mixing in the vicinity of the tropopause and elsewhere in the troposphere–stratosphere from observational data would be important in improving the skill of NWP and climate models.

The space–time intermittency of the turbulence aloft gives rise in the literature to dissipation rates that can be several orders of magnitude different. Given this lack of agreement, atmospheric models use parameterizations that are rather ad hoc. Because of the statically stable nature of the free atmosphere, Richardson number–based parameterizations are common. For example, Hong and Pan (1996) use the mixing length formulation
i1520-0426-25-6-833-e1
where Rig = (g/ρ)(dρ/dz)/(dU/dz)2 is the gradient Richardson number and (1/ℓ) = (1/κz) + (1/ℓ0) is the mixing length; and ρ is density, U is the magnitude of the horizontal velocity, z is the vertical coordinate, g is the gravitational acceleration, and κ is the von Kármán constant. Far above the PBL, ℓ ≃ ℓ0, with the ℓ0 value taken arbitrarily as 30 m based on aircraft measurements in the atmosphere (Kim 1991), although a value of 250 m has also been used.

Inferring turbulent mixing characteristics in the free atmosphere has been difficult. In situ measurements using turbulence probes on research aircraft are both costly and few in number. Radar probing using Doppler radars is more convenient but often subject to uncertainties (see, e.g., Cohn 1995; Hooper and Thomas 1998). One radar technique that uses backscattered power to estimate the turbulence kinetic energy (TKE) dissipation rate requires auxiliary data in the form of simultaneous soundings of the atmospheric column. The spectral broadening approach suffers from ambiguities in removing contamination from causes other than turbulence. Both methods are not always reliable in the troposphere (see section 3 below).

Consequently, accurate and reliable turbulence measurements in the free atmosphere have been very hard to make (e.g., Cohn 1995; Hooper and Thomas 1998). Routine monitoring, an important goal of CAT research, has been all but impossible. This paper suggests a simple alternative technique that applies concepts used in ocean mixing studies to high-resolution radiosonde data to derive the TKE (eddy) dissipation rate ɛ and the eddy diffusivity (viscosity) K from sounding data. This technique would be revolutionary, because it will permit not only routine near-real-time monitoring of turbulence in the free atmosphere, but also the exploration of its spatiotemporal characteristics based on the extensive archive of high-resolution (2 s or better) soundings from around the world. Section 2 describes the concept in some detail for the benefit of atmospheric scientists who may not be familiar with it. Section 3 describes techniques currently used for inferring dissipation rate in the atmosphere, while section 4 provides a proof of concept. Section 5 discusses the possibility of synergistic use of radiosondes and radars for studying free atmosphere turbulence, and section 6 concludes with a few remarks.

2. Concepts related to mixing in the ocean

Mixing in the deep regions away from the OML and the bottom boundary layer (BBL) is thought to be primarily due to the internal wave field prevalent there. The sources of this internal wave field are presumed to be mostly winds acting at the ocean surface and tides acting in the interior. Flow over underwater topography, such as ridges and seamounts, is also a source of internal waves. Immediately below the active OML, shear instabilities also generate turbulence.

Nonlinear interactions among internal waves of different wavenumbers is ubiquitous and quite rapid (Phillips 1977; Kantha and Clayson 2000), with a consequent redistribution of energy in spectral space that produces a roughly universal spectral shape, known as the Garrett–Munk spectrum (e.g., Garett and Munk 1975), away from the immediate vicinity of sources such as underwater ridges. Many observations of deep-ocean mixing have been made over the past few decades using a dedicated instrument called the microstructure profiler. As the instrument descends at the rate of around 0.5 m s−1, the microstructure sensor measures the force fluctuations resulting from fluctuations in the lateral component of velocity, which can then be related to the fluctuations of vertical shear in the water column. Assuming isotropy of turbulence at small scales and using a theoretical shear spectrum to fill the spectral space not resolved by the probe, it is possible to derive ɛ and deduce K, provided care is taken to prevent contamination by sources other than oceanic turbulence.

The microstructure instrument also carries high-resolution temperature and salinity sensors, which make it possible to derive a high-resolution density profile in the water column. This “instantaneous” density profile is generally statically stable but contains regions of local overturns resulting from turbulence, whatever their origin. Thorpe (1977) came up with a very simple means of estimating the scale of the overturns from the density (or potential temperature, where salinity is not important) profile. The method consists of rearranging the measured density (potential temperature) profile into a monotonic profile that contains no inversions. Suppose the profile contains n samples of density and suppose the sample at depth zn needs to be moved to a depth zm in order to create a stable profile. The resulting displacement d = |zmzn| is known as Thorpe displacement (we will simply call it the displacement or displacement distance), whose root-mean-square (rms) value is called the Thorpe scale LT (we will use the Thorpe scale and the displacement interchangeably henceforth). It is possible to estimate the displacements in the overturn regions using a simple routine to sort the high-resolution density profile from the sensors into a stable monotonic density profile. For a detailed description, see Thorpe (2005, his Fig. 6.2, p. 176).

The Thorpe scale is indicative of the local overturning scale in the water column and decades of oceanic microstructure research have shown it to be highly correlated with the most important length scale of turbulence in stably stratified fluids, the Ozmidov scale Lo = (ɛ/N3)1/2 (with N = (g/ρ)(∂ρ/∂z) being the local buoyancy frequency), which is indicative of the maximum turbulence length scale possible in a stably stratified fluid [see Dillon (1982) for a lucid discussion of the Thorpe scale and related issues],
i1520-0426-25-6-833-e2
where c is an empirical constant. Plots of observed LT versus Lo in the ocean show considerable scatter (Osborn 1980; Dillon 1982; Wesson and Gregg 1994; Moum 1996; Ferron et al. 1998; Gargett 1999) that is typical of all microstructure measurements, and so the value of c does involve some uncertainty. The scatter may very well be due to the noise and the resolution of the sensors, but the procedure and bin size used in determining LT may also be important factors. Also for turbulence generated by shear instability, c would depend on the stage of turbulence and would increase with time (Smyth and Moum 2000). Nevertheless, the proportionality between LT and Lo is well established and beyond dispute.
It follows immediately from the definition of the Ozmidov scale that
i1520-0426-25-6-833-e3
where CK = c2 [see Caldwell (1983), Galbraith and Kelly (1996), and Fer et al. (2004); and see Thorpe (2005) for a discussion of this equation by the inventor of the scale himself]. Thus, an estimate of the Thorpe scale provides a reliable means of inferring ɛ, provided the background value of local N can also be estimated. The best way to deduce the background value of N is from the sorted monotonic density (potential temperature) profile (Dillon 1982; Thorpe 2005).

In Wesson and Gregg (1994) measurements of c range between 0.25 and 4.0, yielding a value ranging between 0.0625 and 16 for CK! Other values are quoted in literature for CK, for example, 0.64 from Dillon (1982), 0.9 from Ferron et al. (1998), 0.91 from Peters et al. (1988), and 1.15 from Gavrilov et al. (2005), taken from very high-resolution atmospheric data. Based on our own measurements cited below and our best judgment, we have chosen to use a value of 0.3 for CK in the radiosonde analyses.

Strictly speaking, the Thorpe scale should be computed as the rms value of an ensemble of measurements. However, logistical considerations preclude costly multiple microstructure measurements in the deep ocean so that true Thorpe-scale profiles cannot always be obtained from displacement profiles. Instead, the rms value either is taken over a vertical bin containing several displacement values or, often more simply, displacement scales from a single microstructure measurement are used instead of the Thorpe scales in Eq. (3). In the latter case, the value of CK may be different than the traditional values quoted above.

Knowing ɛ and N, through assumption of local equilibrium and hence a balance between production and dissipation terms in the TKE equation, it is possible to infer the turbulent or eddy diffusivity (viscosity) from
i1520-0426-25-6-833-e4
where γ is the so-called mixing efficiency. Its value is thought to vary between 0.2 and 1 (Fukao et al. 1994). It is approximately equal to Rif/(1–Rif), where Rif is the flux Richardson number, an important parameter in turbulence that is indicative of the ratio of buoyancy production to shear production. Using Rif = 0.2, the most commonly used value, gives γ = 0.25. A value of 1/3 corresponding to Rif of 0.25 has also been used in microstructure literature.

High intermittency is characteristic of turbulence in the free atmosphere and the deep ocean. All ɛ measurements, irrespective of whether made in the atmosphere using aircraft and radars or in the ocean using microstructure profilers, show one to two decades of variability and so significant scatter is characteristic of these measurements (see, e.g., Fig. 1). Given this scatter, a factor of 2–3 uncertainty in the value of CK is not unexpected.

A more recent example (for an earlier example, see Fig. 10 of Ferron et al. 1998) confirming the validity of Eq. (3) in the oceans (see also Carter and Gregg 2002) is Carniel et al. (2007, manuscript submitted to J. Geophys. Res., hereafter CAR), who have made simultaneous microstructure and high-resolution T and S measurements in the southern Adriatic Sea. Figure 1 compares measured ɛ values with those inferred from the Thorpe displacements below the active mixed layer at various stations along the Italian coast. Note the large scatter typical of these measurements, suggesting the desirability of obtaining proper ensemble averages to circumvent the intermittency problem.

There is no reason why this technique cannot be applied to the atmosphere. High-resolution sounding data can be used to obtain the potential temperature profile, from which Thorpe/displacement scales can be inferred by sorting of the profile, and ɛ and K deduced from Eqs. (3) and (4). The goal of this paper is to do precisely that and demonstrate that the technique can be applied to routinely collected high-resolution sondes. It must be stressed at the outset that this technique is accurate only when the background stratification is statically stable, and does not work well for convectively turbulent regions such as the daytime ABL (Thorpe 2005). Note also that Gavrilov et al. (2005) have applied this technique recently with considerable success to very high-resolution (10 cm) sonde measurements of turbulence during which simultaneous aircraft and radar observations were also made. They report excellent agreement between the aircraft retrievals of the Thorpe scales and those obtained using the above-mentioned technique of reordering the potential temperature profile but have not yet reported on comparisons between the radar and sonde measurements. They ignored humidity and hence their study is valid only in the upper troposphere and stratosphere. The vertical resolution of their sondes was 10 cm, but Thorpe length scales were computed by averaging over a bin of 12.8 m. However, such very high-resolution sondes are not in routine use. The best resolution widely available is around 10 m, obtainable from two second sondes. Even these are comparatively scarce, confined to special observational campaigns and more readily available and archived high-resolution sonde data are from 6-s sondes, about 30 m. The question then is whether this sonde database is helpful in improving our knowledge of the turbulence in the free atmosphere. We believe it is and provide the proof of concept in this paper as detailed below.

It is widely recognized that the turbulence in the free atmosphere, especially the stratosphere, is intermittent, inhomogeneous, and anisotropic, with layers of intense turbulence embedded in regions of weak or no turbulence (Gage et al. 1980; see also Gage 1990). Even more importantly, this turbulence structure in the free atmosphere resembles the turbulence structure in the stably stratified upper ocean remarkably well (e.g., Woods 1968). Kelvin–Helmholtz shear instabilities have been observed in both the inversion at the top of the ABL and the thermocline immediately below the OML. Both the free atmosphere and the deep ocean harbor internal waves (IWs) propagating into the region away from sources in and around the ABL and OML, respectively, The near-universal spectrum in the deep oceans away from IW sources such as underwater ridges may also hold in the free atmosphere (see, e.g., Gage and Gossard 2003), although the free atmosphere differs from the ocean because of the presence of strong winds and consequently strong mean shear. Nevertheless, the many similarities between the free atmosphere and the deep ocean are what give us confidence in applying the methodology and results of observations of turbulence in the ocean to the atmosphere.

3. Turbulence in the free atmosphere above ABL

As mentioned earlier, it has been a very hard task to measure ɛ and K in the free atmosphere; Gage et al. (1980) summarize observations of ɛ in the free atmosphere from a variety of measurement techniques and analyses (reproduced here as Fig. 2). To our knowledge, no modern version of this figure exists in open literature, probably because the emphasis of atmospheric scientists has shifted to deducing ɛ from Doppler wind-profiling radars.

Figure 2 includes estimations of ɛ from a variety of methods, for example, 1) in situ aircraft measurements in the troposphere and stratosphere by Lilly et al. (1974); 2) a study by Kung (1966) of the kinetic energy budget of the large-scale atmosphere; 3) an analysis of the variability of rawinsonde winds by Ellsaesser (1969); 4) a time series analysis of winds measured by an anemometer suspended below a floating balloon by Cadet (1977); and 5) a study of dispersion of smoke and bomb clouds by Kellogg (1956) and Wilkins (1958). Cohn (1995) states that, “The existence of such a variety and the limited experimentation with each method suggest the degree of difficulty of the measurement [of ɛ].” He also notes, “There has been no systematic intercomparison of the many methods.” These statements are still valid today and attest to the sorry state of affairs in free atmosphere turbulence. Spatiotemporal characteristics of turbulence in the global free atmosphere remain largely unknown.

Figure 2 shows that ɛ in the free atmosphere ranges between 10−6 and 10−3 m2 s−3, with higher values in the troposphere and much smaller values (typically around 2 × 10−4) in the more stable stratosphere, for undisturbed conditions. These values are exceeded locally by two to three orders of magnitude in the vicinity of jet streams, thunderstorms, and intense wave activity as, for example, in the lee of mountain ranges. Near convective storms, ɛ can exceed 10−1 m2 s−3 (Gage et al. 1980). Trout and Panofsky (1969) classify clear air turbulence with ɛ values of 3 × 10−3, 8 × 10−3, and 6.75 × 10−2 m2 s−3 as light, moderate, and severe, respectively.

Using Eq. (3), and using a value of 0.005 s−1 for the average buoyancy frequency N and 5 × 10−4 for average ɛ, we calculate a Thorpe scale of roughly 115 m in the troposphere. This scale is well resolved by both 2- and 6-s soundings (10- and 30-m vertical resolution, respectively). Taking a value of 0.015 s−1 for average N and 1.5 × 10−5 for average ɛ, we calculate a Thorpe scale of roughly 5 m in the stratosphere, which makes even the 2-s soundings less than optimal for complete accuracy. Obviously, sonde data should be saved in the future at as high a resolution as possible (modern sondes can provide ½–1-s data, and hence 2.5–5-m resolution), and the ascent velocity could be decreased to provide an even higher resolution. However, our goal here is to show the usefulness of the Thorpe displacement scale in providing reasonable estimates of ɛ, and thus we will show results using both the 2- and 6-s soundings, even though the vertical resolution might be marginal in some cases in that overturns smaller than a few tens of meters cannot be well detected.

Some attempts at understanding turbulence in the free atmosphere have been made since the 1970s, a large part of which involved the development and use of powerful UHF and VHF Doppler (wind profiling) radars (typical frequency of around 50 MHz, compared to 1 GHz or more for conventional weather radars), also called mesosphere–stratosphere–troposphere (MST) and stratosphere–troposphere (ST) radars. The backscattered signals from these radars contain information on the turbulence properties in the atmospheric column. Wilson (2004) reviews the methods of inferring turbulent diffusivity from MST radar measurements. The reader is referred to this excellent review for detailed background information on turbulence scales as well. The reader is also referred to Cohn (1995), Bertin et al. (1997), and Hooper and Thomas (1998) for a detailed discussion of radar methods.

There are basically two radar methods. The first method uses the measured profile of the refractivity turbulence structure parameter C2n but requires additional information on the humidity and stability values in the atmospheric column. Consequently, simultaneous soundings are needed in addition to a well-calibrated radar. Nevertheless,
i1520-0426-25-6-833-e5
relates ɛ to the mean refractivity turbulence structure constant C2n, where F is the fraction of the radar volume that is turbulent and M, the gradient of the radio index of refraction, is given by (e.g., Gage et al. 1980)
i1520-0426-25-6-833-e6
where P is the pressure, T is the temperature, q is the humidity, and θ is the potential temperature (Cohn 1995; Hocking 1985). A major problem with this method is in estimating the fraction of the active turbulence in the volume the radar is measuring.
The second method (Hocking 1985; Cohn 1995; Worthington 1998) relates ɛ to the width of the received Doppler spectrum,
i1520-0426-25-6-833-e7
where δ is the spectral width (Hocking 1985; Worthington 1998). Then from Eq. (4) K = 0.15δ2 N−1, although a value of 0.05 has also been used for the proportionality constant (Whiteway et al. 2003). Thus, the proportionality constant in Eq. (7) should be regarded as uncertain to a factor of at least 2.

The Doppler radar return signal power is spread across a range of Doppler velocities, an effect known as beam broadening, with the mean Doppler shift corresponding to the mean vertical velocity. Additional broadening occurs if the vertical velocity fluctuates as a result of turbulent motions. In principle, therefore, the intensity of the turbulence can be derived from the radar return spectral widths once the effects of beam broadening are corrected. Because the beam broadening contribution to the observed spectral width increases with increasing wind speed, the likelihood of being able to extract turbulence information decreases with increasing wind speed. A narrow beamwidth is very desirable. Also, the turbulence signatures tend to be much clearer in the lower stratosphere than in the troposphere. In the troposphere the spectral widths can be enhanced by factors such as convection (Hooper et al. 2005) and precipitation (McDonald et al. 2004), making identification of turbulence in the radar signals more difficult, thus somewhat dampening the euphoria in the 1990s about the ability to use radars to infer mixing in the free atmosphere. It appears that the technique is most reliable in only the lower stratosphere (D. A. Hooper 2007, personal communication). See Rao et al. (1997) for an example of the study of turbulence in the free atmosphere through the use of an MST radar.

Wilson (2004) and Wilson et al. (2005) point out the distinction between TKE dissipation rate ɛk and the dissipation rate of potential energy ɛp, whose ratio is the so-called mixing efficiency γ in Eq. (4), γ = ɛpk. The structure constant C2n deduced from the “power method” can be related to either through the use of γ. Wilson (2004) also discusses uncertainties in the value of γ, which impact the derivation of diffusivity from the dissipation rate [Eq. (4)]. It is a common practice in oceanographic microstructure studies to consider γ as a constant with a value of around 0.2. Atmospheric studies also consider γ as a constant, with recent values given also as roughly 0.2 (Wilson et al. 2005).

Radar probing of atmospheric turbulence is thus fraught with uncertainties but has many potential advantages over in situ measurements, including routine monitoring. The accuracy of retrievals is uncertain because of the absence of simultaneous in situ measurements needed to calibrate/validate radar retrievals (however, see Whiteway et al. 2003). This is the primary reason why independent retrievals of ɛ from high-resolution data from collocated sounding stations hold the promise of better calibration of radar methods and hence their more routine use in monitoring CAT. Synergy between the two is likely, because the radar probing is most accurate in the stratosphere but less so in the troposphere, while the sounding technique is less accurate in the stratosphere, at least until even higher-resolution soundings become available (see below).

4. Inferring turbulence properties from high-resolution radiosonde data

Radiosonde/rawinsonde soundings have been used for decades to measure the temperature, humidity, and wind velocity profiles in the atmosphere. Because the data from these soundings are indispensable to the initialization of the atmospheric state for numerical weather forecast purposes, there is an extensive network of about 800 radiosonde stations around the globe (around 90 in the United States). Two-thirds of the stations send two or more radiosondes daily (0000, 0600, 1200, 1800 UTC), while one-third launch one radiosonde every day. These data are collected, quality controlled, and decimated to rather coarse standard levels for use in initializing NWP models around the world. The data are also routinely archived and millions of soundings from around the globe are available and have been used to study the temporal and spatial distribution of properties (potential temperature and winds) characterizing the global atmosphere.

Until a few years ago, the original high-resolution radiosonde data, from which the coarse standard level data for NWP model initialization are derived, were routinely discarded. However, studies such as of gravity wave breaking in the free atmosphere (e.g., Whiteway et al. 2003) have given rise to the practice of routine collection and archival of high-resolution radiosonde data. Consequently, abundant 6-s radiosonde data are available over the past several years. A more limited number of 2-s (and even higher resolution) data are also available during special observational campaigns [e.g., Fronts and Atlantic Storm-Tracks Experiment (FASTEX): 2 s; and, more recently, Terrain-Induced Rotor Experiment (T-REX): 1 s]. Often these high-resolution sounding stations are located near VHF Doppler radars (but unfortunately, not necessarily collocated), because knowledge of the atmospheric state is often important to the analysis and interpretation of radar measurements. Inferring TKE dissipation rates from radar signals using the power method (section 3) requires knowing the temperature and humidity in the atmospheric column above the radar.

The routine collection and archiving of high-resolution radiosonde data are extremely fortuitous to the study of turbulence and turbulent mixing in the atmospheric column. Because the method is based on inference of the scale of overturning in the atmospheric column, and because this scale is typically a few tens of meters, decimated data are useless for this purpose. Because the ascent rate of radiosondes is about 5 m s−1, on the average, 2-s data can provide a resolution of about 10 m and hence are sufficient in the troposphere and acceptable, though marginal in the stratosphere. The 6-s data are also useful, although the resolution is only about 30 m and hence not very satisfactory in the stratosphere, where the local stability is high and the overturning scales smaller. The resolution needed for the stratosphere would be around 1/2 s (or 2.5 m). In this study, we have used both 2- and 6-s data and both provide roughly similar results, although the 2-s data are clearly more desirable (6-s data provide spotty results in the stratosphere, as shown below).

The goal of this study is confined to demonstrating the usefulness of the existing sonde database. It is not our intent to explore the spatiotemporal characteristics of turbulence in the free atmosphere around the globe, which is a monumental task involving careful analyses of millions of 6- and 2-s soundings available now from around the world (a task that nevertheless needs to be undertaken). Therefore, we present only two examples here: one using 2-s data collected in 1997 during FASTEX, and the other using 6-s data taken at Tallahassee, Florida, and Denver, Colorado, during 2005.

Instrument noise is an important issue in the application of the Thorpe technique to the deep ocean, because the stratification there is very nearly neutral and hence very weakly stable [see Ferron et al. (1998) for procedures to account for instrument noise]. However, it is less of an issue in the free atmosphere, where the mean overall stratification is quite strong (N ∼ 0.01–0.02 s−1). Consequently, complex techniques based on techniques such as wavelet analyses (Piera et al. 2002), while desirable, are not as useful. To eliminate spurious overturn values, and at the same time keep it simple, we discarded all overturns that were less than 1.1 times the vertical resolution of the sonde. This procedure appears to eliminate spurious retrievals resulting from noisy data. A 10-point running mean was used to damp out any residual noise in the retrieved Thorpe displacement and dissipation rate and diffusivity values.

The temperature and RH were used to determine the potential temperature at each level. The resulting data were sorted to yield a stable monotonic potential temperature profile, and deduce the values of displacement d and N at each level. The values of the TKE dissipation rate ɛ and eddy viscosity K were then calculated at each level using Eqs. (3) and (4). The constant CK was taken to be 0.3. Note that this value could be refined once further simultaneous in situ measurements of ɛ by aircraft-borne turbulence sensors become available in the future. Note also that this constant could be a weak function of the sonde resolution, but we ignore this for the time being for lack of a better alternative, especially in view the one to two decade variability inherent to free atmosphere turbulence measurements.

Approximately 285 two-second FASTEX soundings collected between 6 January and 28 February 1997 at station 03953 (Ireland’s Valentia Observatory, located at 51.93°N and 10.25°W) were used to derive ɛ and K. Figure 3 shows atmospheric column properties derived from the sounding on 26 February 1997. The tropopause can be seen at a height of about 12 km. The free atmospheric column is susceptible to shear instability (as indicated by the gradient Richardson number falling below 0.25) at around 8- and 10-km altitudes. This is also the region of large displacements, which range from 100 to 200 m. The TKE dissipation rate is naturally high, reaching roughly 10−3 m2 s−3, indicating strong CAT at these levels. The corresponding eddy diffusivity is also high, reaching values of 1 m2 s−1. The deduced dissipation rates are consistent with past observations (see Fig. 2). It is also worth noting that the Hong and Pan (1996) formulation yields consistently higher values as can also be seen from further examples below. The high mixing region is also not well depicted in this formulation.

The ABL is well mixed, and the sounding shows a double-layer structure at the bottom, typical of boundary layers topped by a stratocumulus cloud deck driving the turbulence in the upper layer (Johansson et al. 2005). Note the large Thorpe displacements in the upper mixed layer. In unstable and well-mixed regions, the sorting technique yields displacements that are of the same magnitude as the thicknesses of these regions. The technique is, therefore, less reliable in well-mixed regions with weak or negative stability, as opposed to regions with strong background stable stratification. In weakly stable regions, there is also the possibility of potential contamination by noise, instrument or otherwise (Piera et al. 2002).

Figure 4 shows atmospheric column properties derived from the sounding on 24 January 1997. As indicated by the gradient Richardson number falling below 0.25, the region immediately below the tropopause is shear unstable, with displacements as large as 100 m leading to TKE dissipation rates exceeding 10−3 m2 s−3, indicating intense CAT at these levels. The corresponding eddy diffusivity exceeds 1 m2 s−1.

Figures 3 and 4 illustrate the feasibility of retrieving the TKE dissipation rate and eddy diffusivity profiles in the free atmosphere from 2-s radiosonde sounding data. Unfortunately, there are no concurrent independent observations of turbulence either from Doppler radars or aircraft, and while the values appear reasonable, there is no way to assess the accuracy of the results. It is worth noting that the value of the constant CK is uncertain to perhaps a factor of 2–5, and simultaneous, independent, and accurate measurements of ɛ are urgently needed to calibrate this constant. Such a campaign should be undertaken at the earliest opportunity.

To further explore the viability of the technique, we retrieved turbulence properties from 209 of the 285 two-second FASTEX soundings that reached above 20-km altitude. Figure 5 shows the cumulative plot. The overall means are also shown. The buoyancy frequency varies around 0.005 s−1 in the troposphere, with values reaching 0.04 s−1 near the tropopause and around 0.015 s−1 in the stratosphere. It can be seen that the mean value of ɛ in the upper troposphere is around 10−4 m2 s−3, falling to roughly 1/3 this value in the lower stratosphere. These values are close to those in Fig. 2, thus confirming the efficacy of the technique. What is also interesting is the high variability of ɛ, ranging from 10−6 to 10−2 m2 s−3 in the troposphere. The values in the stratosphere are somewhat less, reaching about 10−3 m2 s−3. The mean eddy diffusivity in the troposphere ranges between 1 and 10 m2 s−1, with values decreasing to below 10−1 m2 s−1 in the stratosphere, consistent with some in situ measurements by Bertin et al. (1997) who found a range of 10−3 < Kθ < 10−1 m2 s−1 in the stratosphere. In contrast, the Hong and Pan (1996) formulation used in some NWP models yields a mean value of about 2 m2 s−1 throughout the atmospheric column and is also inconsistent with the notion that mixing levels in the stratosphere are usually an order of magnitude below those in the troposphere.

Figure 6 shows all 285 sounding retrievals plotted as a time series extending from 6 January to 28 February. Large displacements resulting in high levels of turbulence can be seen in the vicinity of the tropopause and less often in the troposphere. Turbulence levels intermittently reach high values in the stratosphere also, but stay well below the levels in the troposphere, consistent with what is known about turbulence levels in the troposphere and the stratosphere.

The 10-m resolution obtainable from 2-s data is satisfactory, but 1–2-m resolution would be even more desirable and technically feasible. However, 2-s data are limited. On the other hand, 6-s radiosonde data are abundant and routinely archived. This translates to roughly 30-m resolution, which is clearly not as useful in retrieving turbulence characteristics in the stratosphere, as can be seen from Fig. 7, which shows a sample retrieval from a 6-s sounding taken at Tallahassee on 31 July 2005. The retrieval is spotty in the stratosphere, although it appears to be acceptable in the troposphere itself. The shear-unstable region just below the tropopause is clearly seen, with displacements reaching as high as 100 m. As shown in Fig. 8, the mean values of displacements, and the TKE dissipation rate and eddy diffusivities over 609 soundings taken at Tallahassee during the year 1995, are quite consistent with those derived from the 209 two-second FASTEX soundings (cf. Figs. 5 and 8). This suggests that the vast archive of 6-s sounding data do have significant utility in exploring the spatiotemporal characteristics of turbulence in the free atmosphere, including the stratosphere, albeit to a lesser degree of accuracy.

Figure 9 shows the time series of turbulence at Tallahassee as derived from all the soundings in 2005 irrespective of the altitude to which the soundings reach (Fig. 8 uses only those that reached 20 km or more). Clearly, during summer, high turbulence levels exist in the 10–12-km band. One high-level event can be seen above the tropopause around midyear. This figure shows the utility of the technique in understanding the temporal variability of mixing in the free atmosphere at any given site where such soundings are available.

Data from the Denver station taken during 2005 shows the expected differences from Tallahassee soundings (Figs. 10 and 11). The summer of 2005 in Denver was of significant interest climatologically; for instance, in July 8 days set new records for high temperatures (with one tying for the highest temperature ever recorded in Denver). In the Denver region there is a clear maximum in displacement and dissipation and mixing nearer the surface from roughly 100 to 600 m. As can be seen from the time series, there is a seasonal signal with a greater incidence of mixing during the summertime, possibly related to the seasonal maximum of upslope winds that occur during the summer months (e.g., Losleben and Pepin 2000). There is a secondary maximum occurring in dissipation rate in a layer around 2000 m. However, unlike the FASTEX data, and more similarly to the Tallahassee data, this maximum occurs less from a fairly well defined and consistent layer of increased mixing but rather more from substantial but fairly inconsistent and diffuse episodes of increased mixing. In general, the pronounced maximum at the top of the troposphere evident in the Tallahassee soundings is not as evident here due to the much higher dissipation nearer the surface.

5. Comparison with radar measurements of ɛ

Finally, it may seem useful to compare the technique of retrieving ɛ in the free atmosphere from high-resolution radiosondes with the currently most prevalent method: radars. Unfortunately, the accuracy and reliability of radar retrievals of the vertical profiles of ɛ have not been thoroughly explored in the relevant parameter space of ambient conditions by direct comparison with independent, collocated, simultaneous in situ measurements by aircraft. One attempt by Whiteway et al. (2003) met with rather limited success. More recently, Gavrilov et al. (2005) made simultaneous measurements with radar, high-resolution radiosondes, and aircraft, but they have not yet reported on the intercomparison of their radar and aircraft measurements.

Simultaneous collocated radar and radiosonde observations are also rare, with exceptions being noted by Ghosh et al. (2003), Luce et al. (1997), and Zink et al. (2004). However, none of these studies made use of the Thorpe-scale concept to infer ɛ from the sonde measurements and compare with radar retrievals. Both Luce et al. (2002) and Zink et al. (2004) used high-resolution thermosondes to measure only the thermal structure. Worthington (1998) presents data from the Aberystwyth MST radar facility in Wales showing the high turbulence in the vicinity of the tropopause resulting from breaking mountain waves on two occasions between 1030 UTC 7 January and 1510 UTC 9 January. Unfortunately, the 2-s soundings are available only from Aberporth, 45 km to the southwest of Aberystwyth. The horizontal scale of wave breaking appears to be smaller, and hence collocated measurements are essential for a proper assessment of radar retrievals.

Nastrom and Eaton (1997) have made continuous measurement of the TKE dissipation rate ɛ in the free atmosphere at heights of 5–20 km for a 5-yr period between 1991 and 1995 using the 50-MHz Doppler radar at White Sands Missile Range, New Mexico. They studied its variability on time scales ranging from diurnal to seasonal. Nastrom and Eaton (2005) made similar measurements at Vandenberg Air Force Base, California, and compared it with MST measurements at Gadanki, India (Rao et al. 1997, 2001), and Kyoto, Japan (Fukao et al. 1994; Kurosaki et al. 1996). All of these studies find that ɛ lies between 10−4 and 10−3 m2 s−3 in the upper troposphere–lower stratosphere. The largest values at Vandenberg Air Force Base occurred between 8 and 16 km. However, while the results are believable, none of these studies have been corroborated by independent in situ measurements.

Cohn (1995) presents ɛ profiles derived from UHF radar observations at Millstone Hill, Massachusetts, using both the power and spectrum-broadening techniques. He uses the Cross-Chain Loran Atmospheric Sounding System (CLASS) soundings from Hanscom Air Force Base, Massachusetts, 25 km to the southeast decimated to 300-m resolution to derive the auxiliary information on the atmospheric column needed to use the “power method.” Figure 12 shows the retrieved profiles by the two methods for three periods: 2010–2057 UTC 24 February (top panel), 0222–0301 UTC 25 February (middle panel), and 1719–1813 UTC 25 February (bottom panel; see also Cohn 1995; his Figs. 5, 3, and 7, respectively). The corresponding sondes were launched at 2014 UTC 24 February, 0212 UTC 25 February, and 1719 UTC 25 February (Cohn 1995; his Figs. 4, 2, and 6, respectively).

We were able to obtain computer printouts of the original 50–60-m-resolution sounding data and retrieve ɛ independently using the technique discussed earlier. Figure 13 shows the ɛ profile in the atmosphere from 0- to 10-km altitude, along with the gradient Richardson and displacement distance profile. Because of the rather coarse resolution (50–60 m), the accuracy of the retrieval suffers except in strongly overturning, shear-unstable regions near the tropopause. These regions are clearly delineated by the gradient Ri falling below the canonical value of 0.25. The high ɛ regions correlate very well with these regions, the levels also being reasonable. Because of the poor sonde resolution, we were unable to retrieve ɛ in the lower stratosphere. The atmospheric column appears to be mostly stable from the top of the ABL to about 5 km; however, higher-resolution (2 s) data might have been able to pick up local turbulent regions. While these results are consistent with radar retrievals, the comparison is rather inconclusive, especially since the sondes were not collocated with the radar. The need for high-resolution sondes collocated with the radar is quite clear.

6. Concluding remarks

Mixing in the free atmosphere above the PBL is of great importance to the fate of both natural trace gases and anthropogenic greenhouse gases and pollutants in the troposphere and the stratosphere. Yet, direct observations of turbulence by in situ probes are very scarce. Radar measurements are also few and often subject to uncertainties. Atmospheric scientists have long lamented the difficulty of making turbulence measurements in the free atmosphere. Consequently, routine monitoring of atmospheric turbulence, even in nonreal time, has proved elusive. Spatiotemporal distribution of turbulence characteristics in the global atmosphere above the PBL remains largely unknown.

NWP models use mostly Richardson number–modified mixing length parameterizations, with arbitrary choices for the mixing length, and with consequent impact on weather predictions and long-term runs for applications to climate study. Clearly, a reliable, accurate, easy, and routine means of determining turbulent mixing in the free atmosphere is highly desirable and potentially useful in improving long-term simulations of the atmospheric state.

In this paper, we have made use of turbulence scaling concepts developed over the past few decades for application to mixing in the ocean to suggest an alternative technique for retrieving turbulence properties in the free atmosphere from high-resolution soundings. This technique enables high-resolution radiosondes, which have become quite routine in the past few years, to be used to monitor turbulence in the free atmosphere in near–real time, as well as to study the spatiotemporal characteristics of turbulent mixing in the troposphere and the stratosphere from the abundant archives of high-resolution soundings from around the world. Synergistic use of radars and radiosondes to monitor turbulence in the upper troposphere/lower troposphere could have potential benefits for air travelers in that high CAT regions can be more readily and reliably identified. A better understanding of turbulent mixing in the atmospheric column would also be useful in assessing and improving the skill of NWP and climate models. In addition to the temperature data from sondes, turbulence data can also be assimilated into NWP models to increase their skill.

Intermittency remains a vexing problem in the understanding and predicting turbulence in the free atmosphere. All radar and aircraft measurements in the free atmosphere (and microstructure measurements in the deep ocean) show one to two decades of variability in the observed values of the eddy dissipation rate. As such, it is worth noting that a probe, such as a microstructure sensor in the ocean and a radiosonde in the atmosphere, provides only one realization of the turbulence field (the spatiotemporal sampling is also quite complex). Given the statistical nature of turbulence, it is essential to make several soundings in close proximity both in time and space, and create a proper ensemble average to obtain a clearer picture of the turbulence field, an additional step seldom under taken because of logistics and expense considerations. Availability of cheap radiosondes with automated launching could overcome this problem. However, it would be necessary to modify the sonde transmitter–receiver system so that it is possible to launch multiple sondes simultaneously, and receive and process the data concurrently to generate the needed ensemble averages.

Dropsondes deployed from aircraft and drifting high-altitude balloons are also used to measure atmospheric properties during weather events such as cyclogenesis and hurricanes. High-resolution dropsondes would enable turbulence properties also to be measured during these events. In fact, any method, remote or in situ, that provides high-resolution potential temperature profiles can be used to retrieve turbulent mixing profiles.

A 1–2-m-resolution radiosonde is highly desirable for delineating turbulence zones in the stratosphere (5–10 m is quite adequate in the troposphere). This is technically feasible because the sondes can be slowed down to a 2–3 m s−1 ascent rate and the temperature and humidity can be sampled at 1/2-s intervals. It is, however, essential to make sure that the temperature and humidity sensors have a sufficiently fast response to enable this sampling rate to be achieved. It is possible to build and launch millions of simple, cheap turbusondes every year, which only need to contain pressure–temperature–humidity (PTU) sensors. This would revolutionize the field by making it possible to routinely monitor and map the turbulence characteristics in the troposphere and the stratosphere, and possibly even the mesosphere. It could also help significantly improve the skill of NWP and climate models through assimilation of mixing data or simply through assessment of the adequacy of the mixing models incorporated into NWP models and appropriate refinements.

Acknowledgments

Our thanks to Centre National de Recherches Meteorologiques of Meteo-France for giving us access to the 2-s high-resolution FASTEX soundings and NSF-sponsored SPARC archive at SUNY for access to 6-s data. Thanks to Dr. Laura Bianco for providing us with references to radar measurements. The help of Dr. David Hooper from Wales and Dr. Steve Cohn of NCAR during the early stages of this work in educating us about the nuances of MST radar measurements is gratefully acknowledged. The efforts of Dr. Cohn and the NCAR staff in retrieving the long-forgotten Bedford soundings are appreciated. Thanks to Dr. Rod Frehlich for bringing to our attention the recent use of the Thorpe-scale concepts by Gavrilov et al. (2005).

REFERENCES

  • Bertin, F., , Barat J. , , and Wilson R. , 1997: Energy dissipation rates, eddy diffusivity and the Prandtl number: An in situ experimental approach and its consequences on radar estimate of turbulent parameters. Radio Sci., 32 , 791804.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cadet, D., 1977: Energy dissipation within intermittent clear air turbulence patches. J. Atmos. Sci., 34 , 137142.

  • Caldwell, D. R., 1983: Oceanic turbulence: Big bangs or continuous creation? J. Geophys. Res., 88 , 74537550.

  • Carter, G. S., , and Gregg M. C. , 2002: Intense and variable mixing near the head of the Monterey submarine canyon. J. Phys. Oceanogr., 32 , 31453165.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cohn, S. A., 1995: Radar measurements of turbulent eddy dissipation rate in the troposphere: A comparison of techniques. J. Atmos. Oceanic Technol., 12 , 8595.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dillon, T. M., 1982: Vertical overturns: A comparison of Thorpe and Ozmidov length scales. J. Geophys. Res., 87 , 96019613.

  • Ellsaesser, H. W., 1969: A climatology of epsilon (atmospheric dissipation). Mon. Wea. Rev., 97 , 415423.

  • Fer, L., , Skogseth R. , , and Haugan P. M. , 2004: Mixing of the Storfjorden overflow (Svalbard Archipelago) inferred from density overturns. J. Geophys. Res., 109 .C01005, doi:10.1029/2003JC001968.

    • Search Google Scholar
    • Export Citation
  • Ferron, B., , Mercier H. , , Speer K. , , Gargett A. , , and Polzin K. , 1998: Mixing in the Romanche fracture zone. J. Phys. Oceanogr., 28 , 19291945.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fukao, S., and Coauthors, 1994: Seasonal variability of vertical eddy diffusivity in the middle atmosphere, 1. Three-year observations by the middle and upper atmosphere radar. J. Geophys. Res., 99 , 1897318987.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gage, K. S., 1990: Radar observations of the free atmosphere: Structure and dynamics. Radar in Meteorology: Battan Memorial and 40th Anniversary Radar Meteorology Conference, D. Atlas, Ed., Amer. Meteor. Soc., 534–565.

    • Search Google Scholar
    • Export Citation
  • Gage, K. S., , and Gossard E. E. , 2003: Recent developments in observation, modeling, and understanding atmospheric turbulence and waves. Radar and Atmospheric Science: A Collection of Essays in Honor of David Atlas, Meteor. Monogr., No. 30, Amer. Meteor. Soc., 139–150.

    • Search Google Scholar
    • Export Citation
  • Gage, K. S., , Green J. L. , , and VanZandt T. E. , 1980: Use of Doppler radar for the measurement of atmospheric turbulence parameters from the intensity of clear-air echoes. Radio Sci., 15 , 407416.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Galbraith, P. S., , and Kelly D. E. , 1996: Identifying overturns in CTD profiles. J. Atmos. Oceanic Technol., 13 , 688702.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Garett, C., , and Munk W. , 1975: Space-time scales of internal waves. Progress report. J. Geophys. Res., 80 , 291297.

  • Gargett, A. E., 1999: Velcro measurements of turbulent kinetic energy dissipation rate. J. Atmos. Oceanic Technol., 16 , 19731993.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gavrilov, N. M., , Luce H. , , Crochet M. , , Dalaudier F. , , and Fukao S. , 2005: Turbulence parameter estimations from high-resolution balloon temperature measurements of the MUTSI-2000 campaign. Ann. Geophys., 23 , 24012413.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ghosh, A. K., , Jain A. R. , , and Sivakumar V. , 2003: Simultaneous MST radar and radiosonde measurements at Gadanki (13.5°N, 79.2°E) 2. Determination of various atmospheric turbulence parameters. Radio Sci., 38 .1014, doi:10.1029/2000RS002528.

    • Search Google Scholar
    • Export Citation
  • Hocking, W. K., 1985: Measurement of turbulent eddy dissipation rates in the middle atmosphere by radar techniques: A review. Radio Sci., 20 , 14031422.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hong, S-Y., , and Pan H-L. , 1996: Nonlocal boundary layer vertical diffusion in a medium-range forecast model. Mon. Wea. Rev., 124 , 23222339.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hooper, D. A., , and Thomas L. , 1998: Complementary criteria for identifying regions of intense atmospheric turbulence using lower VHF radar. J. Atmos. Sol. Terr. Phys., 60 , 4961.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hooper, D. A., , McDonald A. J. , , Pavelin E. , , Carey-Smith T. K. , , and Pascoe C. L. , 2005: The signature of mid-latitude convection observed by VHF wind-profiling radar. Geophys. Res. Lett., 32 .L04808, doi:10.1029/2004GL020401.

    • Search Google Scholar
    • Export Citation
  • Johansson, C., , Hennemuth B. , , Bosenberg J. , , Linne H. , , and Smedman A. , 2005: Double-layer structure in the boundary layer over the Baltic Sea. Bound.-Layer Meteor., 114 , 389412.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kantha, L. H., 2003: On an improved model for the turbulent PBL. J. Atmos. Sci., 60 , 22392246.

  • Kantha, L. H., , and Clayson C. A. , 1994: An improved mixed layer model for geophysical applications. J. Geophys. Res., 99 , 2523525266.

  • Kantha, L. H., , and Clayson C. A. , 2000: Small Scale Processes in Geophysical Fluid Flows. Academic Press, 888 pp.

  • Kantha, L., , and Clayson C. A. , 2004: On the effect of surface gravity waves on mixing in an oceanic mixed layer. Ocean Modell., 6 , 101124.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kellogg, W. W., 1956: Diffusion of smoke in the stratosphere. J. Meteor., 13 , 241250.

  • Kim, J., 1991: Turbulent and gravity wave transport in the free atmosphere. Ph.D. dissertation, Oregon State University, 103 pp.

  • Kung, E. C., 1966: Large-scale balance of kinetic energy in the atmosphere. Mon. Wea. Rev., 94 , 627640.

  • Kurosaki, S., , Yamanaka M. D. , , Hashiguchi H. , , Sato T. , , and Fukao S. , 1996: Vertical eddy diffusivity in the lower and middle atmosphere: A climatology based on the MU radar observations during 1986–1992. J. Atmos. Terr. Phys., 58 , 727734.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lilly, D. K., , Waco D. E. , , and Adelfang S. I. , 1974: Stratosphere mixing estimated from high altitude turbulence measurements. J. Appl. Meteor., 13 , 488493.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Losleben, M., , and Pepin N. , 2000: Thermal wind frequency on the east slope of the Colorado Front Range. Eos, Trans. Amer. Geophys. Union, 82 .(Fall Meeting Suppl.), Abstract U22A-03.

    • Search Google Scholar
    • Export Citation
  • Luce, H., , Dalaudier F. , , Crochet M. , , and Sidi S. , 1997: An improved interpretation of VHF oblique radar echoes by a direct balloon Cn 2 estimation using a horizontal pair of sensors. Radio Sci., 32 , 12611268.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Luce, H., , Fuka S. , , Dalandier F. , , and Crochet M. , 2002: Strong mixing events observed near the tropopause with the MU radar and high-resolution balloon techniques. J. Atmos. Sci., 59 , 28852896.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • McDonald, A. J., , Carey-Smith T. K. , , Hooper D. A. , , Fraser G. J. , , and Lublow B. P. , 2004: The effect of precipitation on wind-profiler clear air returns. Ann. Geophys., 22 , 39593970.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mellor, G. L., , and Yamada T. , 1982: Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. Space Phys., 20 , 851875.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moeng, C-H., , and Sullivan P. P. , 1994: Comparison of shear- and buoyancy-driven planetary boundary layer flows. J. Atmos. Sci., 51 , 9991022.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moum, J. N., 1996: Energy-containing scales of turbulence in the ocean thermocline. J. Geophys. Res., 101 , 1409514101.

  • Nastrom, G. D., , and Eaton F. D. , 1997: Turbulence eddy dissipation rates from radar observations at 5–20 km at White Sands Missile Range, New Mexico. J. Geophys. Res., 102 , 1949519505.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nastrom, G. D., , and Eaton F. D. , 2005: Seasonal variability of turbulence parameters at 2 to 21 km from MST radar measurements at Vandenberg Air Force Base, California. J. Geophys. Res., 110 .D19110, doi:10.1029/2005JD005782.

    • Search Google Scholar
    • Export Citation
  • Osborn, T. R., 1980: Estimates of the local rate of vertical diffusion from dissipation measurements. J. Phys. Oceanogr., 10 , 8389.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peters, H., , Gregg M. C. , , and Toole J. M. , 1988: On the parameterization of equatorial turbulence. J. Geophys. Res., 93 , 11991218.

  • Phillips, O. M., 1977: The Dynamics of the Upper Ocean. Cambridge University Press, 336 pp.

  • Piera, J., , Roget E. , , and Catalan J. , 2002: Turbulent patch identification in microstructure profiles: A method based on wavelet denoising and Thorpe displacement analysis. J. Atmos. Oceanic Technol., 19 , 13901402.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rao, N. D., , Kishore P. , , Rao T. N. , , Rao S. V. B. , , Reddy K. K. , , Yarraiah M. , , and Hareesh M. , 1997: Studies on refractivity constant, eddy dissipation rate and momentum flux at a tropical latitude. Radio Sci., 32 , 13751389.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rao, N. D., , Rao T. N. , , Venkataratnam M. , , Tulasiraman S. , , and Rao S. V. B. , 2001: Diurnal and seasonal variability of turbulence parameters observed with Indian MST radar. Radio Sci., 36 , 14391457.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Smyth, W. D., , and Moum J. N. , 2000: Length scales of turbulence in stably stratified mixing layers. Phys. Fluids, 12 , 13271342.

  • Thorpe, S. A., 1977: Turbulence and mixing in a Scottish Loch. Philos. Trans. Roy. Soc. London, 286A , 125181.

  • Thorpe, S. A., 2005: The Turbulent Ocean. Cambridge University Press, 439 pp.

  • Trout, D., , and Panofsky H. A. , 1969: Energy dissipation near the tropopause. Tellus, 21 , 355358.

  • Wesson, J. C., , and Gregg M. C. , 1994: Mixing in the Camarinal Sill in the Strait of Gibraltar. J. Geophys. Res., 99 , 98479878.

  • Whiteway, J. A., , Pavelin E. G. , , Busen R. , , Hacker J. , , and Vosper S. , 2003: Airborne measurements of gravity wave breaking at the tropopause. Geophys. Res. Lett., 30 .2070, doi:10.1029/2003GL018207.

    • Search Google Scholar
    • Export Citation
  • Wilkins, E. M., 1958: Effective coefficients of diffusivity for atom bomb clouds at one thousand to two thousand miles. Eos, Trans. Amer. Geophys. Union, 39 , 5859.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilson, R., 2004: Turbulent diffusivity in the free atmosphere inferred from MST radar measurements: A review. Ann. Geophys., 22 , 38693887.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilson, R., , Dalaudier F. , , and Bertin F. , 2005: Estimate of the turbulent fraction in the free atmosphere from MST radar measurements. J. Atmos. Oceanic Technol., 22 , 13261339.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Woods, J. D., 1968: Wave-induced shear instability in the summer thermocline. J. Fluid Mech., 32 , 791800.

  • Worthington, R. M., 1998: Tropopausal turbulence caused by the breaking of mountain waves. J. Atmos. Sol. Terr. Phys., 60 , 15431547.

  • Zink, F., , Vincent R. A. , , Murphy E. , , and Cote O. , 2004: Comparison of radar and in situ measurements of atmospheric turbulence. J. Geophys. Res., 109 .D11108, doi:10.1029/2003JD003991.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Variation of the ɛ (x axis) with depth below the surface (y axis). The unit of ɛ, W kg−1 is the same as m2 s−3. Careful microstructure profiler-measured dissipation rates (thin lines) at station B90 in the southern Adriatic Sea at three different periods during the Dynamics of the Adriatic in Real Time campaign in August 2006 (CAR) compared with theoretical values. The thick green line denotes the overall (ensemble) mean of the multiple observations made at the same site. The thick black line denotes the mean theoretical values. Below the active mixed layer near the surface (characterized by nearly uniform dissipation rate profile), LT has been used along with Eq. (3) to obtain the theoretical values. Within the active mixed layer, indicated by the straight portion of the black curve near the surface, the inferred surface momentum and buoyancy fluxes from shipboard meteorological data have been used to deduce theoretical values (see CAR, see also Kantha and Clayson 2000). The dotted line denotes extrapolation of such theoretical values to regions below the mixed layer (figure adapted from CAR).

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 2.
Fig. 2.

Dissipation rate inferred by various studies are adapted from Gage et al. (1980). Circles refer to mean values. Solid curves refer to minimum detectable ɛ for the Sunset radar. Clear air turbulence categories are from Trout and Panofsky (1969).

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 3.
Fig. 3.

Atmospheric column properties inferred from the 2-s FASTEX sounding on 26 Feb 1997: (a) potential temperature, (b) wind velocity components U (blue) and V (red), (c) N, (d) mean shear, (e) Rig, (f) LT (m) inferred from the potential temperature profile, (g) ɛ inferred from the LT and buoyancy frequency N, and (h) K derived from ɛ (blue), and inferred from the Hong and Pan (1996) formulation (red). Note the high CAT region around 8- and 10-km altitudes that corresponds to high LT and Ri < 1/4, which is shown by a thick red line. Note that (c) shows the N deduced from the sorted profile; therefore, it shows no regions of negative stability. However Ri is computed from the unsorted profile and hence shows regions with Ri < 0.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 4.
Fig. 4.

As in Fig. 3, but on 24 Jan 1997. Note the high CAT region just below the tropopause, which corresponds to large LT, and Ri < 1/4, which is indicated by a thick red line.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 5.
Fig. 5.

Cumulative atmospheric column properties inferred from the 209 two-second FASTEX soundings taken between 6 Jan and 28 Feb 1997, reaching higher than 20-km altitude: (a) potential temperature, (b) N, (c) wind velocity components (m s−1) U (blue) and V (red), (d) Rig, (e) LT (m) inferred from the potential temperature, (f) ɛ (m3 s−2) inferred from the LT and N, (g) K (m2 s−1) derived from ɛ, and (h) K (m2 s−1) inferred from the Hong and Pan (1996) formulation. Thick green lines denote the overall means.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 6.
Fig. 6.

Time series of atmospheric column properties inferred from 285 two-second FASTEX soundings taken between 6 Jan and 28 Feb 1997: (a) Rig, (b) LT (m) inferred from the potential temperature, (c) log10 ɛ (m3 s−2) inferred from the LT and N, and (d) log10 K (m2 s−1) derived from ɛ. Note intermittent but high turbulence levels in the vicinity of the tropopause when the Ri value falls below 1/4 and less often in the troposphere. Turbulence reaches high levels in the stratosphere also (but significantly less than in the tropopause).

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 7.
Fig. 7.

As in Fig. 3, but for the 6-s sounding at Tallahassee at 1200 UTC 7 Aug 2005.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 8.
Fig. 8.

As in Fig. 5, but from 609 six-second soundings taken in 2005 at Tallahassee.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 9.
Fig. 9.

As in Fig. 6, but from 690 six-second soundings taken twice a day at Tallahassee in 2005.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 10.
Fig. 10.

As in Fig. 5, but from 717 six-second soundings taken in 2005 at Denver.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 11.
Fig. 11.

As in Fig. 6, but from 717 six-second soundings taken twice a day at Denver in 2005.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 12.
Fig. 12.

The ɛ using the (a), (c), (e) power and (b), (d), (f) spectral broadening methods for (a), (b) 2010–2057 UTC 24 Feb, (c), (d) 0222–0301 UTC 25 Feb, and (e), (f) 1719–1813 UTC 25 Feb (from Cohn 1995).

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Fig. 13.
Fig. 13.

Vertical profiles of the (a) Rig, (b) displacements, and (c) retrieved ɛ. Note the high correlation between the high displacements and the low Ri regions. The ɛ levels below the tropopause are consistent with radar measurements. No radar data are available in the ABL for comparison. Note the high stability of the atmospheric column from the top of the ABL to about 5 km.

Citation: Journal of Atmospheric and Oceanic Technology 25, 6; 10.1175/2007JTECHA992.1

Save