• Andra, D. L., , E. M. Quoetone, , and W. F. Bunting, 2002: Warning decision making: The relative roles of conceptual models, technology, strategy, and forecaster expertise on 3 May 1999. Wea. Forecasting, 17 , 559566.

    • Search Google Scholar
    • Export Citation
  • Atkins, N. T., , M. L. Weisman, , and L. J. Wicker, 1999: The influence of preexisting boundaries on supercell evolution. Mon. Wea. Rev., 127 , 29102927.

    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3 , 396409.

  • Barnes, S. L., 1978: Oklahoma thunderstorms on 29–30 April 1970. Part I: Morphology of a tornadic storm. Mon. Wea. Rev., 106 , 673684.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., , and S. G. Gaddy, 2001: Airborne pseudo-dual-Doppler analysis of a rear-inflow jet and deep convergence zone within a supercell. Mon. Wea. Rev., 129 , 22702289.

    • Search Google Scholar
    • Export Citation
  • Bolton, D., 1980: The computation of equivalent potential temperature. Mon. Wea. Rev., 108 , 10461053.

  • Brandes, E. A., 1981: Finestructure of the Del City-Edmond tornadic mesocirculation. Mon. Wea. Rev., 109 , 635647.

  • Brandes, E. A., 1984: Relationships between radar-derived thermodynamic variables and tornadogenesis. Mon. Wea. Rev., 112 , 10331052.

  • Brooks, H. E., , and C. A. Doswell III, 2002: Deaths in the 3 May 1999 Oklahoma City Tornado from a historical perspective. Wea. Forecasting, 17 , 354361.

    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., , C. A. Doswell III, , and R. Davies-Jones, 1993: Environmental helicity and the maintenance and evolution of low-level mesocyclones. The Tornado: Its Structure, Dynamics, Prediction and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 143–159.

  • Brooks, H. E., , C. A. Doswell III, , and J. Cooper, 1994a: On the environments of tornadic and nontornadic mesocyclones. Wea. Forecasting, 9 , 606618.

    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., , C. A. Doswell III, , and R. B. Wilhelmson, 1994b: The role of midtropospheric winds in the evolution and maintenance of low-level mesocyclones. Mon. Wea. Rev., 122 , 126136.

    • Search Google Scholar
    • Export Citation
  • Brown, S., , P. Archer, , E. Kruger, , and S. Mallonee, 2002: Tornado-related deaths and injuries in Oklahoma due to the 3 May 1999 tornadoes. Wea. Forecasting, 17 , 343353.

    • Search Google Scholar
    • Export Citation
  • Burgess, D. W., , R. A. Brown, , L. R. Lemon, , and C. R. Safford, 1977: Evolution of a tornadic thunderstorm. Preprints, 10th Conf. on Severe Local Storms, Omaha, NE, Amer. Meteor. Soc., 84–89.

  • Burgess, D. W., , M. A. Magsig, , J. Wurman, , D. C. Dowell, , and Y. Richardson, 2002: Radar observations of the 3 May 1999 Oklahoma City Tornado. Wea. Forecasting, 17 , 456471.

    • Search Google Scholar
    • Export Citation
  • Darkow, G. L., 1969: An analysis of over sixty tornado proximity soundings. Preprints, Sixth Conf. on Severe Local Storms, Chicago, IL, Amer. Meteor. Soc., 218–221.

  • Davies-Jones, R., 1984: Streamwise vorticity: The origin of updraft rotation in supercell storms. J. Atmos. Sci., 41 , 29913006.

  • Davies-Jones, R., , R. J. Trapp, , and H. B. Bluestein, 2001: Tornadoes and tornadic storms. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 167–222.

  • Doswell III, C. A., , D. V. Baker, , and C. A. Liles, 2002: Recognition of negative mesoscale factors for severe-weather potential: A case study. Wea. Forecasting, 17 , 937954.

    • Search Google Scholar
    • Export Citation
  • Dowell, D. C., , and H. B. Bluestein, 1997: The Arcadia, Oklahoma, storm of May 1981: Analysis of a supercell during tornadogenesis. Mon. Wea. Rev., 125 , 25622582.

    • Search Google Scholar
    • Export Citation
  • Edwards, R., , S. F. Corfidi, , R. L. Thompson, , J. S. Evans, , J. P. Craven, , J. P. Racy, , D. W. McCarthy, , and M. D. Vescio, 2002: Storm Prediction Center forecasting issues related to the 3 May 1999 tornado outbreak. Wea. Forecasting, 17 , 544558.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 1994: Atmospheric Convection. Oxford University Press, 580 pp.

  • Fujita, T. T., 1975: New evidence from the April 3–4, 1974 tornadoes. Preprints, Ninth Conf. on Severe Local Storms, Norman, OK, Amer. Meteor. Soc., 248–255.

  • Gilmore, M. S., , and L. J. Wicker, 2002: Influences of the local environment on supercell cloud-to-ground lightning, radar characteristics, and severe weather on 2 June 1995. Mon. Wea. Rev., 130 , 23492372.

    • Search Google Scholar
    • Export Citation
  • Grzych, M. L., , B. D. Lee, , C. A. Finley, , and J. L. Schroeder, 2004: Thermodynamic characterization of supercell rear-flank downdrafts in Project ANSWERS 2003. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, P11.1.

  • Hane, C. E., , and P. S. Ray, 1985: Pressure and buoyancy fields derived from Doppler radar data in a tornadic thunderstorm. J. Atmos. Sci., 42 , 1835.

    • Search Google Scholar
    • Export Citation
  • Klemp, J. B., 1987: Dynamics of tornadic thunderstorms. Annu. Rev. Fluid Mech., 19 , 369402.

  • Klemp, J. B., , and R. Rotunno, 1983: A study of the tornadic region within a supercell thunderstorm. J. Atmos. Sci., 40 , 359377.

  • Koch, S. E., , M. desJardins, , and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor., 22 , 14871503.

    • Search Google Scholar
    • Export Citation
  • Lee, B. D., , C. A. Finley, , and P. Skinner, 2004: Thermodynamic and kinematic analysis of multiple RFD surges for the 24 June 2003 Manchester, SD cyclic tornadic supercell during Project ANSWERS 2003. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, P11.2.

  • Lemon, L. R., , and C. A. Doswell III, 1979: Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Mon. Wea. Rev., 107 , 11841197.

    • Search Google Scholar
    • Export Citation
  • Ludlum, F. H., 1963: Severe local storms: A review. Severe Local Storms, Meteor. Monogr., No. 27, Amer. Meteor. Soc., 1–30.

  • Markowski, P. M., 2002a: Hook echoes and rear-flank downdrafts: A review. Mon. Wea. Rev., 130 , 852876.

  • Markowski, P. M., 2002b: Mobile mesonet observations on 3 May 1999. Wea. Forecasting, 17 , 430444.

  • Markowski, P. M., , and Y. P. Richardson, 2004: Multiple-Doppler radar observations of vertical wind profile heterogeneity in convective boundary layers. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, P13.1.

  • Markowski, P. M., , and J. Y. Harrington, 2005: A simulation of a supercell thunderstorm with emulated radiative cooling beneath the anvil. J. Atmos. Sci., 62 , 26072617.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , E. N. Rasmussen, , J. M. Straka, , and D. C. Dowell, 1998a: Observations of low-level baroclinity generated by anvil shadows. Mon. Wea. Rev., 126 , 29422958.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , E. N. Rasmussen, , and J. M. Straka, 1998b: The occurrence of tornadoes in supercells interacting with boundaries during VORTEX-95. Wea. Forecasting, 13 , 852859.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , J. M. Straka, , and E. N. Rasmussen, 2002: Direct surface thermodynamic measurements within the rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 16921721.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , J. M. Straka, , and E. N. Rasmussen, 2003: Tornadogenesis resulting from the transport of circulation by a downdraft: Idealized numerical simulations. J. Atmos. Sci., 60 , 795823.

    • Search Google Scholar
    • Export Citation
  • Marshall, T. P., 2002: Tornado damage survey at Moore, Oklahoma. Wea. Forecasting, 17 , 582598.

  • McCarthy, D. H., 2002: The role of ground-truth reports in the warning-decision making process during the 3 May 1999 Oklahoma tornado outbreak. Wea. Forecasting, 17 , 647649.

    • Search Google Scholar
    • Export Citation
  • Mielke, P. W., , K. J. Berry, , and G. W. Brier, 1981: Application of multi-response permutation procedures for examining seasonal changes in monthly mean sea-level pressure patterns. Mon. Wea. Rev., 109 , 120126.

    • Search Google Scholar
    • Export Citation
  • Pan, K., , P. Montpellier, , and M. Zadeh, 2002: Engineering observations of the 3 May 1999 Oklahoma tornado damage. Wea. Forecasting, 17 , 599610.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., 2003: Refined supercell and tornado forecast parameters. Wea. Forecasting, 18 , 530535.

  • Rasmussen, E. N., , and J. M. Straka, 1996: Mobile mesonet observations of tornadoes during VORTEX-95. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 1–5.

  • Rasmussen, E. N., , and D. O. Blanchard, 1998: A baseline climatology of sounding-derived supercell and tornado forecast parameters. Wea. Forecasting, 13 , 11481164.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., , J. M. Straka, , R. Davies-Jones, , C. A. Doswell III, , F. H. Carr, , M. D. Eilts, , and D. R. MacGorman, 1994: Verification of the origins of rotation in tornadoes experiment: VORTEX. Bull. Amer. Meteor. Soc., 75 , 9951006.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., , S. J. Richardson, , J. M. Straka, , P. M. Markowski, , and D. O. Blanchard, 2000: The association of significant tornadoes with a baroclinic boundary on 2 June 1995. Mon. Wea. Rev., 128 , 174191.

    • Search Google Scholar
    • Export Citation
  • Roebber, P. J., , D. M. Schultz, , and R. Romero, 2002: Synoptic regulation of the 3 May 1999 tornado outbreak. Wea. Forecasting, 17 , 399429.

    • Search Google Scholar
    • Export Citation
  • Rotunno, R., 1981: On the evolution of thunderstorm rotation. Mon. Wea. Rev., 109 , 577586.

  • Rotunno, R., , and J. B. Klemp, 1985: On the rotation and propagation of simulated supercell thunderstorms. J. Atmos. Sci., 42 , 271292.

    • Search Google Scholar
    • Export Citation
  • Speheger, D. A., , C. A. Doswell III, , and G. J. Stumpf, 2002: The tornadoes of 3 May 1999: Event verification in central Oklahoma and related issues. Wea. Forecasting, 17 , 362381.

    • Search Google Scholar
    • Export Citation
  • Srivastava, R. C., 1987: A model of intense downdrafts driven by the melting and evaporation of precipitation. J. Atmos. Sci., 44 , 17521774.

    • Search Google Scholar
    • Export Citation
  • Stensrud, D. J., , and S. J. Weiss, 2002: Mesoscale model ensemble forecasts of the 3 May 1999 tornado outbreak. Wea. Forecasting, 17 , 526543.

    • Search Google Scholar
    • Export Citation
  • Straka, J. M., , E. N. Rasmussen, , and S. E. Fredrickson, 1996: A mobile mesonet for fine-scale meteorological observations. J. Atmos. Oceanic Technol., 13 , 921936.

    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , and R. Edwards, 2000: An overview of environmental conditions and forecast implications of the 3 May 1999 tornado outbreak. Wea. Forecasting, 15 , 682699.

    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , R. Edwards, , J. A. Hart, , K. L. Elmore, , and P. Markowski, 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting, 18 , 12431261.

    • Search Google Scholar
    • Export Citation
  • Trapp, R. J., 1999: Observations of nontornadic low-level mesocyclones and attendant tornado-genesis failure during VORTEX. Mon. Wea. Rev., 127 , 16931705.

    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., 2001: Convectively driven high wind events. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 255–298.

  • Wakimoto, R. M., , and H. Cai, 2000: Analysis of a nontornadic storm during VORTEX 95. Mon. Wea. Rev., 128 , 565592.

  • Wakimoto, R. M., , H. V. Murphey, , D. C. Dowell, , and H. B. Bluestein, 2003: The Kellerville Tornado during VORTEX: Damage survey and Doppler radar analyses. Mon. Wea. Rev., 131 , 21972221.

    • Search Google Scholar
    • Export Citation
  • Weisman, M. L., , and J. B. Klemp, 1982: The dependence of numerically simulated convective storms on vertical wind shear and buoyancy. Mon. Wea. Rev., 110 , 504520.

    • Search Google Scholar
    • Export Citation
  • Wicker, L. J., 1996: The role of near surface wind shear on low-level mesocyclone generation and tornadoes. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 115–119.

  • Wilhelmson, R. B., , and L. J. Wicker, 2001: Numerical modeling of severe local storms. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 123–166.

  • Wurman, J., 2002: The multiple vortex structure of a tornado. Wea. Forecasting, 17 , 473505.

  • Yuan, M., , M. Dickens-Micozzi, , and M. A. Magsig, 2002: Analysis of tornado damage tracks from the 3 May tornado outbreak using multispectral satellite imagery. Wea. Forecasting, 17 , 382398.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (left) Schematic representation of a supercell thunderstorm, adapted from the conceptual model presented by Lemon and Doswell (1979). The dark gray shading approximates the precipitation region, at least as viewed by radar (e.g., radar reflectivity factor >30 dBZ). Within this region, the locations of the FFD and RFD are indicated. The light gray shading approximates the main updraft region (U). The typical location of a tornado, if one occurs, also is indicated (T). The outflow boundary is indicated by the barbed contour. A few storm-relative streamlines are drawn. (right) Photograph of a tornadic supercell that closely resembles the schematic shown on the left. The approximate vantage point of the photograph is indicated by the star in the schematic, roughly east of the main updraft. The labels in the photograph correspond to the same features depicted in the schematic. (Photograph courtesy of E. Rasmussen)

  • View in gallery

    Three-dimensional schematic of a numerically simulated supercell thunderstorm in westerly mean shear, viewed from the southeast, at a stage when low-level rotation is intensifying. The cylindrical arrows depict the storm-relative winds. The thin lines are vortex lines, with the sense of rotation indicated by the circular arrows. The heavy barbed line marks the outflow boundary. The orientation of the horizontal buoyancy gradient, hB, also is indicated. [Adapted from Klemp (1987)]

  • View in gallery

    Retrieved virtual potential temperature perturbations (θυ; K) at 1.3 km AGL in the (left) Del City, OK, supercell (0045 UTC 21 May 1977), and in the (right) Harrah, OK, supercell (2143 UTC 8 Jun 1974) studied by Brandes (1984). The gray region indicates where radar reflectivity factor exceeds 30 dBZ (40 dBZ) in the 21 May 1977 (8 Jun 1974) case. Positive (negative) perturbations are contoured using thin solid (dashed) lines (note that the contour interval is 2 K on 21 May 1977 and 1 K on 8 Jun 1974). The heavy dashed lines enclose regions where vertical vorticity exceeds 0.01 s−1. [Adapted from Brandes (1984). The figure actually displays fields of “quasi-virtual” potential temperature perturbations, which Brandes (1984) defines as B = θ′ + 0.61qυ, where θ′, , and qυ are the potential temperature perturbation, base-state potential temperature, and specific humidity perturbation, respectively. According to Brandes (1984), B values typically are within 0.1 K of θυ values]

  • View in gallery

    Schematic diagram illustrating how the FFD outflow region was objectively defined. The radar echo is shaded dark gray, the updraft is shaded light gray, and the FFD outflow region is defined by the hatching. The FFD outflow region was defined as the region of low-level radar reflectivity on the forward side of a line drawn orthogonal to the major axis of the echo, through the radar-observed mesocyclone center (M).

  • View in gallery

    The relationship between radar reflectivity factor (dBZ) and parameterized condensate, qc, and the difference between the virtual and density potential temperature, θυθρ, assuming a potential temperature, θ, of 300 K.

  • View in gallery

    Density potential temperature fields for all 12 cases analyzed. Unshaded areas indicate that observations were unavailable. The heavy solid line encloses the 40-dBZ radar echo. The filled circles indicate the locations of the mesocyclone centers at an altitude of approximately 1–2 km AGL.

  • View in gallery

    Subjective analyses at 2349 UTC 22 May 1995 of (a) virtual potential temperature perturbations (θυ; 1-K contour interval), (b) density potential temperature perturbations (θρ; 1-K contour interval), (c) equivalent potential temperature perturbations (θe; 2-K contour interval), and (d) the height on an inflow sounding where the surface θe value is equal to that observed at the surface (zo; 0.4-km contour interval). The reference values are υ = ρ = 310.9 K and e = 353.0 K. The analyses have been overlaid on objectively analyzed radar reflectivity data at 0.5 km AGL from the NOAA P3 lower fuselage radar (see legend). The filled circle indicates the location of the mesocyclone center at an altitude of approximately 1–2 km AGL. Station models display storm-relative winds (a full barb equals 5 m s−1; a half-barb equals 2.5 m s−1) and the observed values of the variables analyzed, except in (d), in which station models display θe values as in (c). To avoid clutter, only a small number of station models are plotted (data were recorded at a frequency of 2 Hz, which corresponds to an along-track horizontal data spacing of 40 m given a typical vehicle speed of 20 m s−1).

  • View in gallery

    Subjective analyses of density potential temperature perturbations (θρ; 1-K contour interval) at 2338 and 2359 UTC 22 May 1995 (approximately 10 min before and after the analysis time in Fig. 7) (ρ = 310.9 K). All symbols are identical to those appearing in Fig. 7.

  • View in gallery

    Same as in Fig. 7, but for 0116 UTC 21 May 1998. The reference values are υ = ρ = 312.0 K and e = 342.1 K. The zo field is contoured every 0.2 km. The radar reflectivity data are from the lowest elevation scan of the KGLD WSR-88D (the data are at approximately 0.8 km AGL).

  • View in gallery

    Same as in Fig. 7, but for 0058 UTC 3 Jun 1995. The reference values are υ = ρ = 308.9 K and e = 352.9 K. The zo field is contoured every 0.2 km. The radar reflectivity data are at 0.5 km AGL and are from the NOAA P3 lower fuselage radar.

  • View in gallery

    Subjective analyses of density potential temperature perturbations (θρ; 1-K contour interval) at 0048 and 0108 UTC 3 Jun 1995 (10 min before and after the analysis time in Fig. 10) (ρ = 308.9 K). All symbols are identical to those appearing in Fig. 10.

  • View in gallery

    Same as in Fig. 7, but for 2356 UTC 3 May 1999. The field of zo is not displayed because a suitable proximity sounding was not available. The reference values are υ = ρ = 304.8 K and e = 352.0 K. The radar reflectivity data are from the lowest elevation scan of the KTLX WSR-88D (the data are at approximately 0.6 km AGL).

  • View in gallery

    Scatter diagram of minimum density potential temperature perturbation, θρmin, within the FFD outflow sampled vs surface dewpoint depression, Tddsfc, in the near-storm inflow. Units are K. A best-fit line also has been added. The linear correlation coefficient, r, appears at the top right.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 15 15 11
PDF Downloads 15 15 12

Surface In Situ Observations within the Outflow of Forward-Flank Downdrafts of Supercell Thunderstorms

View More View Less
  • 1 Department of Meteorology, The Pennsylvania State University, University Park, Pennsylvania
© Get Permissions
Full access

Abstract

In the long-standing conceptual model of a supercell thunderstorm, the forward-flank downdraft (FFD) and its associated negative buoyancy originate from precipitation loading and the latent chilling of air due to the melting and evaporation of precipitation. The horizontal buoyancy gradient within the outflow of the FFD has been identified as an important source of low-level, streamwise vorticity in three-dimensional numerical simulations of supercells. These simulations have demonstrated that the formation of low-level mesocyclones is critically dependent on the baroclinic generation of horizontal vorticity within the FFD outflow.

Despite the implied dynamical importance of the FFD outflow in the evolution of supercell thunderstorms, only a very limited number of thermodynamic observations have been obtained within FFD outflow. The range of thermodynamic conditions within FFD outflow is not well known, nor is it known whether any systematic relationship exists between the thermodynamic characteristics of FFD outflow and the intensity of low-level mesocyclones and/or tornadogenesis. In this paper, in situ observations obtained at the ground by a mobile mesonet within FFD outflow are used to investigate whether any relationship exists between the thermodynamic characteristics of the outflow and low-level mesocyclogenesis and/or tornadogenesis. The data were obtained within both tornadic and nontornadic supercells (12 cases total) during the Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX) from 1994 to 1995, and in smaller field campaigns during the 1997–99 period.

* Current affiliation: AccuWeather, Inc., State College, Pennsylvania

Corresponding author address: Dr. Paul Markowski, The Pennsylvania State University, 503 Walker Building, University Park, PA 16802. Email: pmarkowski@psu.edu

Abstract

In the long-standing conceptual model of a supercell thunderstorm, the forward-flank downdraft (FFD) and its associated negative buoyancy originate from precipitation loading and the latent chilling of air due to the melting and evaporation of precipitation. The horizontal buoyancy gradient within the outflow of the FFD has been identified as an important source of low-level, streamwise vorticity in three-dimensional numerical simulations of supercells. These simulations have demonstrated that the formation of low-level mesocyclones is critically dependent on the baroclinic generation of horizontal vorticity within the FFD outflow.

Despite the implied dynamical importance of the FFD outflow in the evolution of supercell thunderstorms, only a very limited number of thermodynamic observations have been obtained within FFD outflow. The range of thermodynamic conditions within FFD outflow is not well known, nor is it known whether any systematic relationship exists between the thermodynamic characteristics of FFD outflow and the intensity of low-level mesocyclones and/or tornadogenesis. In this paper, in situ observations obtained at the ground by a mobile mesonet within FFD outflow are used to investigate whether any relationship exists between the thermodynamic characteristics of the outflow and low-level mesocyclogenesis and/or tornadogenesis. The data were obtained within both tornadic and nontornadic supercells (12 cases total) during the Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX) from 1994 to 1995, and in smaller field campaigns during the 1997–99 period.

* Current affiliation: AccuWeather, Inc., State College, Pennsylvania

Corresponding author address: Dr. Paul Markowski, The Pennsylvania State University, 503 Walker Building, University Park, PA 16802. Email: pmarkowski@psu.edu

1. Introduction

Supercell thunderstorms are well known for the variety of threats they pose to life and property, such as large hail, damaging straightline winds, and tornadoes. Although not all supercells are tornadic, nearly all large, violent tornadoes are associated with supercells. This is certainly one of the reasons why supercells have been a focal point of decades of severe storms observational, theoretical, and numerical modeling research [recent reviews have been written by Davies-Jones et al. (2001) and Wilhelmson and Wicker (2001)].

For more than two decades, the supercell conceptual model presented by Lemon and Doswell (1979) has remained relatively unaltered in terms of its main components, which include one main updraft region and two downdraft regions—one on the rear flank of the storm and another on the forward flank (Fig. 1). Many past studies [see Markowski (2002a) for a review] have directed considerable attention to the rear-flank downdraft (RFD), the development or intensification of which has been closely associated with tornadogenesis on many occasions (e.g., Ludlam 1963; Fujita 1975; Burgess et al. 1977; Barnes 1978; Brandes 1981). Of interest in this paper is the forward-flank downdraft (FFD), which forms downwind (with respect to the midlevel, storm-relative flow) of the updraft as condensate produced in the updraft is advected downstream (Lemon and Doswell 1979). As the precipitation falls into subsaturated air, latent chilling gives rise to negative buoyancy, leading to the formation of the FFD. Precipitation loading also contributes to the negative buoyancy within the FFD. The negatively buoyant air descends to the surface and a gust front separates the relatively cool FFD outflow1 from the warm environmental inflow.

The potential dynamical importance of the FFD and its outflow has been revealed by three-dimensional numerical simulations (Klemp and Rotunno 1983; Rotunno and Klemp 1985). These simulations have shown that baroclinic horizontal vorticity, generated by the horizontal buoyancy gradient along the forward-flank gust front, can be tilted into the vertical and stretched, just as environmental horizontal vorticity associated with the mean vertical wind shear is tilted and stretched (Fig. 2). Furthermore, the baroclinic vorticity tends to be streamwise because storm-relative streamlines approaching the updraft from the forward flank are generally normal to the horizontal buoyancy gradient. Whereas the tilting of environmental horizontal vorticity has been shown to be fundamental to the formation of midlevel mesocyclones in supercells (Barnes 1978; Rotunno 1981; Davies-Jones 1984), the tilting of baroclinic vorticity originating within the FFD and its outflow has been implicated in the formation of low-level mesocyclones (Klemp and Rotunno 1983; Rotunno and Klemp 1985), where “low level” has typically referred to approximately the lowest 1 km above ground level (AGL).

Simulations have demonstrated that the magnitude of the baroclinic horizontal vorticity can be comparable to or exceed the magnitude of the environmental horizontal vorticity. This result is not surprising given the scale analysis presented by Klemp and Rotunno (1983), whereby the change in horizontal vorticity following an air parcel moving parallel to the isentropes (a proxy for buoyancy isopleths), Δωs, is
i1520-0493-134-5-1422-e1
where s and n represent directions parallel and perpendicular to the flow, respectively; ωs is the streamwise vorticity component; g is the gravitational acceleration; θ is the potential temperature; and υs is the storm-relative wind speed. When a normal potential temperature gradient of ∂θ/∂n = 1°C km−1 and inflow having υs = 15 m s−1 is assumed (Klemp and Rotunno 1983), a streamwise horizontal vorticity of Δωs = 10−2 s−1 is generated over a distance of Δs = 5 km.

Although baroclinic vorticity generated within the FFD outflow generally might be expected to augment environmental horizontal vorticity, some numerical simulations (Wicker 1996) have suggested that the orientation of the baroclinic vorticity relative to the environmental vorticity might influence the intensity and longevity of low-level mesocyclones. The spatial distribution of precipitation, controlled in large part by the storm-relative winds, affects the spatial distribution of negative buoyancy and baroclinic vorticity generation, and therefore also likely exerts a significant influence on the characteristics of low-level mesocyclones (Brooks et al. 1993, 1994b). An additional complication is that the relationship between low-level mesocyclones and tornadoes seems rather tenuous. The presence of a low-level mesocyclone is not a sufficient condition for tornadogenesis, nor does the likelihood of tornadogenesis (or the intensity of an attendant tornado) necessarily increase with increasing low-level mesocyclone intensity or longevity (e.g., Trapp 1999; Wakimoto and Cai 2000; Wakimoto et al. 2003).

Thermodynamic observations within FFDs and their associated outflow have been relatively rare. Indirect observations obtained through buoyancy retrievals based on dual-Doppler wind syntheses (Brandes 1984; Fig. 3) might lead some to question whether the dynamical significance of the FFD is the same in all supercells. Brandes (1984) found that streamwise horizontal vorticity was produced baroclinically in the FFD outflow of the extensively studied tornadic supercell that struck Del City, Oklahoma, on 20 May 1977, just as a numerical simulation of the same storm revealed (Klemp and Rotunno 1983). Buoyancy retrievals by Hane and Ray (1985) supported Brandes’s findings for the Del City storm, as their buoyancy retrievals (using a slightly different approach) also indicated a favorable orientation of buoyancy gradients. However, in the Harrah, Oklahoma, storm of 8 June 1974, Brandes found that the buoyancy gradient was reversed along inflow trajectories, resulting in an unfavorable (anti-streamwise) orientation of baroclinic horizontal vorticity. It is not clear whether the conflicting findings of Brandes (1984) regarding the importance of baroclinic vorticity generation in FFD outflow are indicative of real differences among supercell thunderstorms, or whether his conflicting results are indicative of the limitations of buoyancy retrievals. Buoyancy retrievals are sensitive to both temporal and spatial derivatives of the horizontal and vertical velocity field, the latter of which is itself derived from spatial derivatives of the horizontal velocity field. The accuracy of the temporal derivative calculations is perhaps the most questionable aspect of dual-Doppler-based buoyancy retrievals, given the times to complete a volume scan (typically 4–10 min), during which time significant evolution can occur. Furthermore, buoyancy retrievals are usually unreliable in the lowest few 100 m AGL because of radar horizon issues.

In situ (direct) surface thermodynamic observations within FFDs and FFD outflow have been even harder to come by. The relatively rare passages of supercells over fixed observing systems combined with the possible hazards of acquiring observations using mobile observation systems have made it difficult to gather dense observations. Only a handful of direct measurements have been obtained within FFDs and FFD outflow, most of which are limited in spatial coverage. For example, the Arcadia, Oklahoma, storm on 17 May 1981 passed over a 444-m-tall instrumented tower (Dowell and Bluestein 1997). A gradual 5°C temperature drop was observed at the surface over a 1-h period, suggesting that the temperature gradient within the FFD outflow was spread over a broad region. However, the tower data did not permit an assessment of the spatial variability of buoyancy within the outflow. The capabilities of a mobile mesonet (e.g., Straka et al. 1996) observing system have been demonstrated in several recent studies (e.g., Rasmussen and Straka 1996; Markowski et al. 2002; Grzych et al. 2004; Lee et al. 2004). Markowski et al. (2002) used mobile mesonet observations to document the surface thermodynamic fields in close proximity of tornadic and nontornadic low-level mesocyclones to determine if differences existed at the surface between tornadic and nontornadic supercells. They concluded that tornado likelihood, intensity, and longevity increase as the surface buoyancy and equivalent potential temperature within the hook echo and RFD outflow increase.

Regardless of whether or not the FFD outflow and associated buoyancy gradient contributes significantly to low-level mesocyclones, at least some baroclinity is likely present in all supercells, given the virtually ubiquitous presence of precipitation on the forward flank. The goal of this paper is to further document surface conditions within the outflow of FFDs, extending our understanding of the near-ground thermodynamic differences between tornadic and nontornadic supercells. This work has been motivated in part by a recent review by Wakimoto (2001), who states: “detailed observations of the forward-flank downdraft and gust front are still lacking. These analyses would be useful complements to the high-resolution numerical simulations presented in the literature.” It is believed that past observations have barely scratched the surface with regards to capturing the range of thermodynamic properties within FFDs, their relationship to storm evolution, and their association with the large-scale environment.

This research relies on surface data collected by a mobile mesonet within supercell thunderstorms during the 1994–99 period. A description of data and analysis methods appears in section 2. This is followed by a summary of the characteristics of FFD outflow in nontornadic and tornadic supercells in section 3, which are related to characteristics of the larger-scale environment in section 4. Section 5 discusses possible implications of the findings. Section 6 presents the conclusions and some closing remarks.

2. Data and analysis methods

a. Overview of the available data

Data from an automobile-borne observing system—a “mobile mesonet”—were analyzed for 12 cases (7 tornadic, 5 nontornadic) occurring on 10 different days during the period spanning 1994–99 (Table 1). In the tornadic cases, the tornado damage ratings ranged from F0 to F4 on the Fujita scale and tornado durations ranged from 1 to 46 min. The cases acquired in 1994 and 1995 (9 cases) occurred during the Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX; Rasmussen et al. 1994).

The mobile mesonet employed herein was developed by Straka et al. (1996) for use during VORTEX and several smaller, post-VORTEX field experiments during the 1997–99 period. Time, latitude, longitude, temperature, relative humidity, pressure, and wind velocity were recorded at 2-s intervals by anywhere from 4 to 16 vehicles. Instrument specifications, quality control techniques, and error analyses are described by Straka et al. (1996) and Markowski et al. (2002). High-frequency noise in the raw data was suppressed using a triangular weighting function with a filter radius of 10 s. The spatial scales retained by such filtering depend on vehicle speed, but for typical vehicle speeds during data collection, this filtering significantly damped features having wavelengths less than 200 m.

Because the shape and size of precipitation fields associated with supercells vary by storm, there was a need to objectively define the FFD outflow region of a supercell. The FFD outflow was defined as the region of low-level radar reflectivity on the forward side of a line drawn orthogonal to the major axis of the echo, through the radar-observed circulation center (Fig. 4). Although no attempt to objectively define the FFD outflow is free of limitations, the consistent application of a single definition from case to case increases one’s ability to make meaningful comparisons of FFD outflow.

The greatest challenge in observing storms using in situ sensors such as these is that the spatial coverage of the data collection is limited by road networks and a finite number of vehicles. To maximize the areal coverage of observations, storms were assumed to be approximately steady for 6-min periods [i.e., approximately the length of time it takes for the Weather Surveillance Radar-1988 Doppler (WSR-88D) to complete a volume scan], during which time time-to-space-converted data contributed to an analysis (Markowski et al. 2002). Storm translation velocities were obtained by maximizing the correlation of phase-shifted radar reflectivity fields obtained from different times. In each case, all times at which data were available were examined in order to determine the periods having the best spatial coverage of observations within the FFD outflow. At the times when the sampling was deemed to be best, the quality of the sampling was quantified (Table 2). In 9 of the 12 cases, observations were obtained within 1 km of the forward-flank radar reflectivity maximum, based on objectively analyzed data from the nearest National Weather Service radar (WSR-88D), or, if available, the lower fuselage radar aboard the P3 aircraft operated by the National Oceanic and Atmospheric Administration (NOAA), which was commissioned for VORTEX. One might expect the buoyancy minimum within the FFD outflow to be located near the radar reflectivity maximum.

Although it might be preferable to analyze each case at common times with respect to a storm’s evolutionary stage (e.g., time of maximum low-level rotation, hook echo formation, tornadogenesis, etc.), this goal conflicted with the goal of maximizing observation coverage. Higher priority was put on the latter goal; therefore, not all analyses are at the same time relative to a storm’s evolutionary stage. This might introduce some difficulty in comparing observations from case to case, but even greater difficulty would be had if analyses that were constrained to be at identical evolution-relative times had even larger gaps in data coverage. For two of the cases described in section 3, a sequence of analyses over a 20-min time period is presented. Although the data coverage at the auxiliary analysis times is not as extensive, there is some suggestion that the thermodynamic fields in the FFD outflow do not change drastically on 20-min time scales. This gives some support for the approach we have undertaken.

b. Variables analyzed

Fields of virtual potential temperature, density potential temperature, equivalent potential temperature, and parcel origins (by examining sounding data and using moist entropy as a tracer) were subjectively analyzed from the mobile mesonet data. The virtual potential temperature, θυ, is defined as
i1520-0493-134-5-1422-e2
where θ is the potential temperature and qυ is the water vapor mixing ratio. The contribution of condensate to negative buoyancy is included in density potential temperature, which is defined as (Emanuel 1994)
i1520-0493-134-5-1422-e3
where qc is the condensate mixing ratio. The condensate mixing ratio was parameterized in terms of the objectively analyzed radar reflectivity following Hane and Ray (1985) (Fig. 5), whereby
i1520-0493-134-5-1422-e4
where a = 0.1× the radar reflectivity factor in dBZ and ρ is the density of air. The radar data were obtained from either the lower fuselage radar of the NOAA P3 aircraft or a nearby WSR-88D. The reflectivity data were objectively analyzed using a single pass of an exponential weight function (Barnes 1964), whereby the weight of the kth datum that contributes to the reflectivity value at the ith grid point, wik, is
i1520-0493-134-5-1422-e5
where rik is the distance between the kth datum and ith grid point and κ is a smoothing parameter that controls the amount of detail retained in the gridded reflectivity field. Based on ranges of the radars from the storms, which varied from case to case, the range of κ values that one might find appropriate (e.g., Koch et al. 1983) is approximately 0.8–6.5 km2. Instead of using a different κ for each case, depending on the range from the radar to the storm, κ was prescribed to be 1.2 km2 for all cases. By fixing κ, the reflectivity fields from case to case had similar detail so that calculations of θρ did not have added uncertainty owing to differences in the spatial scales represented in the objectively analyzed reflectivity fields. In other words, by fixing κ, there was some risk of under- or oversmoothing the radar reflectivity data in some cases; however, higher priority was given to the consistent resolution of reflectivity scales from case to case so that the θρ calculations would not suffer from these differences.

Both θυ and θρ were calculated, as opposed to just one or the other, in order to best capture the magnitude of the true buoyancy and its gradients. In the absence of condensate, buoyancy is proportional to θυ perturbations, whereas in the presence of condensate, buoyancy is proportional to θρ perturbations. The disadvantage of analyzing θυ fields is that they underestimate the negative buoyancy in precipitation regions. But the analyzed θρ fields are not without their own limitations. For example, the condensate mixing ratio at the surface (where the mobile mesonet observations of other state variables are valid) has been derived from radar data that do not extend to the surface. Furthermore, the condensate estimated from (4) may be in error when hail is present, due to radar reflectivity being enhanced or attenuated. It was hoped that the most accurate portrayal of the buoyancy fields could be obtained by analyzing both the θυ and θρ fields. The difference between the θυ and θρ fields is just θqc, which is displayed in Fig. 5 as a function of radar reflectivity.

Analyses of equivalent potential temperature, θe, used Bolton’s (1980) approximation.2 If θe is approximately conserved for adiabatic processes, then one can also estimate from a nearby sounding the height from which surface air in the FFD outflow has descended. The origin of air parcels (zo), assuming θe conservation during descent, was analyzed for each case; however, zo values should be viewed with caution, given the uncertainties of the degree of entrainment with downdraft parcels en route to the surface. As Markowski et al. (2002) cautioned, it might be most appropriate to refer to zo as simply the height on an inflow sounding where the θe value is equal to that observed at the surface within the downdraft, rather than as a measure of parcel origin.

c. Specification of the reference state and buoyancy gradient calculations

To make meaningful comparisons of FFD outflow from one case to another, it was useful to analyze the deviations of variables from some reference state, whereby ξ = ξ + ξ′, where ξ is a thermodynamic variable, ξ is the reference value of this variable, and ξ′ is the deviation of the thermodynamic variable from the reference state. The reference state was defined by the conditions sampled by the mobile mesonet immediately outside of the precipitation region (assumed to be where the radar reflectivity factor exceeded 25 dBZ), rather than with respect to some large-scale spatial average (e.g., Markowski et al. 2002). This approach was taken so that virtual and density potential temperature deficits would be proportional to the baroclinity within the FFD outflow. All reference state specifications are arbitrary. The reference state could also have been defined by some larger-scale spatial average, but such an approach often results in temperature deficits outside of the precipitation regions due to processes unrelated to the FFD (e.g., cloud shading; Markowski et al. 1998a; Markowski and Harrington 2005).

Given the prior emphasis on the baroclinic generation of horizontal vorticity within the FFD outflow by horizontal buoyancy gradients, there was a strong motivation to assess the buoyancy gradients sampled by the mobile mesonet. In general, gradient calculations tend to be volatile and highly dependent on observation density. In some of the cases (e.g., 29 April and 22 May 1995), the spatial coverage and density of the observations allowed for direct calculations of the gradients that were deemed reasonably trustworthy. In other cases sampled less adequately (e.g., 6 May 1994 and 20 May 1998), the horizontal buoyancy gradient was estimated in a bulk sense, based on the maximum density potential temperature deficit and its distance from the edge of the echo.

The magnitude of the streamwise vorticity that could be generated baroclinically by the buoyancy gradients also was estimated using (1), but by replacing θ in (1) with θυ or θρ, and also taking the angle between the storm-relative wind and buoyancy gradient into consideration (it was found that only rarely is the storm-relative wind exactly normal to the buoyancy gradient). The streamwise vorticity generation was estimated only over a 5-km distance, just as in the scale analysis in section 1. For typical storm-relative wind speeds of 10–15 m s−1, this implies an integration of the solenoidal torque on a parcel over a period of 5–8 min. We were simply unable to assess the total baroclinic vorticity generation with any confidence given that this requires knowledge of much longer parcel histories, which requires knowledge of parcel trajectories and the time evolution of the buoyancy gradients. Furthermore, it is not known just how significant the contribution of near-ground baroclinic vorticity generation is to the total vorticity budget of a low-level mesocyclone (baroclinic vorticity generation above the surface could not be assessed).

3. FFD outflow observations

The surface θρ fields for all 12 cases are presented in Fig. 6 in order to provide the reader with an overall sense of the data coverage and range of buoyancy characteristics within the FFD outflow. More detailed analyses of θυ, θρ, θe, and zo are presented for two nontornadic cases (22 May 1995 and 20 May 1998; Figs. 7 and 9, respectively) and two tornadic cases (2 June 1995 and 3 May 1999; Figs. 10 and 12, respectively), in order to give the reader a better feel for the nature of the observations upon which the conclusions will be based. Statistics characterizing the thermodynamic properties of the FFD outflow appear in Table 3, and the mean properties of nontornadic and tornadic supercells are compared in Table 4.

a. Nontornadic supercells

The minimum values of θυ, θρ, and θe (θυmin, θρmin, and θemin, respectively) sampled within the FFD outflow of nontornadic supercells ranged from −3.9 to −7.6, −5.0 to −12.4, and −7.0 to −13.9 K, respectively (Table 3). Values of the maximum altitude of the “origin” of FFD air reaching the surface, zomax ranged from 0.8–2.7 km. The magnitude of the baroclinity due to the horizontal buoyancy gradient was computed using both θυ and θρ for the reasons described in section 2. The maximum horizontal gradient of θυ (θρ), |hθυ|max(|hθρ|max), ranged from 0.4–1.0 K km−1 (0.6–1.8 K km−1). The maximum streamwise vorticity that was estimated to have been baroclinically generated over a 5-km distance based on the θυ(θρ) gradient and observed storm-relative winds, Δωθυsmaxωθρsmax), ranged from 0.004 to 0.015 s−1 (0.005–0.020 s−1).

1) The 22 May 1995 storm

The nontornadic supercell on 22 May 1995 was sampled during VORTEX, and various aspects of the storm have been previously documented by Markowski et al. (1998a), Trapp (1999), and Bluestein and Gaddy (2001). The storm was one of several storms initiated along a dryline in the eastern Texas panhandle. Radar observations revealed a strong midlevel mesocyclone within the storm, consistent with the visually stunning, striated appearance of the updraft (Bluestein and Gaddy 2001). The updraft had extremely large vertical velocities. Although the Bluestein and Gaddy (2001) calculation of >40 m s−1 at 4 km AGL is likely overdone (as they cautioned themselves), the widespread observations of hail having diameters of 8–10 cm suggest peak vertical velocities of >50 m s−1.

Despite the many indicators of the intensity of this storm, the storm possessed only weak low-level (below 1 km AGL) rotation throughout its lifetime. Of the six VORTEX supercells investigated by Trapp (1999), the 22 May 1995 storm was associated with the weakest low-level rotation and low-level vorticity stretching. The storm never developed many of the precursor cloud signatures (per VORTEX observers) that typify tornadic supercells (e.g., a lowered cloud base or wall cloud, and rapid rotation and/or vertical motion within the cloud base).

Mobile mesonet sampling of the storm was best near the time that the storm possessed its strongest rotation (Table 2). At 2349 UTC, 9 min before the time of “tornadogenesis failure” reported by Trapp (1999), the mobile mesonet observed large θυ and θρ deficits (maximum deficits were >7 and >10 K, respectively) north and northeast of the mesocyclone (Figs. 7a,b). The large buoyancy deficit was associated with strong baroclinity, with a |hθυ|max(|hθρ|max) of 0.7 K km−1 (1.4 K km−1). Furthermore, estimates of the streamwise vorticity that could be acquired by inflow parcels passing through this baroclinity were among the largest of any of the supercells sampled in this study (Δωθυsmax and Δωθρsmax were 0.008 and 0.017 s−1, respectively). The FFD outflow also contained large (>13 K) θe deficits at the surface (Fig. 7c). This lowest θe air had the same θe as air that was at 2.7 km AGL on an M-CLASS sounding launched less than 30 km southeast of the storm by a VORTEX mobile sounding unit (Fig. 7d).

Additional analyses of θρ are presented at 2338 and 2359 UTC, the approximate time of most intense low-level rotation (Trapp 1999), in an attempt to provide a sense of how the near-ground thermodynamic fields varied in time within the FFD outflow (Fig. 8). Despite the fact that the data coverage 2338 and 2359 UTC was not as extensive as at 2349 UTC, the subjective analyses of the θρ fields suggest that the maximum density excess was relatively constant throughout the 2338–2359 UTC period. Furthermore, large forward-flank baroclinity and streamwise vorticity generation are present throughout the period.

2) The 20 May 1998 storm

On 20 May 1998 conditions were conducive for supercell development over a large portion of the High Plains due to moist, easterly upslope flow at low levels and relatively strong (>20 m s−1) westerly flow in the mid- to upper troposphere. A small armada of five instrumented vehicles sampled a nontornadic supercell near Yuma, Colorado, that developed during the early evening. The spatial coverage of observations within the FFD outflow was greatest at 0116 UTC (Table 2), which was 30 min prior to the time of maximum low-level rotation detected by the WSR-88D in Goodland, Kansas.

As was the case for the 22 May 1995 storm, the mobile mesonet observed relatively large θυ and θρ deficits (maximum deficits were >6 and >12 K, respectively) within the FFD outflow, especially north of the mesocyclone, where radar reflectivity was largest (Figs. 9a,b). Not surprisingly, the FFD outflow also was associated with large baroclinity and streamwise vorticity generation estimates (e.g., |hθρ|max ∼ 1.8 K km−1 and Δωθρsmax ∼ 0.016 s−1). Deficits of θe (up to 8 K) were not as large as in the 22 May 1995 case (Fig. 9c). The lowest θe air detected at the surface within the FFD outflow had the same θe as air at 0.8 km AGL on the 0000 UTC North Platte, Nebraska, sounding (Fig. 9d).

b. Tornadic supercells

The values of θυmin, θρmin, and θemin sampled within the FFD outflow of tornadic supercells ranged from −2.5 to −5.3, −3.0 to −6.0, and −2.1 to −12.5 K, respectively (Table 3). Values of zomax ranged from 0.5 to 1.9 km. Values of |hθυ|max(|hθρ|max) ranged from 0.3 to 0.7 K km−1 (0.4–1.0 K km−1). Estimates of Δωθυsmaxωθρsmax) ranged from 0.002 to 0.007 s−1 (0.003–0.009 s−1).

1) The 2 June 1995 storm

The supercell that produced a tornado near Dimmitt, Texas, on 2 June 1995 was perhaps the best-sampled tornadic storm during VORTEX (this is the “2 June 1995b” case in Table 2). The supercell developed near the Texas–New Mexico border and became tornadic shortly after crossing an outflow boundary that had been produced by thunderstorms several hours earlier. The mesoscale aspects of this study have been documented by Rasmussen et al. (2000) and Gilmore and Wicker (2002).

The best mobile mesonet sampling of the FFD outflow was near the time of tornadogenesis (Table 2). At 0058 UTC 3 June, 1 min after tornadogenesis, the mobile mesonet observed only modest negative buoyancy within the FFD outflow (θυmin and θρmin were −2.5 and −3.1 K, respectively; Figs. 10a,b). The magnitude of the baroclinity was correspondingly smaller than in the 22 May 1995 and 20 May 1998 storms, as was the estimated streamwise vorticity generation (e.g., |hθρ|max ∼ 0.4 K km−1 and Δωθρsmax ∼ 0.003 s−1). The fact that the storm-relative wind direction was closely aligned with the buoyancy gradient within much of the FFD outflow also contributed to the relatively small estimates of streamwise vorticity generation. The θe deficits within the FFD outflow (2–5 K) were not particularly large either (Fig. 10c). Comparison of the θe values observed at the surface within the FFD outflow with those on an M-CLASS sounding launched within 10 km of the rear of the storm suggested that the air reaching the surface within the FFD outflow originated in the lowest 500 m AGL (Fig. 10d).

Analyses of θρ are also presented at an earlier (0048 UTC) and later (0108 UTC) analysis time, as was done for the 22 May 1995 nontornadic case, in order to give the reader a sense of the evolution of the near-ground thermodynamic fields within the FFD outflow (Fig. 11). There is some hint that the forward-flank baroclinity may be weaker at 0048 UTC than at 0058 UTC. At 0108 UTC, at which time a large, mature tornado is in progress, the observation density within the FFD outflow is relatively sparse; thus, the amplitude of the θρ deficit within the FFD outflow and the associated baroclinity cannot easily be assessed. The θρ analysis at this time also reveals a curious θρ excess south and southeast of the tornado, which has been previously documented by Markowski et al. (2002).

2) The 3 May 1999 storm

A significant tornado outbreak affected central Oklahoma and southern Kansas on 3 May 1999. Many aspects of this outbreak have been discussed in the literature, including the mesoscale and synoptic-scale conditions (Thompson and Edwards 2000; Roebber et al. 2002), forecasting and nowcasting issues (Edwards et al. 2002; Andra et al. 2002; Stensrud and Weiss 2002; McCarthy 2002), damage and casualty patterns (Brooks and Doswell 2002; Brown et al. 2002; Yuan et al. 2002; Marshall 2002; Pan et al. 2002), and targeted observations obtained from mobile platforms (Burgess et al. 2002; Wurman 2002; Markowski 2002b).

The mobile mesonet collected data over several hours within a tornadic supercell that moved northeastward from near Anadarko, Oklahoma, to the northwest of Oklahoma City by sunset (Markowski 2002b).3 A total of 20 tornadoes were identified with the storm in the 2236–0345 UTC period (Speheger et al. 2002). The tornadoes ranged from F0 to F4 in damage intensity. Although the storm was a prolific tornado producer, the storm produced mainly brief and weak tornadoes during the first few hours of its life (8 tornadoes were reported from 2236 to 0040 UTC, but all were rated F0 or F1 and only one persisted longer than 4 min; Speheger et al. 2002). It was during this time period that mobile mesonet sampling of the storm was best.

Figure 12 displays mobile mesonet data at 2356 UTC, at which time one of the aforementioned weak tornadoes was developing (Table 2). As was the case in the 2 June 1995 storm, the mobile mesonet sampled only small θυ, θρ, and θe deficits (θυmin, θρmin, and θemin were −2.5, −3.0, and −4.0 K, respectively; Figs. 12a–c). However, the magnitude of the maximum buoyancy gradient was larger than might be implied by the relatively modest negative buoyancy, with a |hθυ|max(|hθρ|max) of 0.7 K km−1 (1.0 K km−1). The relatively large buoyancy gradient was confined to a narrow corridor along the forward-flank radar reflectivity gradient (Fig. 12b). The estimated Δωθυsmax and Δωθρsmax values were 0.006 and 0.008 s−1, respectively (Table 3). The zo field could not be determined. The θe values at the surface within the FFD outflow were larger than the largest θe value on the only available sounding, which was launched at 0000 UTC at Norman, Oklahoma, approximately 75 km east of the tornadic storm. Although this sounding was not representative of the environment of the storm sampled by the mobile mesonet, it seems unlikely that air from above the boundary layer reached the surface within the FFD given the small θe deficits.

c. Comparison of FFD outflow properties in nontornadic and tornadic supercells

On average, the FFD outflow of nontornadic supercells was observed to have larger θυ, θρ, and θe deficits compared to tornadic supercells. The average θυmin, θρmin, and θemin was 2.6, 4.5, and 4.8 K, respectively, warmer in tornadic supercells (Table 4). The average height of zomax was 0.8 km higher in nontornadic cases than tornadic cases. Nontornadic supercells also were associated with larger buoyancy gradients and baroclinic generation of streamwise vorticity compared to tornadic supercells (|hθρ|max and Δωθρsmax were twice as large in nontornadic supercells). These results might seem surprising given the long-standing notion, reviewed in section 1, that the FFD and its associated baroclinic vorticity generation are critical to low-level mesocyclogenesis. We will attempt to reconcile the mobile mesonet observations in section 5.

Although the sample size of 12 supercells is microscopic in climatological terms, most of the differences between the mean FFD outflow properties of nontornadic and tornadic supercells were found to have statistical significance. The common test for significance is the t test, which assumes that the data sample follows the normal distribution. But since many atmospheric parameters are highly skewed and not normally distributed, a more appropriate test is the nonparametric test referred to as Multiresponse Permutation Procedures (MRPP; Mielke et al. 1981), which makes no assumption about the distribution of the population. MRPP gives a result in the form of a P value of statistical significance, where the P value is the probability that two sets of observations come from the same population (or 1 − P is the probability that two populations are different). The P values for most of the FFD outflow properties were <0.05 (Table 4).

4. Relationships between FFD outflow observations and the ambient environment

Relationships between FFD outflow observations and the ambient state of the atmosphere also were investigated (Table 5). The larger-scale parameters included the lifting condensation level (LCL), convective available potential energy (CAPE), surface dewpoint depression (Tddsfc), magnitude of the 0–6-km wind shear vector (|Δv|0−6), and storm-relative (sr) wind at 6 km AGL. With the exception of Tddsfc, which was determined from nearby mobile mesonet observations outside of the precipitation region of the storms, the parameters were obtained from the nearest sounding available in space and time that was in the inflow of the storm. Although the representativeness of a sounding is perhaps always in question (e.g., Brooks et al. 1994a; Markowski and Richardson 2004), it is worth noting that most of the soundings used herein satisfied Darkow’s (1969) proximity sounding criteria. LCL and CAPE values were computed by lifting a parcel having the mean potential temperature and specific humidity of the lowest 100 mb. This is obviously not an all-inclusive list of parameters that might be suspected a priori to exert some influence on the properties of FFD outflow, but it is hypothesized that the parameters cited above probably capture most of the ways by which the ambient environment might influence the outflow properties (e.g., midlevel entrainment, evaporation below cloud base, etc.). It might have been desirable to include sr wind and vertical shear data from altitudes above 6 km, but a large fraction of the special soundings launched on VORTEX case days lost wind data in the mid- to upper troposphere due to strong electric fields (D. Rust 2003, personal communication).

The largest correlations, in terms of magnitude, were found between measures of the buoyancy of FFD outflow (i.e., θυmin and θρmin) and the relative humidity of the near-storm inflow at the surface (Tddsfc) (Table 5). The correlation coefficient between θυmin and Tddsfc (θρmin and Tddsfc) was −0.75 (−0.81), with outflow buoyancy (and to a lesser extent, θemin) decreasing with the inflow relative humidity (Fig. 13). Not surprisingly, the horizontal buoyancy gradient and estimated streamwise baroclinic vorticity generation increased with decreasing inflow relative humidity. What was somewhat surprising was that the relationship between LCL and FFD outflow properties was not as pronounced (the magnitudes of the correlation coefficients were <0.4 for all outflow properties)—a result of differences between the thermodynamic properties of near-storm inflow at the surface and those sampled by soundings farther from the storm in space and time, and averaged over the lowest 100 mb.

What also might be surprising is the relatively weak relationship between CAPE and the buoyancy of FFD outflow (e.g., the CAPE–θρmin correlation was 0.27; Table 5), and the fact that outflow buoyancy increased as CAPE increased, which conflicts with the numerical simulation results of Weisman and Klemp (1982). Furthermore, as CAPE increased, the magnitude of the baroclinity and associated streamwise vorticity generation within the FFD outflow decreased. It is not known to what extent these discrepancies are a result of overly simplistic microphysics parameterizations within the numerical simulations versus the smallness or representativeness of the sample of observed cases.

As the 0–6-km vertical wind shear increased, θυmin, θρmin, and θemin increased, that is, warmer FFD outflow was associated with larger shear, although the correlations were not particularly large (0.20–0.32; Table 5). The magnitudes of the horizontal buoyancy gradients and the streamwise vorticity generation decreased with increasing vertical wind shear, with the correlations being similar in magnitude (−0.24 to −0.37; Table 5). Last, as the sr wind speed at 6 km AGL increased, θemin(zomax) tended to decrease (increase); however, the correlations were not notably large (∼0.3 in magnitude; Table 5). A weak tendency for the buoyancy gradients (but not the streamwise vorticity generation) to decrease as the sr wind increased also was observed.

5. Discussion

The finding that the FFD outflow of nontornadic supercells tends to be more negatively buoyant and have larger θe deficits than the FFD outflow of tornadic supercells is similar to the observations within the outflow of supercell RFDs reported by Markowski et al. (2002). The results herein are also consistent with prior climatological studies that have shown that supercells are more likely to be tornadic when the boundary layer relative humidity is large (e.g., Rasmussen and Blanchard 1998; Thompson et al. 2003). The fact that FFD outflow tends to be more negatively buoyant in relatively dry ambient low-level environments than in relatively humid ambient conditions is not surprising given the long-standing notion that the FFD arises from latent chilling (much of which is evaporation) in the region where hydrometeors fall to the ground downstream of the updraft. This finding also suggests that graupel melting, which increases as the relative humidity increases (Srivastava 1987), tends to be less important than evaporation in the production of negative buoyancy within the near-ground FFD outflow.

The mobile mesonet observations lead to obvious questions, however, regarding the role of the FFD and its outflow in low-level mesocyclogenesis and tornadogenesis. Numerical simulations of supercells have demonstrated the importance of baroclinity and streamwise vorticity generation in the outflow of the FFD in the genesis of low-level mesocyclones (e.g., Klemp and Rotunno 1983; Rotunno and Klemp 1985), yet the baroclinity and estimated streamwise vorticity generation were largest in the present nontornadic cases. This might lead one to consider the awkward possibility that the strongest low-level mesocyclones are associated with nontornadic supercells; that is, could it be that the processes that give rise to low-level mesocyclogenesis also disrupt the processes that intensify vorticity into a smaller-scale tornado, as questioned by Trapp (1999)?4 On the other hand, some of strongest forward-flank baroclinity was observed by the mobile mesonet in the 22 May 1995 storm, yet this nontornadic supercell also possessed the weakest low-level mesocyclone in the Trapp (1999) study.

The observations suggest that the outflow of FFDs may play a lesser dynamical role in tornadic supercells, but it probably cannot be concluded that the outflow of the FFD is unimportant in the amplification of low-level vorticity that may lead to tornadogenesis (i.e., “lesser” and “unimportant” are not synonymous). It cannot be known how the tornadic supercells would have evolved in the absence of their weaker baroclinity (relative to nontornadic supercells)—the mobile mesonet observations do not refute the hypothesis that the FFD outflow provides an important source of low-level streamwise vorticity to the updraft. In general, it is difficult to isolate the dynamical importance of the FFD outflow in an observational study because each environment possesses different amounts and distributions of ambient vorticity, not to mention the ambient thermodynamic differences from case to case. These environmental differences by themselves could account for almost innumerable differences in storm morphology and evolution, masking the differences resulting from the dynamical processes directly associated with the FFD outflow. It also is difficult to know how best to compare FFD outflow characteristics with mesocyclone characteristics (e.g., altitude, evolutionary stage, and the time lag between baroclinic vorticity generation and mesocyclone response).

We speculate that the substantial baroclinic vorticity generation in the FFD outflow might be unfavorable for low-level mesocyclones and tornadogenesis not for direct reasons, but for indirect reasons, due to the fact that large baroclinity is unavoidably accompanied by large density excesses. As earlier indicated, there is growing evidence, direct and indirect, that strong cold pools and excessive negative buoyancy are associated with storms that have difficulty maintaining low-level mesocyclones (Brooks et al. 1993, 1994b) and producing tornadoes (Rasmussen and Blanchard 1998; Markowski et al. 2002; Thompson et al. 2003; Grzych et al. 2004; Lee et al. 2004). Perhaps the relative orientation of the baroclinic vorticity produced in the FFD outflow and the ambient horizontal vorticity associated with the mean vertical shear is important, as suggested by Wicker (1996). The relative orientation would control how much total horizontal vorticity is ingested by the updraft (i.e., how much the baroclinic vorticity augments the ambient horizontal vorticity).

It may also be noteworthy that environmental, low-level, streamwise vorticity tends to be larger in tornadic supercell environments than in nontornadic supercell environments (Markowski et al. 2003; Rasmussen 2003; Thompson et al. 2003); therefore, perhaps additional horizontal vorticity generation within the FFD outflow is unnecessary for these storms to acquire strong low-level mesocyclones upon tilting horizontal vorticity. Within the most favorable environments for tornadic supercells (i.e., those with large ambient, low-level streamwise vorticity and relative humidity values), high relative humidity would tend to promote weak baroclinity within the FFD outflow, and the presence of large ambient streamwise vorticity might obviate the need for baroclinic streamwise vorticity production in the FFD outflow. Thus, an updraft could ingest potentially large amounts of streamwise vorticity without having to overcome the seemingly detrimental effects of excessive negative buoyancy in its own outflow.

At first glance it also may seem difficult to reconcile the present findings with past observations that supercell interactions with preexisting mesoscale boundaries often are favorable for tornadogenesis (Markowski et al. 1998b; Atkins et al. 1999; Rasmussen et al. 2000). These past studies hypothesized that the enhancement of the ambient horizontal vorticity by baroclinic generation along the boundaries promoted tornadogenesis in many storm–boundary interactions. Yet the present results suggest that storms with strong baroclinity are less likely to be tornadic. One possibility is that the mesoscale boundaries that are most favorable for promoting tornadogenesis following a storm–boundary interaction are those that separate the ambient air mass from an air mass that has been substantially modified, such that the air masses on both sides of the boundary are characterized by relatively small convective inhibition (e.g., Rasmussen et al. 2000). Not all storm–boundary interactions are favorable for tornadogenesis (Markowski et al. 1998b; Doswell et al. 2002). Perhaps the mesoscale boundaries involved in storm interactions that fail to lead to tornadogenesis (or even lead to storm demise) have stability characteristics on their cold sides that are more similar to the outflow of the FFDs of nontornadic supercells.

6. Conclusions

This paper has examined in situ observations obtained at the surface by a mobile mesonet within the FFD outflow of 12 supercells. The observations support the following conclusions.

  1. The FFD outflow of the nontornadic supercells was more negatively buoyant than the FFD outflow of the tornadic supercells.
  2. The FFD outflow of the nontornadic supercells was associated with larger horizontal buoyancy gradients and baroclinic generation of streamwise vorticity than the FFD outflow of the tornadic supercells.
  3. Of the environmental wind and thermodynamic parameters evaluated, the best predictor of the buoyancy, buoyancy gradients, and streamwise vorticity generation within the FFD outflow was the relative humidity (dewpoint depression) at the surface in the near-storm inflow.
  4. The θe deficits sampled at the surface within the FFD outflow were larger in the nontornadic supercells than in the tornadic supercells.
  5. The altitude at which θe values observed at the surface within the FFD outflow matched the environmental θe was higher in the nontornadic supercell cases than in the tornadic supercell cases.

The results raise some questions regarding the relationships among the baroclinic generation of horizontal vorticity, low-level mesocyclones, and tornadoes. Previous numerical simulations of supercells have demonstrated the importance of baroclinity and streamwise vorticity generation within the FFD outflow in the genesis of low-level mesocyclones, yet the baroclinity and estimated streamwise vorticity generation were largest in the nontornadic storms (which were also found by prior investigators to possess weak low-level mesocyclones) sampled by the mobile mesonet. We anxiously await the results of future numerical simulations utilizing more sophisticated microphysics, and perhaps even radiative transfer schemes, as well as future field experiments taking advantage of steadily improving observational capabilities. These should improve our representations and diagnoses of the buoyancy fields of supercell thunderstorms. More complete depictions of the buoyancy fields, in conjunction with the three dimensionality and temporal resolution lacking in this study, should lead to more accurate assessments of the role of the FFD and its outflow in the intensification of near-ground rotation.

Acknowledgments

We are grateful to all of the volunteers (too many to mention) who were involved in the data collection that made this work possible. Drs. Yvette Richardson and John Clark provided reviews of an earlier version of this paper. We also thank the two anonymous reviewers for their suggestions for improving our presentation. Objective analyses of radar reflectivity data were created using the REORDER software developed at the National Center for Atmospheric Research. This research was supported in part by National Science Foundation Grant ATM-0338661 made to The Pennsylvania State University.

REFERENCES

  • Andra, D. L., , E. M. Quoetone, , and W. F. Bunting, 2002: Warning decision making: The relative roles of conceptual models, technology, strategy, and forecaster expertise on 3 May 1999. Wea. Forecasting, 17 , 559566.

    • Search Google Scholar
    • Export Citation
  • Atkins, N. T., , M. L. Weisman, , and L. J. Wicker, 1999: The influence of preexisting boundaries on supercell evolution. Mon. Wea. Rev., 127 , 29102927.

    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3 , 396409.

  • Barnes, S. L., 1978: Oklahoma thunderstorms on 29–30 April 1970. Part I: Morphology of a tornadic storm. Mon. Wea. Rev., 106 , 673684.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., , and S. G. Gaddy, 2001: Airborne pseudo-dual-Doppler analysis of a rear-inflow jet and deep convergence zone within a supercell. Mon. Wea. Rev., 129 , 22702289.

    • Search Google Scholar
    • Export Citation
  • Bolton, D., 1980: The computation of equivalent potential temperature. Mon. Wea. Rev., 108 , 10461053.

  • Brandes, E. A., 1981: Finestructure of the Del City-Edmond tornadic mesocirculation. Mon. Wea. Rev., 109 , 635647.

  • Brandes, E. A., 1984: Relationships between radar-derived thermodynamic variables and tornadogenesis. Mon. Wea. Rev., 112 , 10331052.

  • Brooks, H. E., , and C. A. Doswell III, 2002: Deaths in the 3 May 1999 Oklahoma City Tornado from a historical perspective. Wea. Forecasting, 17 , 354361.

    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., , C. A. Doswell III, , and R. Davies-Jones, 1993: Environmental helicity and the maintenance and evolution of low-level mesocyclones. The Tornado: Its Structure, Dynamics, Prediction and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 143–159.

  • Brooks, H. E., , C. A. Doswell III, , and J. Cooper, 1994a: On the environments of tornadic and nontornadic mesocyclones. Wea. Forecasting, 9 , 606618.

    • Search Google Scholar
    • Export Citation
  • Brooks, H. E., , C. A. Doswell III, , and R. B. Wilhelmson, 1994b: The role of midtropospheric winds in the evolution and maintenance of low-level mesocyclones. Mon. Wea. Rev., 122 , 126136.

    • Search Google Scholar
    • Export Citation
  • Brown, S., , P. Archer, , E. Kruger, , and S. Mallonee, 2002: Tornado-related deaths and injuries in Oklahoma due to the 3 May 1999 tornadoes. Wea. Forecasting, 17 , 343353.

    • Search Google Scholar
    • Export Citation
  • Burgess, D. W., , R. A. Brown, , L. R. Lemon, , and C. R. Safford, 1977: Evolution of a tornadic thunderstorm. Preprints, 10th Conf. on Severe Local Storms, Omaha, NE, Amer. Meteor. Soc., 84–89.

  • Burgess, D. W., , M. A. Magsig, , J. Wurman, , D. C. Dowell, , and Y. Richardson, 2002: Radar observations of the 3 May 1999 Oklahoma City Tornado. Wea. Forecasting, 17 , 456471.

    • Search Google Scholar
    • Export Citation
  • Darkow, G. L., 1969: An analysis of over sixty tornado proximity soundings. Preprints, Sixth Conf. on Severe Local Storms, Chicago, IL, Amer. Meteor. Soc., 218–221.

  • Davies-Jones, R., 1984: Streamwise vorticity: The origin of updraft rotation in supercell storms. J. Atmos. Sci., 41 , 29913006.

  • Davies-Jones, R., , R. J. Trapp, , and H. B. Bluestein, 2001: Tornadoes and tornadic storms. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 167–222.

  • Doswell III, C. A., , D. V. Baker, , and C. A. Liles, 2002: Recognition of negative mesoscale factors for severe-weather potential: A case study. Wea. Forecasting, 17 , 937954.

    • Search Google Scholar
    • Export Citation
  • Dowell, D. C., , and H. B. Bluestein, 1997: The Arcadia, Oklahoma, storm of May 1981: Analysis of a supercell during tornadogenesis. Mon. Wea. Rev., 125 , 25622582.

    • Search Google Scholar
    • Export Citation
  • Edwards, R., , S. F. Corfidi, , R. L. Thompson, , J. S. Evans, , J. P. Craven, , J. P. Racy, , D. W. McCarthy, , and M. D. Vescio, 2002: Storm Prediction Center forecasting issues related to the 3 May 1999 tornado outbreak. Wea. Forecasting, 17 , 544558.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 1994: Atmospheric Convection. Oxford University Press, 580 pp.

  • Fujita, T. T., 1975: New evidence from the April 3–4, 1974 tornadoes. Preprints, Ninth Conf. on Severe Local Storms, Norman, OK, Amer. Meteor. Soc., 248–255.

  • Gilmore, M. S., , and L. J. Wicker, 2002: Influences of the local environment on supercell cloud-to-ground lightning, radar characteristics, and severe weather on 2 June 1995. Mon. Wea. Rev., 130 , 23492372.

    • Search Google Scholar
    • Export Citation
  • Grzych, M. L., , B. D. Lee, , C. A. Finley, , and J. L. Schroeder, 2004: Thermodynamic characterization of supercell rear-flank downdrafts in Project ANSWERS 2003. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, P11.1.

  • Hane, C. E., , and P. S. Ray, 1985: Pressure and buoyancy fields derived from Doppler radar data in a tornadic thunderstorm. J. Atmos. Sci., 42 , 1835.

    • Search Google Scholar
    • Export Citation
  • Klemp, J. B., 1987: Dynamics of tornadic thunderstorms. Annu. Rev. Fluid Mech., 19 , 369402.

  • Klemp, J. B., , and R. Rotunno, 1983: A study of the tornadic region within a supercell thunderstorm. J. Atmos. Sci., 40 , 359377.

  • Koch, S. E., , M. desJardins, , and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor., 22 , 14871503.

    • Search Google Scholar
    • Export Citation
  • Lee, B. D., , C. A. Finley, , and P. Skinner, 2004: Thermodynamic and kinematic analysis of multiple RFD surges for the 24 June 2003 Manchester, SD cyclic tornadic supercell during Project ANSWERS 2003. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, P11.2.

  • Lemon, L. R., , and C. A. Doswell III, 1979: Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Mon. Wea. Rev., 107 , 11841197.

    • Search Google Scholar
    • Export Citation
  • Ludlum, F. H., 1963: Severe local storms: A review. Severe Local Storms, Meteor. Monogr., No. 27, Amer. Meteor. Soc., 1–30.

  • Markowski, P. M., 2002a: Hook echoes and rear-flank downdrafts: A review. Mon. Wea. Rev., 130 , 852876.

  • Markowski, P. M., 2002b: Mobile mesonet observations on 3 May 1999. Wea. Forecasting, 17 , 430444.

  • Markowski, P. M., , and Y. P. Richardson, 2004: Multiple-Doppler radar observations of vertical wind profile heterogeneity in convective boundary layers. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, P13.1.

  • Markowski, P. M., , and J. Y. Harrington, 2005: A simulation of a supercell thunderstorm with emulated radiative cooling beneath the anvil. J. Atmos. Sci., 62 , 26072617.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , E. N. Rasmussen, , J. M. Straka, , and D. C. Dowell, 1998a: Observations of low-level baroclinity generated by anvil shadows. Mon. Wea. Rev., 126 , 29422958.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , E. N. Rasmussen, , and J. M. Straka, 1998b: The occurrence of tornadoes in supercells interacting with boundaries during VORTEX-95. Wea. Forecasting, 13 , 852859.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , J. M. Straka, , and E. N. Rasmussen, 2002: Direct surface thermodynamic measurements within the rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 16921721.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , J. M. Straka, , and E. N. Rasmussen, 2003: Tornadogenesis resulting from the transport of circulation by a downdraft: Idealized numerical simulations. J. Atmos. Sci., 60 , 795823.

    • Search Google Scholar
    • Export Citation
  • Marshall, T. P., 2002: Tornado damage survey at Moore, Oklahoma. Wea. Forecasting, 17 , 582598.

  • McCarthy, D. H., 2002: The role of ground-truth reports in the warning-decision making process during the 3 May 1999 Oklahoma tornado outbreak. Wea. Forecasting, 17 , 647649.

    • Search Google Scholar
    • Export Citation
  • Mielke, P. W., , K. J. Berry, , and G. W. Brier, 1981: Application of multi-response permutation procedures for examining seasonal changes in monthly mean sea-level pressure patterns. Mon. Wea. Rev., 109 , 120126.

    • Search Google Scholar
    • Export Citation
  • Pan, K., , P. Montpellier, , and M. Zadeh, 2002: Engineering observations of the 3 May 1999 Oklahoma tornado damage. Wea. Forecasting, 17 , 599610.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., 2003: Refined supercell and tornado forecast parameters. Wea. Forecasting, 18 , 530535.

  • Rasmussen, E. N., , and J. M. Straka, 1996: Mobile mesonet observations of tornadoes during VORTEX-95. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 1–5.

  • Rasmussen, E. N., , and D. O. Blanchard, 1998: A baseline climatology of sounding-derived supercell and tornado forecast parameters. Wea. Forecasting, 13 , 11481164.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., , J. M. Straka, , R. Davies-Jones, , C. A. Doswell III, , F. H. Carr, , M. D. Eilts, , and D. R. MacGorman, 1994: Verification of the origins of rotation in tornadoes experiment: VORTEX. Bull. Amer. Meteor. Soc., 75 , 9951006.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., , S. J. Richardson, , J. M. Straka, , P. M. Markowski, , and D. O. Blanchard, 2000: The association of significant tornadoes with a baroclinic boundary on 2 June 1995. Mon. Wea. Rev., 128 , 174191.

    • Search Google Scholar
    • Export Citation
  • Roebber, P. J., , D. M. Schultz, , and R. Romero, 2002: Synoptic regulation of the 3 May 1999 tornado outbreak. Wea. Forecasting, 17 , 399429.

    • Search Google Scholar
    • Export Citation
  • Rotunno, R., 1981: On the evolution of thunderstorm rotation. Mon. Wea. Rev., 109 , 577586.

  • Rotunno, R., , and J. B. Klemp, 1985: On the rotation and propagation of simulated supercell thunderstorms. J. Atmos. Sci., 42 , 271292.

    • Search Google Scholar
    • Export Citation
  • Speheger, D. A., , C. A. Doswell III, , and G. J. Stumpf, 2002: The tornadoes of 3 May 1999: Event verification in central Oklahoma and related issues. Wea. Forecasting, 17 , 362381.

    • Search Google Scholar
    • Export Citation
  • Srivastava, R. C., 1987: A model of intense downdrafts driven by the melting and evaporation of precipitation. J. Atmos. Sci., 44 , 17521774.

    • Search Google Scholar
    • Export Citation
  • Stensrud, D. J., , and S. J. Weiss, 2002: Mesoscale model ensemble forecasts of the 3 May 1999 tornado outbreak. Wea. Forecasting, 17 , 526543.

    • Search Google Scholar
    • Export Citation
  • Straka, J. M., , E. N. Rasmussen, , and S. E. Fredrickson, 1996: A mobile mesonet for fine-scale meteorological observations. J. Atmos. Oceanic Technol., 13 , 921936.

    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , and R. Edwards, 2000: An overview of environmental conditions and forecast implications of the 3 May 1999 tornado outbreak. Wea. Forecasting, 15 , 682699.

    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , R. Edwards, , J. A. Hart, , K. L. Elmore, , and P. Markowski, 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting, 18 , 12431261.

    • Search Google Scholar
    • Export Citation
  • Trapp, R. J., 1999: Observations of nontornadic low-level mesocyclones and attendant tornado-genesis failure during VORTEX. Mon. Wea. Rev., 127 , 16931705.

    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., 2001: Convectively driven high wind events. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 255–298.

  • Wakimoto, R. M., , and H. Cai, 2000: Analysis of a nontornadic storm during VORTEX 95. Mon. Wea. Rev., 128 , 565592.

  • Wakimoto, R. M., , H. V. Murphey, , D. C. Dowell, , and H. B. Bluestein, 2003: The Kellerville Tornado during VORTEX: Damage survey and Doppler radar analyses. Mon. Wea. Rev., 131 , 21972221.

    • Search Google Scholar
    • Export Citation
  • Weisman, M. L., , and J. B. Klemp, 1982: The dependence of numerically simulated convective storms on vertical wind shear and buoyancy. Mon. Wea. Rev., 110 , 504520.

    • Search Google Scholar
    • Export Citation
  • Wicker, L. J., 1996: The role of near surface wind shear on low-level mesocyclone generation and tornadoes. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 115–119.

  • Wilhelmson, R. B., , and L. J. Wicker, 2001: Numerical modeling of severe local storms. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 123–166.

  • Wurman, J., 2002: The multiple vortex structure of a tornado. Wea. Forecasting, 17 , 473505.

  • Yuan, M., , M. Dickens-Micozzi, , and M. A. Magsig, 2002: Analysis of tornado damage tracks from the 3 May tornado outbreak using multispectral satellite imagery. Wea. Forecasting, 17 , 382398.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

(left) Schematic representation of a supercell thunderstorm, adapted from the conceptual model presented by Lemon and Doswell (1979). The dark gray shading approximates the precipitation region, at least as viewed by radar (e.g., radar reflectivity factor >30 dBZ). Within this region, the locations of the FFD and RFD are indicated. The light gray shading approximates the main updraft region (U). The typical location of a tornado, if one occurs, also is indicated (T). The outflow boundary is indicated by the barbed contour. A few storm-relative streamlines are drawn. (right) Photograph of a tornadic supercell that closely resembles the schematic shown on the left. The approximate vantage point of the photograph is indicated by the star in the schematic, roughly east of the main updraft. The labels in the photograph correspond to the same features depicted in the schematic. (Photograph courtesy of E. Rasmussen)

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 2.
Fig. 2.

Three-dimensional schematic of a numerically simulated supercell thunderstorm in westerly mean shear, viewed from the southeast, at a stage when low-level rotation is intensifying. The cylindrical arrows depict the storm-relative winds. The thin lines are vortex lines, with the sense of rotation indicated by the circular arrows. The heavy barbed line marks the outflow boundary. The orientation of the horizontal buoyancy gradient, hB, also is indicated. [Adapted from Klemp (1987)]

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 3.
Fig. 3.

Retrieved virtual potential temperature perturbations (θυ; K) at 1.3 km AGL in the (left) Del City, OK, supercell (0045 UTC 21 May 1977), and in the (right) Harrah, OK, supercell (2143 UTC 8 Jun 1974) studied by Brandes (1984). The gray region indicates where radar reflectivity factor exceeds 30 dBZ (40 dBZ) in the 21 May 1977 (8 Jun 1974) case. Positive (negative) perturbations are contoured using thin solid (dashed) lines (note that the contour interval is 2 K on 21 May 1977 and 1 K on 8 Jun 1974). The heavy dashed lines enclose regions where vertical vorticity exceeds 0.01 s−1. [Adapted from Brandes (1984). The figure actually displays fields of “quasi-virtual” potential temperature perturbations, which Brandes (1984) defines as B = θ′ + 0.61qυ, where θ′, , and qυ are the potential temperature perturbation, base-state potential temperature, and specific humidity perturbation, respectively. According to Brandes (1984), B values typically are within 0.1 K of θυ values]

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 4.
Fig. 4.

Schematic diagram illustrating how the FFD outflow region was objectively defined. The radar echo is shaded dark gray, the updraft is shaded light gray, and the FFD outflow region is defined by the hatching. The FFD outflow region was defined as the region of low-level radar reflectivity on the forward side of a line drawn orthogonal to the major axis of the echo, through the radar-observed mesocyclone center (M).

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 5.
Fig. 5.

The relationship between radar reflectivity factor (dBZ) and parameterized condensate, qc, and the difference between the virtual and density potential temperature, θυθρ, assuming a potential temperature, θ, of 300 K.

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 6.
Fig. 6.

Density potential temperature fields for all 12 cases analyzed. Unshaded areas indicate that observations were unavailable. The heavy solid line encloses the 40-dBZ radar echo. The filled circles indicate the locations of the mesocyclone centers at an altitude of approximately 1–2 km AGL.

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 7.
Fig. 7.

Subjective analyses at 2349 UTC 22 May 1995 of (a) virtual potential temperature perturbations (θυ; 1-K contour interval), (b) density potential temperature perturbations (θρ; 1-K contour interval), (c) equivalent potential temperature perturbations (θe; 2-K contour interval), and (d) the height on an inflow sounding where the surface θe value is equal to that observed at the surface (zo; 0.4-km contour interval). The reference values are υ = ρ = 310.9 K and e = 353.0 K. The analyses have been overlaid on objectively analyzed radar reflectivity data at 0.5 km AGL from the NOAA P3 lower fuselage radar (see legend). The filled circle indicates the location of the mesocyclone center at an altitude of approximately 1–2 km AGL. Station models display storm-relative winds (a full barb equals 5 m s−1; a half-barb equals 2.5 m s−1) and the observed values of the variables analyzed, except in (d), in which station models display θe values as in (c). To avoid clutter, only a small number of station models are plotted (data were recorded at a frequency of 2 Hz, which corresponds to an along-track horizontal data spacing of 40 m given a typical vehicle speed of 20 m s−1).

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 8.
Fig. 8.

Subjective analyses of density potential temperature perturbations (θρ; 1-K contour interval) at 2338 and 2359 UTC 22 May 1995 (approximately 10 min before and after the analysis time in Fig. 7) (ρ = 310.9 K). All symbols are identical to those appearing in Fig. 7.

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 9.
Fig. 9.

Same as in Fig. 7, but for 0116 UTC 21 May 1998. The reference values are υ = ρ = 312.0 K and e = 342.1 K. The zo field is contoured every 0.2 km. The radar reflectivity data are from the lowest elevation scan of the KGLD WSR-88D (the data are at approximately 0.8 km AGL).

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 10.
Fig. 10.

Same as in Fig. 7, but for 0058 UTC 3 Jun 1995. The reference values are υ = ρ = 308.9 K and e = 352.9 K. The zo field is contoured every 0.2 km. The radar reflectivity data are at 0.5 km AGL and are from the NOAA P3 lower fuselage radar.

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 11.
Fig. 11.

Subjective analyses of density potential temperature perturbations (θρ; 1-K contour interval) at 0048 and 0108 UTC 3 Jun 1995 (10 min before and after the analysis time in Fig. 10) (ρ = 308.9 K). All symbols are identical to those appearing in Fig. 10.

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 12.
Fig. 12.

Same as in Fig. 7, but for 2356 UTC 3 May 1999. The field of zo is not displayed because a suitable proximity sounding was not available. The reference values are υ = ρ = 304.8 K and e = 352.0 K. The radar reflectivity data are from the lowest elevation scan of the KTLX WSR-88D (the data are at approximately 0.6 km AGL).

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Fig. 13.
Fig. 13.

Scatter diagram of minimum density potential temperature perturbation, θρmin, within the FFD outflow sampled vs surface dewpoint depression, Tddsfc, in the near-storm inflow. Units are K. A best-fit line also has been added. The linear correlation coefficient, r, appears at the top right.

Citation: Monthly Weather Review 134, 5; 10.1175/MWR3131.1

Table 1.

Cases whereby mobile mesonet observations were obtained within the FFD outflow of supercells. All times are UTC. The date given for each case is with respect to local, not UTC, time. Tornado times and intensity were based on reports made by field experiment personnel and VORTEX data, where available, and may not agree with those published in Storm Data by the National Oceanic and Atmospheric Administration. The time of maximum low-level rotation is tabulated for nontornadic supercell cases. For the 29 Apr 1995, 12 May 1995, and 22 May 1995 cases, the times listed correspond to the times of “tornadogenesis failure” defined by Trapp (1999). For the 8 Jun 1995a and 20 May 1998 cases, the time of maximum low-level rotation corresponds to the time of peak azimuthal wind shear in the radial velocity data obtained from the 0.5° elevation angle scan of the nearest WSR-88D.

Table 1.
Table 2.

Analysis times (UTC) of best FFD outflow sampling and quality of the observations, as defined by the fraction of the FFD outflow region within 1 km of a mobile mesonet observation. The FFD outflow regions having an objectively analyzed radar reflectivity factor, Z, exceeding 30 and 45 dBZ were both considered. In the rightmost column, a Y (N) indicates that mobile mesonet observations were (not) available within 1 km of the forward-flank reflectivity maximum, which might be expected to be at or near the buoyancy minimum. The numerals in parentheses after the analysis times indicate the number of minutes between the analysis time and the time of tornadogenesis (tornadic cases) or maximum rotational velocity (nontornadic cases; refer to Table 1). Negative (positive) values indicate that the analysis time is prior to (after) the time of tornadogenesis or maximum rotational velocity.

Table 2.
Table 3.

Thermodynamic characteristics of FFD outflow sampled by the mobile mesonet: minimum virtual potential temperature perturbation (θυmin; K); minimum density potential temperature perturbation (θρmin; K); minimum equivalent potential temperature perturbation (θemin; K); maximum altitude of origin of FFD air (zomax; km), maximum magnitude of the horizontal gradient of virtual potential temperature perturbations (|hθυ|max; K km−1); maximum magnitude of the horizontal gradient of density potential temperature perturbations (|hθρ|max; K km−1); estimated maximum streamwise vorticity acquired from baroclinic vorticity generation over a 5-km distance based on the θυ gradient (i.e., neglecting condensate loading) (Δωθυsmax; s−1); estimated maximum streamwise vorticity acquired from baroclinic vorticity generation over a 5-km distance based on the θρ gradient (i.e., including parameterized condensate loading) (Δωθρsmax; s−1). Tornadic cases are in bold. “Unknown” appears where either the mobile mesonet data were inadequate or a suitable proximity sounding was unavailable.

Table 3.
Table 4.

Mean thermodynamic characteristics (standard deviations in parentheses) of the FFD outflow sampled by the mobile mesonet: minimum virtual potential temperature perturbation (θυmin; K); minimum density potential temperature perturbation (θρmin; K); minimum equivalent potential temperature perturbation (θemin; K); maximum altitude of origin of FFD air (zomax; km), maximum magnitude of the horizontal gradient of virtual potential temperature perturbations (|hθυ|max; K km−1); maximum magnitude of the horizontal gradient of density potential temperature perturbations (|hθρ|max; K km−1); estimated maximum streamwise vorticity acquired from baroclinic vorticity generation over a 5-km distance based on the θυ gradient (i.e., neglecting condensate loading) (Δωθυsmax; s−1); estimated maximum streamwise vorticity acquired from baroclinic vorticity generation over a 5-km distance based on the θρ gradient (i.e., including parameterized condensate loading) (Δωθρsmax; s−1). All P values were computed using the MRPP technique of Mielke et al. (1981).

Table 4.
Table 5.

Linear correlation coefficients between the thermodynamic characteristics of FFD outflow (see Table 3 caption for definition of variables) and assorted environmental parameters determined by proximity surface observations or soundings. The surface dewpoint depression measured by the mobile mesonet in the near-storm inflow is Tddsfc, the LCL and CAPE were computed from proximity soundings by lifting a parcel having the mean potential temperature and specific humidity of the lowest 100 mb, and the magnitude of the 0–6-km wind shear vector is |Δv|0–6. Correlation magnitudes >0.5 are in bold.

Table 5.
1

It was noted by a reviewer that many references to FFDs really are referring to the outflow of the FFD, which may not be within the FFD at all. For example, FFD outflow is ingested by the supercell updraft to produce “wall clouds” (Rotunno and Klemp 1985). We are careful to distinguish between FFD and FFD outflow throughout the paper.

2

The θe computed is really the “pseudo-equivalent potential temperature,” which Bolton (1980) defined as θep.

3

This storm has been referred to as “storm B” by National Weather Service damage surveyors and in previous papers on the tornado outbreak (e.g., Speheger et al. 2002).

4

Trapp’s (1999) statement was based on the simulations of Brooks et al. (1993, 1994b), which showed that strong cold pools were needed for low-level mesocyclogenesis via the tilting of baroclinic vorticity, but that excessively strong outflow could undercut the updraft. Trapp (1999) inferred that updrafts that have been undercut by outflow do not readily produce tornadoes. It is worth noting, however, that the nontornadic supercells in this study did not appear to be undercut by their own outflow to the same degree as those in the Brooks et al. studies. Trapp (1999), Wakimoto and Cai (2000), and Markowski et al. (2002) also have noted tornadogenesis “failure modes” whereby the updraft does not appear to be undercut significantly by its outflow (e.g., when the gust front leads the updraft by several miles).

Save