• Achenbach, E., 1972: Experiments on the flow past spheres at very high Reynolds numbers. J. Fluid Mech., 54, 565575, https://doi.org/10.1017/S0022112072000874.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Armijo, L., 1969: A theory for the determination of wind and precipitation velocities with Doppler radars. J. Atmos. Sci., 26, 570573, https://doi.org/10.1175/1520-0469(1969)026<0570:ATFTDO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balakrishnan, N., and D. S. Zrnić, 1990: Use of polarization to characterize precipitation and discriminate large hail. J. Atmos. Sci., 47, 15251540, https://doi.org/10.1175/1520-0469(1990)047<1525:UOPTCP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., 1970: Some aspects of a severe, right-moving thunderstorm deduced from mesonetwork rawinsonde observations. J. Atmos. Sci., 27, 634648, https://doi.org/10.1175/1520-0469(1970)027<0634:SAOASR>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, M. M., M. T. Montgomery, and K. A. Emanuel, 2012: Air–sea enthalpy and momentum exchange at major hurricane wind speeds observed during CBLAST. J. Atmos. Sci., 69, 31973222, https://doi.org/10.1175/JAS-D-11-0276.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., E. W. McCaul, G. P. Byrd, and G. R. Woodall, 1988: Mobile sounding observations of a tornadic storm near the dryline: The Canadian, Texas, storm of 7 May 1986. Mon. Wea. Rev., 116, 17901804, https://doi.org/10.1175/1520-0493(1988)116<1790:MSOOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., E. W. McCaul, G. P. Byrd, G. R. Woodall, G. Martin, S. Keighton, and L. C. Showell, 1989: Mobile sounding observations of a thunderstorm near the dryline: The Gruver, Texas, storm complex of 25 May 1987. Mon. Wea. Rev., 117, 244250, https://doi.org/10.1175/1520-0493(1989)117<0244:MSOOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bousquet, O., P. Tabary, and J. Parent du Châtelet, 2008: Operational multiple-Doppler wind retrieval inferred from long-range radial velocity measurements. J. Appl. Meteor. Climatol., 47, 29292945, https://doi.org/10.1175/2008JAMC1878.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Browning, K. A., and F. H. Ludlam, 1962: Airflow in convective storms. Quart. J. Roy. Meteor. Soc., 88, 117135, https://doi.org/10.1002/qj.49708837602.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Browning, K. A., and G. B. Foote, 1976: Airflow and hail growth in supercell storms and some implications for hail suppression. Quart. J. Roy. Meteor. Soc., 102, 499533, https://doi.org/10.1002/qj.49710243303.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brunkow, D., V. N. Bringi, P. C. Kennedy, S. A. Rutledge, V. Chandrasekar, E. A. Mueller, and R. K. Bowie, 2000: A description of the CSU–CHILL national radar facility. J. Atmos. Oceanic Technol., 17, 15961608, https://doi.org/10.1175/1520-0426(2000)017<1596:ADOTCC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bryan, G. H., 2008: getcape. Accessed 1 July 2019, https://www2.mmm.ucar.edu/people/bryan/Code/getcape.F.

  • Chisholm, A. J., 1970: Alberta hailstorms: A radar study and model. Ph.D. thesis, McGill University, 237 pp.

  • Chisholm, A. J., 1973: Alberta hailstorms Part I: Radar case studies and airflow models. Alberta Hailstorms, Meteor. Monogr., No. 36, Amer. Meteor. Soc., 1–36.

    • Crossref
    • Export Citation
  • Collis, S., A. Protat, and K.-S. Chung, 2010: The effect of radial velocity gridding artifacts on variationally retrieved vertical velocities. J. Atmos. Oceanic Technol., 27, 12391246, https://doi.org/10.1175/2010JTECHA1402.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dahl, N. A., A. Shapiro, C. K. Potvin, A. Theisen, J. G. Gebauer, A. D. Schenkman, and M. Xue, 2019: High-resolution, rapid-scan dual-Doppler retrievals of vertical velocity in a simulated supercell. J. Atmos. Oceanic Technol., 36, 14771500, https://doi.org/10.1175/JTECH-D-18-0211.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies-Jones, R. P., 1974: Discussion of measurements inside high-speed thunderstorm updrafts. J. Appl. Meteor., 13, 710717, https://doi.org/10.1175/1520-0450(1974)013<0710:DOMIHS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies-Jones, R. P., and J. H. Henderson, 1975: Updraft properties deduced statistically from Rawin soundings. Pure Appl. Geophys., 113, 787801, https://doi.org/10.1007/BF01592959.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • DiGangi, E. A., D. R. MacGorman, C. L. Ziegler, D. Betten, M. Biggerstaff, M. Bowlan, and C. K. Potvin, 2016: An overview of the 29 May 2012 Kingfisher supercell during DC3. J. Geophys. Res. Atmos., 121, 14 31614 343, https://doi.org/10.1002/2016JD025690.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dolan, B., and S. A. Rutledge, 2010: Using CASA IP1 to diagnose kinematic and microphysical interactions in a convective storm. Mon. Wea. Rev., 138, 16131634, https://doi.org/10.1175/2009MWR3016.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fan, J., and Coauthors, 2017: Cloud-resolving model intercomparison of an MC3E squall line case: Part I—Convective updrafts. J. Geophys. Res. Atmos., 122, 93519378, https://doi.org/10.1002/2017JD026622.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Farley, R. E., 2005: BalloonAscent: 3-D simulation tool for the ascent and float of high-altitude balloons. AIAA Fifth Aviation, Technology, Integration, and Operations Conf., Reston, VA, AIAA, 15 pp.

    • Crossref
    • Export Citation
  • Foote, G. B., and J. C. Fankhauser, 1973: Airflow and moisture budget beneath a northeast Colorado hailstorm. J. Appl. Meteor., 12, 13301353, https://doi.org/10.1175/1520-0450(1973)012<1330:AAMBBA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gal-Chen, T., 1978: A method for the initialization of the anelastic equations: Implications for matching models with observations. Mon. Wea. Rev., 106, 587606, https://doi.org/10.1175/1520-0493(1978)106<0587:AMFTIO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gallice, A., F. G. Wienhold, C. R. Hoyle, F. Immler, and T. Peter, 2011: Modeling the ascent of sounding balloons: Derivation of the vertical air motion. Atmos. Meas. Tech., 4, 22352253, https://doi.org/10.5194/amt-4-2235-2011.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gao, J., M. Xue, A. Shapiro, and K. K. Droegemeier, 1999: A variational method for the analysis of three-dimensional wind fields from two Doppler radars. Mon. Wea. Rev., 127, 21282142, https://doi.org/10.1175/1520-0493(1999)127<2128:AVMFTA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Geerts, B., and Coauthors, 2018: Recommendations for in situ and remote sensing capabilities in atmospheric convection and turbulence. Bull. Amer. Meteor. Soc., 99, 24632470, https://doi.org/10.1175/BAMS-D-17-0310.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilman, D. L., F. J. Fuglister, and J. M. Mitchell, 1963: On the power spectrum of “red noise.” J. Atmos. Sci., 20, 182184, https://doi.org/10.1175/1520-0469(1963)020<0182:OTPSON>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Helmus, J. J., and S. M. Collis, 2016: The Python ARM Radar Toolkit (Py-ART), a library for working with weather radar data in the Python programming language. J. Open Res. Software, 4, e25, https://doi.org/10.5334/jors.119.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., and D. J. Musil, 1982: Case study of a hailstorm in Colorado. Part II: Particle growth processes at mid-levels deduced from in-situ measurements. J. Atmos. Sci., 39, 28472866, https://doi.org/10.1175/1520-0469(1982)039<2847:CSOAHI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • InterMet Systems, 2016: iMet-1-ABxn data sheet. Grand Rapids Michigan, 1 pp., https://www.intermetsystems.com/ee/pdf/202060_iMet-1-ABxn_Data_161006.pdf.

  • Kropfli, R. A., and L. J. Miller, 1976: Kinematic structure and flux quantities in a convective storm from dual-Doppler radar observations. J. Atmos. Sci., 33, 520529, https://doi.org/10.1175/1520-0469(1976)033<0520:KSAFQI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lehmiller, G. S., H. B. Bluestein, P. J. Neiman, F. M. Ralph, and W. F. Feltz, 2001: Wind structure in a supercell thunderstorm as measured by a UHF wind profiler. Mon. Wea. Rev., 129, 19681986, https://doi.org/10.1175/1520-0493(2001)129<1968:WSIAST>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marinescu, P. J., S. C. van den Heever, S. M. Saleeby, and S. M. Kreidenweis, 2016: The microphysical contributions to and evolution of latent heating profiles in two MC3E MCSs. J. Geophys. Res. Atmos., 121, 79137935, https://doi.org/10.1002/2016JD024762.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marinescu, P. J., E. J. T. Levin, D. Collins, S. M. Kreidenweis, and S. C. van den Heever, 2019: Quantifying aerosol size distributions and their temporal variability in the Southern Great Plains, USA. Atmos. Chem. Phys., 19, 11 98512 006, https://doi.org/10.5194/acp-19-11985-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., Y. P. Richardson, S. J. Richardson, and A. Petersson, 2018: Aboveground thermodynamic observations in convective storms from balloonborne probes acting as pseudo-Lagrangian drifters. Bull. Amer. Meteor. Soc., 99, 711724, https://doi.org/10.1175/BAMS-D-17-0204.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marshall, T. C., W. D. Rust, and M. Stolzenburg, 1995: Electrical structure and updraft speeds in thunderstorms over the southern Great Plains. J. Geophys. Res., 100, 10011015, https://doi.org/10.1029/94JD02607.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marwitz, J. D., 1972: The structure and motion of severe hailstorms. Part I: Supercell storms. J. Appl. Meteor., 11, 166179, https://doi.org/10.1175/1520-0450(1972)011<0166:TSAMOS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marwitz, J. D., 1973: Trajectories within the weak echo regions of hailstorms. J. Appl. Meteor., 12, 11741182, https://doi.org/10.1175/1520-0450(1973)012<1174:TWTWER>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marwitz, J. D., and E. X. Berry, 1971: The airflow within the weak echo region of an Alberta hailstorm. J. Appl. Meteor., 10, 487492, https://doi.org/10.1175/1520-0450(1971)010<0487:TAWTWE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, L. J., 1975: Internal airflow of a convective storm from dual-Doppler radar measurements. Pure Appl. Geophys., 113, 765785, https://doi.org/10.1007/BF01592958.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, L. J., and S. M. Fredrick, 1998: CEDRIC custom editing and display of reduced information in Cartesian space. NCAR, 130 pp., https://www.eol.ucar.edu/system/files/cedric_doc.pdf.

  • Mullendore, G. L., D. R. Durran, and J. R. Holton, 2005: Cross-tropopause tracer transport in midlatitude convection. J. Geophys. Res., 110, D06113, https://doi.org/10.1029/2004JD005059.

    • Search Google Scholar
    • Export Citation
  • Musil, D. J., A. J. Heymsfield, and P. L. Smith, 1986: Microphysical characteristics of a well-developed weak echo region in a high plains supercell thunderstorm. J. Climate Appl. Meteor., 25, 10371051, https://doi.org/10.1175/1520-0450(1986)025<1037:MCOAWD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • National Centers for Environmental Information (NCEI), 2016: NCDC storm events database (storm data). NOAA, accessed 13 February 2020, https://www.ncdc.noaa.gov/stormevents/.

  • National Centers for Environmental Information (NCEI), 2017: NCDC storm events database (storm data). NOAA, accessed 13 February 2020, https://www.ncdc.noaa.gov/stormevents/.

  • Nelson, S. P., and R. A. Brown, 1987: Error sources and accuracy of vertical velocities computed from multiple-Doppler radar measurements in deep convective storms. J. Atmos. Oceanic Technol., 4, 233238, https://doi.org/10.1175/1520-0426(1987)004<0233:ESAAOV>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ooyama, K. V., 2002: The cubic-spline transform method: Basic definitions and tests in a 1D single domain. Mon. Wea. Rev., 130, 23922415, https://doi.org/10.1175/1520-0493(2002)130<2392:TCSTMB>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Oue, M., P. Kollias, A. Shapiro, A. Tatarevic, and T. Matsui, 2019: Investigation of observational error sources in multi-Doppler-radar three-dimensional variational vertical air motion retrievals. Atmos. Meas. Tech., 12, 19992018, https://doi.org/10.5194/amt-12-1999-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Palmer, A. D. F., 1912: The Theory of Measurements. McGraw-Hill, 264 pp.

  • Peters, J. M., C. J. Nowotarski, and H. Morrison, 2019: The role of vertical wind shear in modulating maximum supercell updraft velocities. J. Atmos. Sci., 76, 31693189, https://doi.org/10.1175/JAS-D-19-0096.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Potvin, C. K., D. Betten, L. J. Wicker, K. L. Elmore, and M. I. Biggerstaff, 2012: 3DVAR versus traditional dual-Doppler wind retrievals of a simulated supercell thunderstorm. Mon. Wea. Rev., 140, 34873494, https://doi.org/10.1175/MWR-D-12-00063.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Purser, R. J., W. S. Wu, D. F. Parrish, and N. M. Roberts, 2003: Numerical aspects of the application of recursive filters to variational statistical analysis. Part I: Spatially homogeneous and isotropic Gaussian covariances. Mon. Wea. Rev., 131, 15241535, https://doi.org/10.1175//1520-0493(2003)131<1524:NAOTAO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., M. R. Kumjian, S. M. Ganson, and P. Zhang, 2013: Polarimetric radar characteristics of melting hail. Part II: Practical implications. J. Appl. Meteor. Climatol., 52, 28712886, https://doi.org/10.1175/JAMC-D-13-074.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shapiro, A., K. M. Willingham, and C. K. Potvin, 2010: Spatially variable advection correction of radar data. Part II: Test results. J. Atmos. Sci., 67, 34573470, https://doi.org/10.1175/2010JAS3466.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Söder, J., M. Gerding, A. Schneider, A. Dörnbrack, H. Wilms, J. Wagner, and F. J. Lübken, 2019: Evaluation of wake influence on high-resolution balloon-sonde measurements. Atmos. Meas. Tech., 12, 41914210, https://doi.org/10.5194/amt-12-4191-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Son, K., J. Choi, W. P. Jeon, and H. Choi, 2010: Effect of free-stream turbulence on the flow over a sphere. Phys. Fluids, 22, 045101, https://doi.org/10.1063/1.3371804.

    • Crossref
    • 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, https://doi.org/10.1175/1520-0434(2003)018<1243:CPSWSE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Varble, A., and Coauthors, 2014: Evaluation of cloud-resolving and limited area model intercomparison simulations using TWP-ICE observations: 1. Deep convective updraft properties. J. Geophys. Res., 119, 13 89113 918, https://doi.org/10.1002/2013JD021372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J., J. Bian, W. O. Brown, H. Cole, V. Grubišić, and K. Young, 2009: Vertical air motion from T-REX radiosonde and dropsonde data. J. Atmos. Oceanic Technol., 26, 928942, https://doi.org/10.1175/2008JTECHA1240.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Waugh, S., and T. J. Schuur, 2018: On the use of radiosondes in freezing precipitation. J. Atmos. Oceanic Technol., 35, 459472, https://doi.org/10.1175/JTECH-D-17-0074.1.

    • Crossref
    • 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, https://doi.org/10.1175/1520-0493(1982)110<0504:TDONSC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weisman, M. L., and J. B. Klemp, 1984: The structure and classification of numerically simulated convective storms in directionally varying wind shears. Mon. Wea. Rev., 112, 24792498, https://doi.org/10.1175/1520-0493(1984)112<2479:TSACON>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery
    Fig. 1.

    (a) Mean wbuoy during the clear still-air launches from ascending radiosondes (blue) and from descending radiosondes, after the balloons burst (red). Light blue and red lines represent one standard deviation from the mean. Data are not available for most descending radiosondes below 4.5 km MSL. (b) Normalized histogram counts from all wbuoy from ascending radiosondes shown in (a), with the vertical solid line representing the mean value (4.8 m s−1) and dashed lines representing ±1.1 m s−1 from the mean, between which 90% of the data falls. The bin width is 0.2 m s−1.

  • View in gallery
    Fig. 2.

    Radar elevation angles for both the (a),(b) CSU–CHILL and (c),(d) NEXRAD radars during dual-Doppler analysis times for the two C3LOUD-Ex cases. Black lines represent the center of the beams, while gray shading represents the vertical distance covered by the beams. The smaller numbers outside the panels represent the mean elevation angle used for the PPI scan.

  • View in gallery
    Fig. 3.

    (a),(c) Case evolution and (b),(d) environmental soundings from the (a),(b) 2017 case and (c),(d) 2016 case. The white dots represent the locations where the updraft radiosondes were launched. The blue diamonds are the radar locations, and nonoverlapping regions of the black circles indicate where dual-Doppler analyses are possible. The color shading shows radar reflectivity at 1 km AGL at the approximate time of radiosonde launch, gridded and interpolated from the available radars. The gray shading represents MLCAPE from the 2100 UTC operational simulation of High-Resolution Rapid Refresh (HRRR) model for both cases; 1000 J kg−1 is contoured in white. The skew T–logp diagrams of the environmental radiosonde launches in (b) and (d) are described in the text. Hodographs (m s−1) are inlaid and the different colors within the hodographs represent 500-m increases in altitude from the surface to 6 km AGL.

  • View in gallery
    Fig. 4.

    (a) Radiosonde wair from radiosondes that sampled the two C3LOUD-Ex supercell updrafts. (b),(c) The radiosondes’ horizontal wind speed (hspd) and horizontal wind direction (hdir), respectively. In (c), 180° represents winds coming from the south. Data are only shown from the radiosondes’ launch times through to when the radiosondes reached their maximum altitudes. The smaller dots for 2017-1 and 2016-2 represent wair adjusted for the assumption of a burst radiosonde balloon (see Fig. 5). Radiosonde data in this figure and subsequent figures are shown at 1-Hz frequency.

  • View in gallery
    Fig. 5.

    Radiosonde accelerations from each launch (blue, left axis) and radiosonde altitude (red, right axis) as a function of seconds since launch. Yellow vertical lines indicate the strongest negative accelerations, which were assumed to be coincident with the radiosonde balloon bursting.

  • View in gallery
    Fig. 6.

    (a) Radiosonde wair for the 2017-1 launch with uncertainty estimates (gray). The shading from light to dark blue represents the time evolution of the radiosonde from launch to maximum altitude. The smaller red dots take into account adjustments, assuming the radiosonde balloon burst. (b)–(m) The position of the radiosonde (black dots) within the storm at two different times during the radiosonde ascent. (top) Radar reflectivity plan views and vertical cross sections, as denoted by the gray lines in the plan views. Also shown are the plan views and cross sections of (middle) CEDRIC wair and (bottom) SAMURAI wair. The arrows represent storm-relative winds in their respective planes, and black contours indicate 10 m s−1 intervals of wair, excluding the 0 m s−1 contour.

  • View in gallery
    Fig. 7.

    (a) Radiosonde-derived wair for the 2017-2 launch with uncertainty estimates (gray). The shading from light to dark blue represents the time evolution of the radiosonde from launch to maximum altitude. (b)–(g) PPI scans of radar reflectivity that overlapped with the radiosonde within a 15 s window and within 500 m of the radiosonde’s position, as labeled in (a).

  • View in gallery
    Fig. 8.

    As in Fig. 7, but for the 2017-3 radiosonde launch.

  • View in gallery
    Fig. 9.

    As in Fig. 6, but for the radiosonde 2016-1 data. The light blue to dark blue shading in (a) represents the progression of time from launch to when the balloon likely burst.

  • View in gallery
    Fig. 10.

    As in Figs. 6 and 9, but for two later times during the progression of radiosonde 2016-1.

  • View in gallery
    Fig. 11.

    As in Figs. 7 and 8, but for radiosonde 2016-2. The smaller red dots take into account adjustments assuming that the radiosonde balloon burst.

  • View in gallery
    Fig. 12.

    Comparison of radiosonde and dual-Doppler wair for radiosondes (a) 2017-1 and (b) 2016-1, as described in the text. The gray range for the radiosonde data represents the quantified uncertainty in wair. The green and blue dots represent the dual-Doppler analyses interpolated to the radiosonde position. The green and blue horizontal lines represent the range of values within 1 km in the horizontal direction of the radiosonde position within the dual-Doppler analyses.

  • View in gallery
    Fig. A1.

    Mean power spectra for the 13 clear-air (thin gray lines) and 5 updraft (thin red and blue lines) radiosonde launches for (a) wair, (b) hspd, and (c) hdir. The data chunk that was used was 180 s, and the number of chunks that went into each radiosonde launch is shown in the legend. For the 13 clear-air launches, the number of chunks varied from 18 to 49. The thick solid black line represents the mean power spectra for all the data, while the thick solid gray and red lines represent the means of the clear-air and updraft launches, respectively. Estimates of the red noise spectra are also shown as thick dashed lines.

  • View in gallery
    Fig. B1.

    Terminal velocity calculations (m s−1) for ascending C3LOUD-Ex radiosondes with varying drag coefficients and densities (shaded, left axis) and the range (maximum minus minimum) of terminal velocities for each density (right axis).

All Time Past Year Past 30 Days
Abstract Views 377 0 0
Full Text Views 624 400 23
PDF Downloads 525 310 17

Updraft Vertical Velocity Observations and Uncertainties in High Plains Supercells Using Radiosondes and Radars

Peter J. Marinescu Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Peter J. Marinescu in
Current site
Google Scholar
PubMed
Close
,
Patrick C. Kennedy Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Patrick C. Kennedy in
Current site
Google Scholar
PubMed
Close
,
Michael M. Bell Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Michael M. Bell in
Current site
Google Scholar
PubMed
Close
,
Aryeh J. Drager Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Aryeh J. Drager in
Current site
Google Scholar
PubMed
Close
,
Leah D. Grant Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Leah D. Grant in
Current site
Google Scholar
PubMed
Close
,
Sean W. Freeman Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Sean W. Freeman in
Current site
Google Scholar
PubMed
Close
, and
Susan C. van den Heever Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado

Search for other papers by Susan C. van den Heever in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Observations of the air vertical velocities (wair) in supercell updrafts are presented, including uncertainty estimates, from radiosonde GPS measurements in two supercells. These in situ observations were collected during the Colorado State University Convective Cloud Outflows and Updrafts Experiment (C3LOUD-Ex) in moderately unstable environments in Colorado and Wyoming. Based on the radiosonde accelerations, instances when the radiosonde balloon likely bursts within the updraft are determined, and adjustments are made to account for the subsequent reduction in radiosonde buoyancy. Before and after these adjustments, the maximum estimated wair values are 36.2 and 49.9 m s−1, respectively. Radar data are used to contextualize the in situ observations and suggest that most of the radiosonde observations were located several kilometers away from the most intense vertical motions. Therefore, the radiosonde-based wair values presented likely underestimate the maximum values within these storms due to these sampling biases, as well as the impacts from hydrometeors, which are not accounted for. When possible, radiosonde-based wair values were compared to estimates from dual-Doppler methods and from parcel theory. When the radiosondes observed their highest wair values, dual-Doppler methods generally produced 15–20 m s−1 lower wair for the same location, which could be related to the differences in the observing systems’ resolutions. In situ observations within supercell updrafts, which have been limited in recent decades, can be used to improve our understanding and modeling of storm dynamics. This study provides new in situ observations, as well as methods and lessons that could be applied to future field campaigns.

© 2020 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Peter J. Marinescu, peter.marinescu@colostate.edu

Abstract

Observations of the air vertical velocities (wair) in supercell updrafts are presented, including uncertainty estimates, from radiosonde GPS measurements in two supercells. These in situ observations were collected during the Colorado State University Convective Cloud Outflows and Updrafts Experiment (C3LOUD-Ex) in moderately unstable environments in Colorado and Wyoming. Based on the radiosonde accelerations, instances when the radiosonde balloon likely bursts within the updraft are determined, and adjustments are made to account for the subsequent reduction in radiosonde buoyancy. Before and after these adjustments, the maximum estimated wair values are 36.2 and 49.9 m s−1, respectively. Radar data are used to contextualize the in situ observations and suggest that most of the radiosonde observations were located several kilometers away from the most intense vertical motions. Therefore, the radiosonde-based wair values presented likely underestimate the maximum values within these storms due to these sampling biases, as well as the impacts from hydrometeors, which are not accounted for. When possible, radiosonde-based wair values were compared to estimates from dual-Doppler methods and from parcel theory. When the radiosondes observed their highest wair values, dual-Doppler methods generally produced 15–20 m s−1 lower wair for the same location, which could be related to the differences in the observing systems’ resolutions. In situ observations within supercell updrafts, which have been limited in recent decades, can be used to improve our understanding and modeling of storm dynamics. This study provides new in situ observations, as well as methods and lessons that could be applied to future field campaigns.

© 2020 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Peter J. Marinescu, peter.marinescu@colostate.edu

1. Introduction

Supercell updrafts contain some of the most intense vertical air velocities (wair) in the atmosphere (e.g., Musil et al. 1986; Lehmiller et al. 2001; DiGangi et al. 2016). The magnitude and vertical structure of wair within supercell updrafts control many atmospheric processes, including the production of severe hail (e.g., Browning and Foote 1976; Heymsfield and Musil 1982) and the transport of atmospheric constituents from the boundary layer to the upper troposphere and stratosphere (e.g., Foote and Fankhauser 1973; Mullendore et al. 2005). Due to the strong vertical velocities in supercell updrafts, cloud droplets do not have enough time to grow to sizes that can be observed by most radars. Supercell updrafts can therefore be clearly identified in radar data as regions with lower reflectivity in the lower- and middle-tropospheric levels, laterally and vertically bounded by higher reflectivity, known initially as vaults and later as weak-echo regions or bounded weak-echo regions (WERs or BWERs; Browning and Ludlam 1962; Chisholm 1970; Marwitz and Berry 1971; Chisholm 1973). Despite supercell updrafts’ importance for atmospheric processes, these updrafts have seldom been observed in situ.

The first of these infrequent in situ observations of the magnitudes of supercell updraft velocities came from armored aircraft penetrations through the WERs (Marwitz and Berry 1971; Heymsfield and Musil 1982). These observations were usually made near cloud base and in the inflow air ahead of the supercell and were typically taken in the High Plains of the United States and Canada. These initial in situ observations generally resulted in estimates of wair in the 15–30 m s−1 range. One research flight into the WER of a supercell in Montana at ~7 km above mean sea level (MSL) observed wair as high as 50 ± 5 m s−1 (Musil et al. 1986). Despite the continued need for in situ observations of deep convection, the last U.S. storm-penetrating research aircraft was retired without replacement in 2005 (Geerts et al. 2018).

In situ estimates of updraft velocities can also be achieved via releasing sensors or trackable objects into supercell updrafts from the storm’s proximity. Chaff packets have been released from aircraft at thunderstorms’ cloud bases and tracked with radar to estimate vertical velocities within supercells. Results from this approach have generally been consistent with those from in situ aircraft penetrations (Marwitz 1972, 1973). Radiosondes have also been used throughout the past 50 years, albeit infrequently, to estimate the vertical velocities in supercells (Barnes 1970; Davies-Jones 1974; Davies-Jones and Henderson 1975; Bluestein et al. 1988, 1989; Marshall et al. 1995; Markowski et al. 2018). From these radiosonde observations, the greatest reported wair values were 49 m s−1 (Bluestein et al. 1988) and 53 m s−1 (Markowski et al. 2018), which occurred in Texas and Oklahoma, respectively.

Due to the challenges associated with in situ observations of updrafts, such as the hazardous sampling conditions and the difficulty of placing sensors directly within the updraft core, remotely sensed observations have replaced in situ observations as the primary estimates of wair in deep convection in recent decades. The most common method for estimating wair with remote sensing utilizes data from multiple Doppler radars to determine the horizontal components of the wind, and then invokes the mass continuity equation to calculate the vertical component of the wind (e.g., Armijo 1969; Miller 1975; Kropfli and Miller 1976; Gal-Chen 1978). Multi-Doppler retrievals can provide vertical velocities over a relatively large domain and are often conveniently gridded to Cartesian coordinates. However, multi-Doppler estimates also have hard-to-characterize uncertainties due to their sensitivities to analysis specifications, such as how the data are filtered or interpolated (e.g., Nelson and Brown 1987; Miller and Fredrick 1998; Collis et al. 2010; Shapiro et al. 2010) or the temporal and spatial resolution of the data (e.g., Bousquet et al. 2008; Potvin et al. 2012; Oue et al. 2019; Dahl et al. 2019). Because of their availability, these remotely sensed observations have often been used to validate case study model simulations of deep convection in large field campaigns (Varble et al. 2014; Marinescu et al. 2016; Fan et al. 2017). These studies have shown that cloud-resolving models tend to produce stronger vertical velocities than their corresponding radar-derived estimates. However, the errors associated with multi-Doppler wair are largely case-specific and depend on the radar scanning strategy, the type of convection and location of convection with respect to the radars (Oue et al. 2019). Therefore, it is still challenging to attribute the differences in updraft magnitudes from radar-based analyses and cloud-resolving models. In situ observations can thus assist in providing independent estimates of wair.

In this study, we present GPS-radiosonde-based in situ observations and uncertainties of wair within the updraft regions of two supercells. These observations were made during the Colorado State University Convective Cloud Outflows and Updrafts Experiment (C3LOUD-Ex) during 2016 and 2017 in the High Plains of Colorado, Wyoming, and Nebraska (van den Heever et al. 2020, manuscript submitted to Bull. Amer. Meteor. Soc.). Using the radiosonde data, along with radar observations within the C3LOUD-Ex domain, we 1) provide our best in situ estimates of wair within the two supercell updrafts, 2) contextualize and compare these observations to other available wair estimates for the two cases, and 3) offer insights for future efforts toward obtaining in situ observations within supercell updrafts.

2. C3LOUD-ex observations

a. Radiosondes

During C3LOUD-Ex, the iMet-1-ABxn radiosonde was used, which included a pressure, temperature, and humidity sensor, as well as a GPS receiver (InterMet Systems 2016). The radiosonde package was attached via a dereeler (30-m length) to a 200-g balloon that was filled with enough helium to reduce the helium tank’s gauge pressure by approximately 3447 kPa (500 psi). For this study, the most essential radiosonde data were from the GPS receiver, which has a horizontal position accuracy of 10 m and an altitude accuracy of 15 m. GPS positions were received from the radiosonde at a rate of approximately 1 Hz and linearly interpolated to create a 1-Hz record.

Using the GPS altitude data, the vertical velocity of the radiosonde was estimated every second using a centered-in-time derivative:
wsonde=ΔzΔt,
where wsonde is the representative vertical velocity of the radiosonde system over the time interval Δt, and Δz is the vertical distance traveled by the radiosonde during Δt. For this study, Δt is chosen to be 12 s, which for 10–60 m s−1 updrafts equates to vertical distances of 120–720 m, comparable to current numerical model simulation grid spacings and/or observational grids. This Δt is chosen in order to reduce the periodic signals that were present in this dataset on the time scales of 12 s and less, as described in more detail in appendix A. These periodic signals were likely associated with pendulum motions, which are theoretically estimated to have periods between 11 and 12 s for a dereeler length of 30 m. The periodic signals could also be associated with other self-induced balloon motions (e.g., Wang et al. 2009; Söder et al. 2019) that can occur on these small time scales. The error in this wsonde, denoted ϵw,sonde, was calculated using error propagation methods (e.g., Palmer 1912). Because the relative error in the GPS time measurement was several orders of magnitude smaller than the error in GPS position measurement ϵw,GPS can be simplified to the following:
ϵw,sonde=|wsonde|(2ϵzΔz),
where ϵz is the error in the GPS altitude from the radiosonde (15 m). For a fixed Δt = 12 s and due to the linear relationship between wsonde and Δz, ϵw,sonde is always ±1.8 m s−1. For the cases presented in this study, each increase of 2 s in Δt, for Δt between 8 and 16 s, reduces the maximum vertical velocity observed by on average 0.1–0.5 m s−1 due to smoothing and decreases the uncertainty by ±0.1–0.4 m s−1. Therefore, the results are minimally impacted by the choice of Δt.
While wsonde was directly observed by the radiosonde, the vertical velocity of the air that the radiosonde sampled (wair) was desired. We decompose wsonde into the following components:
wsonde=wair+wbuoy+wupd-drag+wupd-hydro,
where wbuoy is the vertical velocity arising from the buoyancy of the radiosonde system (balloon and radiosonde) in clear-sky, still-air conditions; wupd-drag is the vertical velocity associated with changes to the drag force on the radiosonde system within an updraft as compared to clear, still air; and wupd-hydro is the forcing from hydrometeors impacting or accumulating on the radiosonde system. Ultimately, by observing wsonde, whose uncertainty (ϵw,sonde) is known, and estimating wbuoy, wupd-drag, wupd-hydro, and their associated uncertainties (ϵw,buoy, ϵw,upd-drag, ϵw,upd-hydro), an estimate of wair and its uncertainty (ϵw,air) can be determined.

Implicit in these definitions is that in clear-sky, still-air conditions wair, wupd-drag, and wupd-hydro are all ~0 m s−1 and hence, wsonde = wbuoy. Therefore, we estimated wbuoy from the wsonde measurements obtained from thirteen radiosondes that were launched at the Colorado State University Foothills Campus in clear conditions with weak vertical motions throughout the troposphere. These radiosondes were launched during synoptic-scale ridges, which provided weak subsidence throughout the region. Seven launches took place overnight to minimize the influence of boundary layer vertical motions, as well as to eliminate the impacts of solar radiation on the balloon, which could affect the buoyancy of the radiosonde system (Farley 2005). Vertical profiles of wbuoy for the clear-sky, still-air launches are shown in Fig. 1a. The radiosonde descent rates (red), which occur after the radiosondes’ balloons burst, vary with altitude and have a greater spread than the ascent rates (blue), which are approximately constant throughout the troposphere and lower stratosphere. Figure 1b shows a normalized histogram of the ascent rates from the rising radiosondes. The mean upward vertical velocity from these experiments is 4.8 m s−1 (wbuoy), with 90% of the data falling within ±1.1 m s−1, which we define here as ϵw,buoy. We also estimated the wbuoy following the theoretical basis from Wang et al. (2009) and using a height-invariant drag coefficient of 0.5 for seven clear-air launches in which the free-lift weights were directly measured prior to launch. The theoretical wbuoy varied with height, increasing from ~4.1–4.8 m s−1 near the surface to ~5.0–6.0 m s−1 at ~13 km MSL. These theoretical values overlap with the height-invariant estimate of wbuoy obtained from observed wsonde from the clear-air launches (4.8 ± 1.1 m s−1).

Fig. 1.
Fig. 1.

(a) Mean wbuoy during the clear still-air launches from ascending radiosondes (blue) and from descending radiosondes, after the balloons burst (red). Light blue and red lines represent one standard deviation from the mean. Data are not available for most descending radiosondes below 4.5 km MSL. (b) Normalized histogram counts from all wbuoy from ascending radiosondes shown in (a), with the vertical solid line representing the mean value (4.8 m s−1) and dashed lines representing ±1.1 m s−1 from the mean, between which 90% of the data falls. The bin width is 0.2 m s−1.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

It is unknown whether and how the drag force on the radiosonde system within supercell updrafts differs from that in clear air, and we therefore assume that the wupd-drag is 0 m s−1 (i.e., no systematic shifts in the radiosonde-based wair due to different drag forces within the updraft). Using the relationship between terminal velocity and the drag coefficient, however, we estimate that the uncertainty associated with variable drag forces on the radiosonde system within updraft conditions (ϵw,upd-drag) is ±1.6 m s−1 (see appendix B).

The forcing from hydrometeor impacts (wupd-hydro) will typically be downward and can be caused by collisions with or accumulation of condensate mass (e.g., riming) on the radiosonde system. Because of the uncertainties in quantifying the presence and magnitude of these processes from the data available during C3LOUD-Ex, we did not attempt to estimate wupd-hydro or its uncertainty in this study. Therefore, the radiosonde wair is expected to be most accurate in scenarios where there is little to no impact from hydrometeors on the radiosonde system (i.e., outside of regions with hydrometeors). In such situations, the radiosonde wair has an uncertainty (ϵw,air) of ±2.6 m s−1, where ϵw,air is the summation in quadrature of ϵw,sonde (±1.8 m s−1), ϵw,buoy (±1.1 m s−1), and ϵw,upd-drag (±1.6 m s−1), following error propagation methods. In regions with hydrometeors, such as the cloudy regions of the supercell updraft, however, since wupd-hydro is negative for a rising balloon, the radiosonde wair represents a lower bound on the actual wair. It is important to note here that these estimates also assume that the balloon has not burst. Using the radiosonde accelerations and the radar observations (as described in section 4), we estimated the times at which the balloons burst and made corresponding adjustments for those situations to provide a more realistic estimate of wair.

b. Radars

Because the radiosondes provided localized measurements within the broad supercell updrafts, we used radar data to contextualize the in situ observations. Additionally, the radar data provided an independent estimate of wair using dual-Doppler methods. Three radars were primarily utilized during C3LOUD-Ex: the CSU–CHILL radar (Brunkow et al. 2000), located in Greeley, Colorado; the Cheyenne, Wyoming NEXRAD (KCYS); and the Denver, Colorado NEXRAD (KFTG). KCYS is located ~79 km to the north of CSU–CHILL, and CSU–CHILL is located ~74 km to the north of KFTG. Plan position indicator (PPI) scans from all radars, as well as additional range height indicator (RHI) scans from CSU–CHILL, provided detailed views of the storm structure and the relative position of the radiosonde within the storms. During C3LOUD-Ex, the NEXRAD radars (KCYS and KFTG) had prescribed volume coverage patterns (VCP212) that each lasted ~5 min, while the CSU–CHILL radar was manually operated and synchronized with the relevant NEXRAD radar during updraft-targeted radiosonde launches. Figure 2 shows an example of radar elevation angles for the NEXRAD and CSU–CHILL radars for one radar volume for the two cases examined in this study.

Fig. 2.
Fig. 2.

Radar elevation angles for both the (a),(b) CSU–CHILL and (c),(d) NEXRAD radars during dual-Doppler analysis times for the two C3LOUD-Ex cases. Black lines represent the center of the beams, while gray shading represents the vertical distance covered by the beams. The smaller numbers outside the panels represent the mean elevation angle used for the PPI scan.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

Reflectivity, velocity, and some dual-polarization data from all three radars were used. These radar data were first quality-controlled using the dual-polarization data. Specifically, we excluded all radar gates where the standard deviation of the differential propagation phase was greater than 21° over a range of 11 gates. We found that this threshold eliminated noise and ground clutter, while retaining more data near features of interest (e.g., the WER), which were otherwise eliminated when using correlation coefficient as a threshold. The radar velocity data were dealiased using the region-based method in the Python-ARM Radar Toolkit (Py-ART; Helmus and Collis 2016), and the storm motion for both cases was estimated for each 5-min radar volume scan using the Py-ART grid displacement algorithm on the radar reflectivity between 3 and 8 km AGL. These estimated storm motions were calculated for each radar volume and used for corrections related to storm translation in the dual-Doppler analyses, as well as for advecting the radar analyses in time for comparisons with the 1-Hz radiosonde data. Although these processing steps were largely automated, all quality-controlled and processed data were also manually checked.

Two analysis programs were then used to synthesize the radial velocity data and produce radar-based wair estimates. These programs were the Custom Editing and Display of Reduced Information in Cartesian space (CEDRIC; Miller and Fredrick 1998) and the Spline Analysis at Mesoscale Utilizing Radar and Aircraft Instrumentation (SAMURAI; Bell et al. 2012). While these programs both solve the basic radar equations, CEDRIC uses column-by-column vertical integration of the mass continuity equation to produce local solutions for each vertical column, while SAMURAI uses a 3D-variational approach (Gao et al. 1999) and produces a global solution for the entire analysis domain via a cost minimization function. The 3D-variational approach has been shown to produce better vertical velocity solutions for a supercell case than other methods (Potvin et al. 2012). These analyses were completed on 1-km and 500-m Cartesian grids for the 26 May 2017 and 17 July 2016 cases, respectively, due to the relative locations of each storm with respect to the radars as shown in the following section. The top boundaries in the analyses were set to 17 km MSL (5–6 km above the tropopause) and the vertical velocities were set to 0 at the top boundary in SAMURAI and at half a vertical grid level above the highest level where divergence was calculated in each column in CEDRIC. For the CEDRIC analyses shown here, the variational vertical integration method was used, whereby downward integration was first completed, residual errors were spread throughout the column in an iterative manner and last, variationally adjusted integration was applied (e.g., Wvar in Dolan and Rutledge 2010). A linear, least squares two-dimensional filter was also used on the horizontal winds in the CEDRIC analyses (Miller and Fredrick 1998). Low-pass filters with approximate scales of 4 and 2 km for the 1-km and 500-m Cartesian grids, respectively, were applied in the SAMURAI analyses (Ooyama 2002, Purser et al. 2003).

3. C3LOUD-ex cases

During C3LOUD-Ex, there were seven cases in which the updrafts of supercell storms were successfully sampled with radiosondes (van den Heever et al. 2020, manuscript submitted to Bull. Amer. Meteor. Soc.). In this study, we focus on the two cases that had successful radiosonde sampling of updrafts within the regions where dual-Doppler estimates of wair could also be made. These occurred on 26 May 2017 and 17 July 2016 and are briefly described in the following two sections and summarized in Fig. 3.

Fig. 3.
Fig. 3.

(a),(c) Case evolution and (b),(d) environmental soundings from the (a),(b) 2017 case and (c),(d) 2016 case. The white dots represent the locations where the updraft radiosondes were launched. The blue diamonds are the radar locations, and nonoverlapping regions of the black circles indicate where dual-Doppler analyses are possible. The color shading shows radar reflectivity at 1 km AGL at the approximate time of radiosonde launch, gridded and interpolated from the available radars. The gray shading represents MLCAPE from the 2100 UTC operational simulation of High-Resolution Rapid Refresh (HRRR) model for both cases; 1000 J kg−1 is contoured in white. The skew T–logp diagrams of the environmental radiosonde launches in (b) and (d) are described in the text. Hodographs (m s−1) are inlaid and the different colors within the hodographs represent 500-m increases in altitude from the surface to 6 km AGL.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

a. 26 May 2017 case study

At 1815 UTC (UTC = local time + 6 h), an environmental sounding (Fig. 3b) was launched at 39.72°N, 104.22°W and showed 0–6-km shear of 26 m s−1, mixed-layer (0–90 hPa AGL) convective available potential energy (MLCAPE) of 491 J kg−1, and surface-based CAPE of 1882 J kg−1.1 By 2000 UTC, terrain-induced scattered convection was moving eastward over the Denver metropolitan region. The destabilized boundary layer and favorable environmental conditions resulted in the development of an isolated supercell by 2200 UTC, located within the dual-Doppler analysis region for the CSU–CHILL and KFTG radars (Fig. 3a). At 2158 UTC, a radiosonde (2017-1) was launched and sampled the updraft of the developing supercell, while 1.5 in. (3.8 cm) diameter hail was reported at the surface nearby (NCEI 2017). Around 2200 UTC, the storm propagation slowed and took a rightward turn toward the east-southeast. Over the next several hours, many instances of hail with diameters of 1–1.5 in. (2.5–3.8 cm) were reported at the ground along the storm’s path, as were 2 weak tornadoes (NCEI 2017). Two additional radiosondes (2017-2 and 2017-3) sampled the supercell updraft between 2200 and 2400 UTC. This long-lived supercell continued into Kansas, outside of the C3LOUD-Ex domain, and subsequently became part of a mesoscale convective system.

b. 17 July 2016 case study

On 17 July 2016 at ~2030 UTC, convection that had initiated over the high terrain of southern Wyoming moved eastward onto the high plains to the northwest of Cheyenne, Wyoming, where it quickly organized into a supercell and subsequently turned toward the southeast (Fig. 3c). Earlier in the day, between 1800 and 1900 UTC, three radiosondes were launched (at 40.67°N, 104.33°W; 41.22°N, 104.35°W; and 41.24°N, 103.70°W) to better capture the environment ahead of this storm. These observations (Fig. 3d) indicate MLCAPE of ~950–1200 J kg−1 and 0–6-km shear of 21–25 m s−1. This supercell propagated southeastward across the C3LOUD-Ex domain, including through the region where dual-Doppler analyses could be conducted using the CSU–CHILL and KCYS radars. This storm had more intense radar reflectivity than did the 2017 case, and there were several reports of 2.0-in. (5.1-cm) diameter hail as well as a few baseball-sized hailstones (diameters of ~7.5 cm; NCEI 2016). As the supercell propagated southeastward, two radiosondes were launched into the supercell’s main updraft region (Fig. 3c). The first, 2016-1, was located within the dual-Doppler analysis region, while the second, 2016-2, was just outside the dual-Doppler lobes in a more unstable environment. By 0130 UTC 18 July 2016, the storm began to lose many of its supercellular characteristics, and it dissipated by 0300 UTC.

We note here that both of these High Plains supercells experienced environments with substantial vertical wind shear (0–6 km; ~21–26 m s−1) and moderate MLCAPE (~1000–1600 J kg−1). These environments had bulk Richardson numbers of ~10–15, well within the range favorable for supercells (Weisman and Klemp 1982), although the MLCAPE values are on the lower end of those conditions supporting weakly tornadic and nontornadic supercells within the broader United States (Thompson et al. 2003). Therefore, these C3LOUD-Ex observations of wair will likely be lower than similar observations of supercells in more unstable air masses, such as those present in the U.S. southern Great Plains.

4. Radiosonde-derived updraft wair

The wair estimated from the five radiosondes that sampled the two supercells’ updrafts are shown in Fig. 4, which for simplicity’s sake only depicts wair from when the radiosonde was launched to when the radiosonde reached its maximum altitude. The horizontal wind speed and direction were also calculated from the raw GPS data every second, using the same Δt of 12 s. These radiosonde data represent point locations within the large supercell updrafts. Despite the radar’s inability to observe the finer-scale motions observed by the radiosondes, the radar data were useful for determining the position of the radiosonde within the updraft and elucidating whether each radiosonde was in the vicinity of the strongest wair within these storms. The radiosondes took many different trajectories throughout the supercells. Only one of these five radiosondes (2017–2) continued to rise into the stratosphere after sampling the supercell updraft. The other radiosonde systems likely experienced conditions within the updraft that robbed them of their positive buoyancy (e.g., radiosonde balloon bursting or significant riming). To identify these events, the radiosonde-derived accelerations were calculated from the difference in the 1-Hz wair data and were examined for the entirety of the radiosondes’ data transmissions (Fig. 5). A 5-s moving average was applied to the calculated acceleration to eliminate noise but still capture significant events. The most intense negative accelerations were assumed to be associated with the radiosonde balloon bursting, whereby wbuoy instantaneously changed from approximately +4.8 m s−1 to anywhere between −15 and −25 m s−1, depending on the radiosonde’s tropospheric altitude (Fig. 1a). It is possible that such intense negative accelerations could also be a result of significant icing and hydrometeor collisions with the balloon, although these effects are unquantifiable in our dataset. The most intense negative accelerations are highlighted in yellow in Fig. 5 and had values between −1.4 and −4.3 m s−2. These values were similar to those associated with the radiosonde balloon bursts during the clear-sky, still-air launches, which all occurred above 16 km MSL and ranged from −2.2 to −4.8 m s−2 (not shown). For radiosondes 2017-2 and 2016-1, the radiosondes’ balloons did not burst until right before their final descents to the surface. However, for radiosondes 2017-1 and 2016-2, it appears that the balloon burst within the radiosonde’s initial ascent through the updraft, based on the assumption that the most negative accelerations represent balloon bursts. While the likelihood of this relationship has yet to be established, radar data (shown in the following sections) suggest that the radiosondes 2017-1 and 2016-2 were entering regions of large hail within intense updrafts during their most negative accelerations, conditions that can cause a radiosonde balloon to burst. If the balloons had not burst, we would have expected these radiosondes to eventually exit the storm and rise into the stratosphere. Therefore, for radiosondes 2017-1 and 2016-2, adjustments were made to the radiosonde-derived wair after the assumed balloon burst event (Fig. 4, smaller dots), taking into account the altitude-dependent, mean terminal velocities of the descending radiosonde system (Fig. 1a). For radiosonde 2017-3, communication was lost with the radiosonde during its ascent within the updraft before any significant negative accelerations occurred, and therefore, no adjustments were necessary.

Fig. 4.
Fig. 4.

(a) Radiosonde wair from radiosondes that sampled the two C3LOUD-Ex supercell updrafts. (b),(c) The radiosondes’ horizontal wind speed (hspd) and horizontal wind direction (hdir), respectively. In (c), 180° represents winds coming from the south. Data are only shown from the radiosondes’ launch times through to when the radiosondes reached their maximum altitudes. The smaller dots for 2017-1 and 2016-2 represent wair adjusted for the assumption of a burst radiosonde balloon (see Fig. 5). Radiosonde data in this figure and subsequent figures are shown at 1-Hz frequency.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

Fig. 5.
Fig. 5.

Radiosonde accelerations from each launch (blue, left axis) and radiosonde altitude (red, right axis) as a function of seconds since launch. Yellow vertical lines indicate the strongest negative accelerations, which were assumed to be coincident with the radiosonde balloon bursting.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

Before the adjustments described above, the maximum wair values measured by the radiosondes for the 2017 and 2016 cases were 36.2 and 25.5 m s−1, respectively. After adjusting for the balloon burst assumption for the 2017-1 and 2016-2 radiosondes, the respective maximum radiosonde wair values were 45.8 and 49.9 m s−1. Here, it is important to emphasize that there is larger uncertainty in wbuoy after the balloon burst, in part due to the larger spread of descent rates that are used for the adjustments and that are based on the clear-air radiosondes after their balloons burst (Fig. 1a). However, making this adjustment provides a more realistic estimate of wair, assuming the radiosonde balloon does burst. Additional testing would be needed to quantify the uncertainties for these adjusted wair estimates. In the next sections, we present the radiosonde wair for each launch in the context of the radar data.

a. 2017 case

Radiosonde 2017-1 was launched at 2158 UTC, shortly after the supercell formed and within the dual-Doppler analysis region for the CSU–CHILL and KFTG radars. Figure 6 depicts the radiosonde wair along with two snapshots of the radiosonde position within the storm based on the radar reflectivity and dual-Doppler-derived wair. Based on the radiosonde humidity data, the radiosonde entered cloud around 2.7 km MSL, at which point wair, the updraft vertical velocity, was 5.5 m s−1. This corresponds to an average rate of acceleration from the ground level to cloud base of 0.034 m s−2. The radiosonde continued to accelerate within the cloudy updraft through ~7.5 km MSL at an average rate of 0.116 m s−2, more than triple the rate below cloud base, and the horizontal winds decreased from ~10 to ~1 m s−1 and shifted from southerly to northerly (Figs. 4b,c).

Fig. 6.
Fig. 6.

(a) Radiosonde wair for the 2017-1 launch with uncertainty estimates (gray). The shading from light to dark blue represents the time evolution of the radiosonde from launch to maximum altitude. The smaller red dots take into account adjustments, assuming the radiosonde balloon burst. (b)–(m) The position of the radiosonde (black dots) within the storm at two different times during the radiosonde ascent. (top) Radar reflectivity plan views and vertical cross sections, as denoted by the gray lines in the plan views. Also shown are the plan views and cross sections of (middle) CEDRIC wair and (bottom) SAMURAI wair. The arrows represent storm-relative winds in their respective planes, and black contours indicate 10 m s−1 intervals of wair, excluding the 0 m s−1 contour.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

During this time period, the radiosonde was located within the main updraft, along the western part of the weak echo region. At 7.5 km MSL (Figs. 6b–g), the radiosonde decelerated for ~15–20 s as it entered a region of higher reflectivity (>50 dbZ) and low correlation coefficients (<0.9, not shown), suggesting large hail (e.g., Balakrishnan and Zrnić 1990; Ryzhkov et al. 2013). Although the dual-Doppler analyses do not resolve the winds on the scales observed by the radiosonde, the fact that both dual-Doppler analyses (Figs. 6e,g) depict increasing wair with height (which would suggest positive balloon acceleration as opposed to deceleration) supports our hypothesis that the balloon burst. As such, above 7.5 km MSL, adjustments were made to the wair estimates using a wbuoy corresponding to a burst radiosonde balloon, as described in the prior section. At 9.7 km MSL (Figs. 6h–m), the wair after adjustments reached its peak value (45.8 m s−1). At this time, the radiosonde was within the primary updraft region but was nevertheless located ~5 km to the southwest of the most intense radar-derived updrafts (Figs. 6j,l), suggesting that the maximum wair in this storm was likely even higher than that estimated from the radiosonde. We note that the adjusted radiosonde wair values are more intense than those from the radar analyses, and the wair estimates from the different observing platforms are compared in section 5a. The radiosonde reached its maximum altitude of 10.6 km MSL after experiencing north-northwesterly winds for 2–3 min, which advected the radiosonde to the southern periphery of the updraft, where the wair was no longer strong enough to suspend the radiosonde system.

Approximately 1 h later (2251 UTC), another radiosonde (2017-2) was launched into the supercell updraft. Although the supercell was no longer within the region where dual-Doppler estimates could be made, both radar RHIs (not shown) and PPIs were used to contextualize the radiosonde measurements. Figure 7 shows PPI snapshots throughout the radiosonde trajectory at times when the radiosonde location was simultaneously sampled by one of the radars. The 2017-2 radiosonde was launched to the southwest of the WER (Fig. 7b), was advected northward in the inflow, and observed wair of ~14 m s−1 before entering the cloud at 3.7 km MSL, which was above cloud base. A maximum wair of 36.2 m s−1 was obtained at approximately 10.4 km MSL (Fig. 7e). Despite observing strong wair throughout its trajectory, the radiosonde was consistently located ~5–10 km to the southwest of where the strongest wair was likely located: the WER in the lower and middle troposphere (Figs. 7c,d) and the higher reflectivity regions in the upper troposphere (Fig. 7e). After reaching the top of the storm, the radiosonde underwent negative acceleration and sampled a minimum wair of −26.1 m s−1, which was likely associated with strong downdrafts south of the main updraft (Fig. 7f). Unlike the 2017-1 radiosonde, 2017-2 eventually exited the storm (Fig. 7g) and rose to an altitude of 22.2 km MSL before the radiosonde balloon burst.

Fig. 7.
Fig. 7.

(a) Radiosonde-derived wair for the 2017-2 launch with uncertainty estimates (gray). The shading from light to dark blue represents the time evolution of the radiosonde from launch to maximum altitude. (b)–(g) PPI scans of radar reflectivity that overlapped with the radiosonde within a 15 s window and within 500 m of the radiosonde’s position, as labeled in (a).

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

At 2359 UTC, a third radiosonde (2017-3; Fig. 8) was launched and subsequently sampled the WER in the middle troposphere (Figs. 8b,c). This radiosonde experienced the strongest vertical velocities between the surface and 6.8 km MSL of all three radiosondes from this case, accelerating at an average rate of 0.113 m s−2 from 4.3 m s−1 at 2 km MSL to a maximum wair of 31.1 m s−1 at 7.1 km MSL. Unfortunately, the thermodynamic sensors were compromised during the radiosonde launch, and thus it is unclear at exactly which point the radiosonde entered cloudy conditions. Above 7.1 km MSL, the radiosonde began to decelerate and likely encountered rain and/or hail (Figs. 8a,d); communication with the radiosonde was lost at 10.8 km MSL.

Fig. 8.
Fig. 8.

As in Fig. 7, but for the 2017-3 radiosonde launch.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

b. 2016 case

Similar analyses were conducted for radiosondes 2016-1 and 2016-2 for the isolated supercell that occurred on 17 July 2016. Because the supercell passed closer to the radar network (Fig. 3c), the dual-Doppler analyses were conducted with 500-m grid spacing, which allowed for a more detailed structure in the wair values, although the analyses were still unable to resolve the finer-scale motions observed by the radiosondes.

At 2224 UTC, the 2016-1 radiosonde (Figs. 9, 10) was launched on the southern side of the supercell, shortly after the cold pool associated with the rear flank downdraft passed the launch location, resulting in negative-to-neutral wair and northwesterly winds near the surface (Figs. 9a, 5). A radiosonde-based wair of ~23 m s−1 was observed twice during the radiosonde’s ascent through the storm (at 6.8 and 9.1 km MSL; Fig. 9). In both instances, the radiosonde was in the extreme southwest edge of the updraft region, and ~10 km to the west of the WER (Figs. 9b,c,h,i). The radiosonde continued to rise above 12 km MSL and then underwent a 2.5-km descent, during which it observed a minimum wair of −26.8 m s−1 (Figs. 10b–g). This radiosonde, however, experienced its most intense negative acceleration immediately before the radiosonde’s final descent to the surface (Fig. 5d), and therefore, we propose that this first radiosonde descent was associated with nearby, strong upper-level downdrafts that were diagnosed by both the SAMURAI analysis (Figs. 10f,g) and, to a lesser extent, the CEDRIC analysis (Figs. 10d,e) rather than with the balloon bursting. The radiosonde then experienced several vertical oscillations, ascending and descending three times around 10–11 km MSL and ~15 km to the southeast of the main updraft (Figs. 10h–m). These oscillations were likely associated with gravity waves in the anvil, which are evident in the CEDRIC analyses (Figs. 10j,k), but less so in the SAMURAI analyses (Figs. 10l,m) due to the filtering scales and different approaches used (section 2b). The relatively weak vertical motions in the anvil (Figs. 10j–m) would not have been strong enough to suspend the radiosonde had the balloon burst, providing further evidence that the balloon did not burst until right before the radiosonde’s final descent to the surface.

Fig. 9.
Fig. 9.

As in Fig. 6, but for the radiosonde 2016-1 data. The light blue to dark blue shading in (a) represents the progression of time from launch to when the balloon likely burst.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

Fig. 10.
Fig. 10.

As in Figs. 6 and 9, but for two later times during the progression of radiosonde 2016-1.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

At 2341 UTC, radiosonde 2016-2 was launched to the south of the WER (Fig. 11b) and was likely closer to the regions with the most intense vertical motions than was radiosonde 2016–1. This radiosonde experienced strong southerly winds, particularly between 6 and 8 km MSL, (Figs. 4b,c) that advected it toward the storm’s updraft. At 8 km MSL, however, the radiosonde experienced its most intense negative acceleration (Figs. 11a,c, 5e) and a significant decrease in horizontal wind speeds (Fig. 4b), while the radiosonde was entering a region to the north with high reflectivity (>50 dbZ) and correlation coefficients <0.94, which suggests large hail. Based on this evidence, we suspect that the balloon burst at this time right before being entrained into the storm’s intense updraft. Therefore, adjustments were made to wair to account for this balloon burst assumption. However, we acknowledge that this 2016-2 balloon burst assumption is more uncertain than that for the 2017-1 balloon. The radiosonde measured a maximum estimated wair of 49.9 m s−1 at 10.3 km MSL (Fig. 11a). Shortly after this maximum value was reached, the radiosonde was located within the region of maximum reflectivity at 12.1 km MSL (Fig. 11d). This suggests that the radiosonde was near some of the storm’s most intense vertical motions, which were able to loft large hydrometeors to these near-tropopause heights. Considering the assumptions and adjustments for balloon bursting, 49.9 m s−1 was the strongest vertical velocity observed by a radiosonde from these two C3LOUD-Ex cases. This result is consistent with the fact that this radiosonde was launched in the most unstable (i.e., highest CAPE) environment of all the radiosondes (Fig. 3c; Table 1), as will be discussed in section 5b. It is important to restate that none of these estimates considers the impacts of hydrometeors, which would lead to both an underestimation of and an additional uncertainty in the wair values presented.

Fig. 11.
Fig. 11.

As in Figs. 7 and 8, but for radiosonde 2016-2. The smaller red dots take into account adjustments assuming that the radiosonde balloon burst.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

5. Comparisons of radiosonde wair to other platforms

a. Comparisons with dual-Doppler estimates

In addition to contextualizing the radiosonde observations, the radar data also provide an independent estimate of wair for radiosondes 2017-1 and 2016-1. It is important to note the differences in the features that the two types of observing systems can resolve. The values in the dual-Doppler analyses represent the average vertical velocity over a cube with side lengths of 1 km (500 m) for the 2017 (2016) case using data collected over a 5-min interval. The radiosonde values, however, represent averages along a slantwise path corresponding to the radiosonde trajectory over the course of the 12-s averaging period (e.g., horizontal and vertical distances generally between 150 and 700 m). Such differences need to be considered when comparing these estimates of vertical velocity obtained using these different platforms.

A comparison of radiosonde wair with the dual-Doppler wair from SAMURAI and CEDRIC is shown in Fig. 12. The dual-Doppler analyses for each radar volume were calculated at the volume-scan midpoint time and were advected in time using the calculated storm motion for each radar volume to create a 4D dataset. These 4D data were interpolated in time and space to the same position as the radiosonde for this comparison. To account for shifts in position within the dual-Doppler analyses that may be due to small advection errors, we also show the range of values in the surrounding grid boxes that are 1 km from the radiosonde location in the horizontal plane. This spread does not, however, represent any underlying uncertainty in the radar dual-Doppler analyses, which can come from a variety of sources as described in the introduction. In particular, the distance of these C3LOUD-Ex storms from the radars, combined with the fixed NEXRAD radar scanning patterns, as well as the homogeneous advection corrections (e.g., Shapiro et al. 2010) could produce significant sources of error. However, these errors can only be quantified from additional observation system simulation experiments (OSSEs; e.g., Potvin et al. 2012; Oue et al. 2019; Dahl et al. 2019).

Fig. 12.
Fig. 12.

Comparison of radiosonde and dual-Doppler wair for radiosondes (a) 2017-1 and (b) 2016-1, as described in the text. The gray range for the radiosonde data represents the quantified uncertainty in wair. The green and blue dots represent the dual-Doppler analyses interpolated to the radiosonde position. The green and blue horizontal lines represent the range of values within 1 km in the horizontal direction of the radiosonde position within the dual-Doppler analyses.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

Comparisons cannot be made below 3.7 km MSL (Fig. 12a) and 6.0 km MSL (Fig. 12b) for the 2017 and 2016 cases, respectively, due to the lack of quality radar data at the radiosonde locations. This demonstrates one benefit of the radiosonde observations, namely their ability to sample vertical motions where radars only observe very low signal-to-noise ratios, such as below cloud base and along cloud edges. Based on the C3LOUD-Ex radiosonde observations, wair can approach 20 m s−1 in these regions.

Both dual-Doppler analyses show consistent trends and similar magnitudes of wair. In both cases and for both dual-Doppler analyses, at the locations where the radiosondes observe the strongest wair, the dual-Doppler wair values was generally 15–20 m s−1 less than those derived from the radiosondes. For radiosonde 2017-1 (Fig. 12a), right before it is assumed that the balloon burst at 7.5 km MSL, the difference between the radiosonde wair and those of both dual-Doppler analyses was ~15–20 m s−1. For radiosonde 2016-1 (Fig. 12b), similar differences were present at 6.8 and 9.1 km MSL. This dual-Doppler underestimation of wair as compared to the most intense radiosonde wair was at least partly due to the radiosonde capturing localized features that were unable to be resolved by the resolution of these radar analyses. However, without a detailed error estimation of the dual-Doppler syntheses obtained from OSSEs for these cases, we are unable to quantify how much of the differences are due to errors associated with the C3LOUD-Ex radar network and scanning patterns (e.g., Oue et al. 2019) versus systematic differences in the observed quantities. Regardless, this comparison does demonstrate that a comprehensive analysis of wair would benefit from in situ measurements that can better capture highly localized conditions.

b. Comparisons with parcel theory

Parcel theory can also be used to estimate the theoretical maximum possible vertical velocity due to its relationship with CAPE (e.g., Weisman and Klemp 1984):
wMLCAPE=2×MLCAPE.

MLCAPE is chosen, as compared to other CAPE variants (e.g., surface-based or most-unstable), because it more realistically represents the air entering deep convective updrafts. The expression shown in Eq. (4) assumes that vertical accelerations are only forced by buoyancy and does not account for the negative impacts from condensate loading and entrainment. Equation (4) also does not consider the impacts of perturbation pressure gradients, which have been shown to decelerate updrafts within the upper levels of supercells where the maximum vertical velocities are achieved (Peters et al. 2019). Therefore, Eq. (4) likely overestimates the maximum vertical velocities in supercell updrafts.

To assess Eq. (4) with respect to the C3LOUD-Ex observations, MLCAPE (0–90 hPa AGL) is calculated for each radiosonde launch. These calculations assume pseudoadiabatic ascent and account for the latent heating associated with freezing above the 0°C level by assuming that ice fraction linearly increases from 0° to −40°C. While the subcloud-layer radiosonde data sampled by the updraft radiosondes are generally representative of the environmental air entering the supercell updraft, the data within the cloudy updraft are no longer representative of the environmental conditions needed to estimate MLCAPE. Therefore, the thermodynamic data from lowest levels of the updraft soundings were merged with data from the middle and upper levels of the environmental soundings (Figs. 3b,d). This concatenation occurred at the altitude where the temperature profiles first overlapped for each pair of soundings, near the inversion of the environmental sounding between 700 and 800 hPa. In cases where the radiosonde was launched in a cold pool or the thermodynamic data were not available (radiosondes 2016–1, 2017–1, and 2017–3), the closest, representative radiosonde launch in time and space was used as a better estimate of the inflow air for that radiosonde launch, since we are interested in estimating the theoretical maximum vertical velocities.

Overall, the wMLCAPE values calculated via parcel theory were larger than the wair values observed by the radiosondes (Table 1). Further, these results highlight the variability of wair within the primary supercell updraft. The ratio of wair to wMLCAPE ranges from 42% to 89%, largely due to the variability in the positions sampled within the supercell updrafts. When the balloon burst assumption is not considered, the ratios for radiosondes 2017-1 and 2016-2 fall from 89% to 67% and from 74% to 38%, respectively. Assuming the correct identification and adjustments for balloon bursts, the radiosonde with the smallest ratio (42%, 2016-1) sampled the extreme western edge of the primary updraft, ~10 km from the WER (Fig. 9). The radiosonde with the largest ratio (89%, 2017-1) sampled close to where the most intense vertical motions were likely located (Fig. 6). While the maximum vertical velocities estimated from these radiosonde data do not reach their theoretical maxima, as predicted by Eq. (4), a larger sample of observations, especially those similar to radiosonde 2017-1 that sampled near the most intense wair, is needed to better observationally assess the relationship shown in Eq. (4).

Table 1.

The 0–90 hPa AGL MLCAPE, the theoretical maximum wMLCAPE based on Eq. (4), and comparisons with the maximum radiosonde wair for each radiosonde launch. For instances where the assumption of a balloon burst was used, two values are shown. The first represents the value including the balloon burst assumption, while the second, in parentheses, represents the value without adjusting for a balloon burst.

Table 1.

6. Implications for future in situ observations of wair within storms

This study has shown that GPS sensors aboard radiosondes can provide useful in situ observations of wair within storms, especially when used in conjunction with radar data. Understanding the position within the updraft being sampled by the radiosonde provided valuable context for interpreting the radiosonde observations. Particularly with GPS radiosondes that can directly transmit their locations while sampling, coordinated scanning of radars through the use of PPIs and RHIs to the exact positions of airborne radiosondes should be considered for future field campaigns. For example, using these collocated radar and radiosonde observations, we demonstrated that most of the radiosonde measurements were likely several kilometers away from the strongest wair in these two supercell updrafts. Obtaining large samples of in situ observations in the locations of strongest wair within storms continues to be challenge, but forgoing cost constraints, this sampling difficulty can be alleviated by launching a high number of GPS sensors into storms (e.g., Markowski et al. 2018) so as to increase the probability of sampling the most intense vertical motions. This would also simultaneously improve the spatial coverage of these in situ measurements.

While several of the uncertainties in the radiosonde-based wair were quantified in this study, we did not quantify the uncertainty associated with hydrometeor collisions and collection on the radiosonde system. Innovative techniques and technologies to minimize or quantify these hydrometeor impacts would improve radiosonde observations within cloud systems. For example, cameras have been placed on radiosondes to assess icing impacts on in situ observations within winter storms (Waugh and Schuur 2018), and similar strategies could potentially be used to observe the possible accumulation of hydrometeors on the radiosonde system within updrafts. Furthermore, we analyzed balloon accelerations and assumed, with contextual support, that the radiosonde balloon burst when it experienced its most negative accelerations, in order to obtain a better estimate of wair. However, this assumption was more uncertain for radiosonde 2016-2 due to the less clear trajectory and more turbulent conditions, as compared with radiosonde 2017-1. Additional sensors could be introduced to the radiosonde system to assist in assessing balloon burst events, which would reduce these uncertainties.

7. Conclusions

One of the goals of the C3LOUD-Ex field campaign was to obtain in situ observations of the vertical velocities of supercell updrafts (wair) with targeted radiosonde launches. In situ observations of supercell vertical velocities have been limited, despite their importance for understanding physical processes within supercells and for verifying simulations as well as other observational platforms with difficult-to-characterize uncertainties. In this study, we present observations of wair from two isolated supercell cases observed during C3LOUD-Ex, which occurred in the High Plains of Colorado, Wyoming, and Nebraska. Radiosonde wair estimates were based on GPS data and were calculated with an uncertainty of ±2.6 m s−1, which considered uncertainties associated with the GPS measurements themselves, the helium balloon buoyancy, and varying drag forces. These estimates, however, did not consider hydrometeor impacts on the radiosonde systems which could be significant and would lead to an underestimation of the wair presented in this study.

In two of the five updraft radiosonde launches assessed in this study, we inferred that the radiosonde balloon burst while within the updraft, based on the extrema in the radiosonde negative accelerations. In these instances, we adjusted the wair estimates to account for the loss of buoyancy associated with balloon bursting. Before these adjustments, the maximum radiosonde wair was 36.2 m s−1 at an altitude of 10.4 km MSL during the 2017 case. After these adjustments, the maximum wair that was observed was 49.9 m s−1 at an altitude of 10.3 km MSL during the 2016 case, which occurred in the most unstable environment. At the lower and middle tropospheric levels, radiosonde 2017-3 captured the greatest wair and was located within the WER, reaching a maximum value of 31.1 m s−1 at 7.1 km MSL. In most of the observations presented, the radar data suggested that the radiosondes were several km away from the strongest wair within the supercell updraft. This fact, along with the potential impacts of hydrometeors on the radiosonde systems, suggests that the maximum wair in these two supercells was likely even larger than the values reported here.

The C3LOUD-Ex radiosonde observations were also compared with other methods of obtaining wair. One radiosonde in each of the two supercell cases sampled the updraft within the regions where dual-Doppler analyses could be performed, allowing for an independent measure of wair. For the locations where the radiosondes observed the greatest wair, the dual-Doppler wair values were generally 15–20 m s−1 less than the radiosonde estimated wair values. This was at least partly due to the different scales being observed by these two platforms, although it was difficult to fully quantify these differences without a detailed assessment of the dual-Doppler errors, such as may be obtained through the use of OSSEs, and which is left for future work. However, these comparisons did demonstrate that radiosondes provide complementary data to multi-Doppler analyses in terms of their ability to sample regions with low signal-to-noise ratios and to provide localized, high-resolution observations, both of which can be challenging in multi-Doppler analyses. When the balloon burst correction was included, the maximum radiosonde-based wair values were 42%–89% of the theoretical maximum wair from parcel theory. The variability in these comparisons was primarily due to the locations within the broad supercell updrafts that were sampled by the radiosondes, which were ascertained using collocated radar data.

Some of the challenges associated with making radiosonde observations of updrafts were highlighted here, and additional ideas on how these challenges can be surmounted were provided. There continues to be large uncertainty in the vertical velocities within deep convection, which are important for understanding many atmospheric processes and improving models. In situ observations of wair can complement remotely sensed estimates both by providing both an independent measure of wair for comparison and by observing finer-scale motions that often cannot be resolved using remote sensing. As such, despite their relative scarcity, in situ observations of wair can contribute to a more comprehensive understanding of storm vertical motions and hence should be considered for future field campaigns.

Acknowledgments

Funding from the Monfort Excellence Fund provided to Susan C. van den Heever as a Monfort Professor at Colorado State University is acknowledged, as well as funding from NSF grants AGS-1409686 and AGS-2019947. Peter Marinescu, Sean Freeman, and Aryeh Drager were also partially supported by NSF Grant No. DGE-1321845 Amend 5 and NSF Grant No. DGE-1840343. Michael Bell was supported by NSF Grants AGS-1701225 and OAC-1661663. We would like to acknowledge the entire C3LOUD-Ex science team for their time and efforts, particularly Dr. Emily Riley Dellaripa for assisting in the initial discussions for this work. We would also like to acknowledge Mark Benoit from InterMet systems for assisting with the radiosonde system and Dr. Stacey Hitchcock for useful discussions. We thank Dr. Nathan Dahl and two anonymous reviewers for their constructive and thorough feedback on this work.

Data availability statement

The radar and radiosonde data analyzed in this manuscript are all available upon request. The HRRR data were obtained from an archive of the High Resolution Rapid Refresh model (doi:10.7278/S5JQ0Z5B).

APPENDIX A

Power Spectra Analysis of Radiosonde Velocity Data

To determine the choice of Δt, power spectra were computed for all the clear air and updraft radiosonde launches presented in this manuscript. The vertical wind speeds (wair), horizontal wind speed (hspd), and horizontal wind direction (hdir) were first calculated with a Δt of 2 s, using the GPS data from the position 1 s before and after the current position. Power spectra analyses, following the methodology in Marinescu et al. (2019), were then conducted on these data. To summarize this methodology, these 1-Hz data were broken down into data chunks that were 180 s long and accurately resolved periodic signals from 2 to 90 s. These data chunking resulted in anywhere from 5 to 49 data chunks for each radiosonde launch, and the power spectra from these data chunks were then averaged together to create a better statistical representation of the periodic signals within each launch’s data (thin lines in Fig. A1). The power spectra from each radiosonde launch were also combined for all the clear air, updraft, and all launches, respectively, and averaged (thick lines in Fig. A1). Red-noise power spectra were estimated using the average lag-1 autocorrelations from these data groups, as a reference for these data without any periodic signals (Gilman et al. 1963). A periodic signal in the observed data is interpreted to be present if the data has more power than the red-noise power spectra for that period. From these analyses, it is clear that periodic signals are present in this data on time scales of ~12 s and less in all the wind data. These results are consistent with the theoretical calculation of the ~11–12-s period of a pendulum with a 30-m string, which represents the length of the radiosonde dereeler using during C3LOUD-Ex. Interestingly, the updraft launches (red, thick line) have more consistent periodic signals with periods between 6.0 and 6.5 s in the wind speeds, while the clear-air launches have more consistent signals between 9–12, 5, and 3.3 s, suggesting slight differences in the radiosondes’ periodic motions between these two conditions. Overall, these power spectra guided the choice of using a Δt of 12 s, which substantially reduced the contribution of these periodic signals with time scales of 12 s and less in the updraft wair calculations, while still allowing for finer-scale observations and error propagation analyses.

Fig. A1.
Fig. A1.

Mean power spectra for the 13 clear-air (thin gray lines) and 5 updraft (thin red and blue lines) radiosonde launches for (a) wair, (b) hspd, and (c) hdir. The data chunk that was used was 180 s, and the number of chunks that went into each radiosonde launch is shown in the legend. For the 13 clear-air launches, the number of chunks varied from 18 to 49. The thick solid black line represents the mean power spectra for all the data, while the thick solid gray and red lines represent the means of the clear-air and updraft launches, respectively. Estimates of the red noise spectra are also shown as thick dashed lines.

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

APPENDIX B

Analysis of ϵw,upd-drag

The term ϵw,upd-drag is the uncertainty in the wair estimate arising from changes in the drag force on the radiosonde system within an updraft as compared to still-air conditions. Because radiosonde systems typically reach their terminal velocity within a couple of seconds and are often close to terminal-velocity balance, we can use the formula for the terminal velocity and its dependence on the drag coefficient (CD) to estimate the uncertainty.

The terminal velocity (υT) of the radiosonde system can be determined as follows (following, e.g., Wang et al. 2009; Gallice et al. 2011):
υT=2g(netfreelift)ρCDA

In Eq. (B1), net free lift (units of kg), when multiplied by acceleration due to gravity g ~ 9.81 m s−2, is the upward buoyant force acting on the radiosonde system. Net free lift is calculated as the difference of two quantities: 1) the mass measured when the helium-filled balloon is attached to a spring scale (typical value of 1.03 kg; range from 0.86 to 1.40 kg); and 2) the combined mass of the radiosonde and dereeler attached to the balloon (0.24 kg). These measurements were taken during the clear-sky, still-air launches described in section 2a. The other variables in Eq. (B1) include the ambient air density ρ, the drag coefficient CD, and balloon cross-sectional area A. The helium inside the balloon is assumed to expand adiabatically as the balloon rises. The initial A of the balloon is approximately 1.33 m2, obtained from the clear-sky, still-air launches. Based on prior laboratory studies using perfect spheres (Achenbach 1972; Son et al. 2010) and on radiosonde observations during relatively calm, nighttime conditions (Gallice et al. 2011), drag coefficients for tropospheric conditions generally fall between 0.2 and 0.5. The drag coefficient within a supercell updraft may fall outside of this range, but we have no way of knowing whether this is the case due to the lack of observations. Using the known range of tropospheric drag coefficients from relatively calm conditions and using a range of tropospheric air densities, we can estimate the uncertainty of υT, and thus wair, due to variations in CD based on Eq. (B1) (Fig. B1). The range of υT as a function of air density (gold line) is at most 3.1 m s−1, which occurs at the lowest density included (0.3 kg m−3, representative of the upper troposphere). Therefore, we estimate that ϵw,upd-drag is ±1.6 m s−1, which is half of the maximum range (3.1 m s−1).

Fig. B1.
Fig. B1.

Terminal velocity calculations (m s−1) for ascending C3LOUD-Ex radiosondes with varying drag coefficients and densities (shaded, left axis) and the range (maximum minus minimum) of terminal velocities for each density (right axis).

Citation: Monthly Weather Review 148, 11; 10.1175/MWR-D-20-0071.1

REFERENCES

  • Achenbach, E., 1972: Experiments on the flow past spheres at very high Reynolds numbers. J. Fluid Mech., 54, 565575, https://doi.org/10.1017/S0022112072000874.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Armijo, L., 1969: A theory for the determination of wind and precipitation velocities with Doppler radars. J. Atmos. Sci., 26, 570573, https://doi.org/10.1175/1520-0469(1969)026<0570:ATFTDO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balakrishnan, N., and D. S. Zrnić, 1990: Use of polarization to characterize precipitation and discriminate large hail. J. Atmos. Sci., 47, 15251540, https://doi.org/10.1175/1520-0469(1990)047<1525:UOPTCP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., 1970: Some aspects of a severe, right-moving thunderstorm deduced from mesonetwork rawinsonde observations. J. Atmos. Sci., 27, 634648, https://doi.org/10.1175/1520-0469(1970)027<0634:SAOASR>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bell, M. M., M. T. Montgomery, and K. A. Emanuel, 2012: Air–sea enthalpy and momentum exchange at major hurricane wind speeds observed during CBLAST. J. Atmos. Sci., 69, 31973222, https://doi.org/10.1175/JAS-D-11-0276.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., E. W. McCaul, G. P. Byrd, and G. R. Woodall, 1988: Mobile sounding observations of a tornadic storm near the dryline: The Canadian, Texas, storm of 7 May 1986. Mon. Wea. Rev., 116, 17901804, https://doi.org/10.1175/1520-0493(1988)116<1790:MSOOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., E. W. McCaul, G. P. Byrd, G. R. Woodall, G. Martin, S. Keighton, and L. C. Showell, 1989: Mobile sounding observations of a thunderstorm near the dryline: The Gruver, Texas, storm complex of 25 May 1987. Mon. Wea. Rev., 117, 244250, https://doi.org/10.1175/1520-0493(1989)117<0244:MSOOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bousquet, O., P. Tabary, and J. Parent du Châtelet, 2008: Operational multiple-Doppler wind retrieval inferred from long-range radial velocity measurements. J. Appl. Meteor. Climatol., 47, 29292945, https://doi.org/10.1175/2008JAMC1878.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Browning, K. A., and F. H. Ludlam, 1962: Airflow in convective storms. Quart. J. Roy. Meteor. Soc., 88, 117135, https://doi.org/10.1002/qj.49708837602.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Browning, K. A., and G. B. Foote, 1976: Airflow and hail growth in supercell storms and some implications for hail suppression. Quart. J. Roy. Meteor. Soc., 102, 499533, https://doi.org/10.1002/qj.49710243303.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brunkow, D., V. N. Bringi, P. C. Kennedy, S. A. Rutledge, V. Chandrasekar, E. A. Mueller, and R. K. Bowie, 2000: A description of the CSU–CHILL national radar facility. J. Atmos. Oceanic Technol., 17, 15961608, https://doi.org/10.1175/1520-0426(2000)017<1596:ADOTCC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bryan, G. H., 2008: getcape. Accessed 1 July 2019, https://www2.mmm.ucar.edu/people/bryan/Code/getcape.F.

  • Chisholm, A. J., 1970: Alberta hailstorms: A radar study and model. Ph.D. thesis, McGill University, 237 pp.

  • Chisholm, A. J., 1973: Alberta hailstorms Part I: Radar case studies and airflow models. Alberta Hailstorms, Meteor. Monogr., No. 36, Amer. Meteor. Soc., 1–36.

    • Crossref
    • Export Citation
  • Collis, S., A. Protat, and K.-S. Chung, 2010: The effect of radial velocity gridding artifacts on variationally retrieved vertical velocities. J. Atmos. Oceanic Technol., 27, 12391246, https://doi.org/10.1175/2010JTECHA1402.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dahl, N. A., A. Shapiro, C. K. Potvin, A. Theisen, J. G. Gebauer, A. D. Schenkman, and M. Xue, 2019: High-resolution, rapid-scan dual-Doppler retrievals of vertical velocity in a simulated supercell. J. Atmos. Oceanic Technol., 36, 14771500, https://doi.org/10.1175/JTECH-D-18-0211.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies-Jones, R. P., 1974: Discussion of measurements inside high-speed thunderstorm updrafts. J. Appl. Meteor., 13, 710717, https://doi.org/10.1175/1520-0450(1974)013<0710:DOMIHS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies-Jones, R. P., and J. H. Henderson, 1975: Updraft properties deduced statistically from Rawin soundings. Pure Appl. Geophys., 113, 787801, https://doi.org/10.1007/BF01592959.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • DiGangi, E. A., D. R. MacGorman, C. L. Ziegler, D. Betten, M. Biggerstaff, M. Bowlan, and C. K. Potvin, 2016: An overview of the 29 May 2012 Kingfisher supercell during DC3. J. Geophys. Res. Atmos., 121, 14 31614 343, https://doi.org/10.1002/2016JD025690.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dolan, B., and S. A. Rutledge, 2010: Using CASA IP1 to diagnose kinematic and microphysical interactions in a convective storm. Mon. Wea. Rev., 138, 16131634, https://doi.org/10.1175/2009MWR3016.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fan, J., and Coauthors, 2017: Cloud-resolving model intercomparison of an MC3E squall line case: Part I—Convective updrafts. J. Geophys. Res. Atmos., 122, 93519378, https://doi.org/10.1002/2017JD026622.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Farley, R. E., 2005: BalloonAscent: 3-D simulation tool for the ascent and float of high-altitude balloons. AIAA Fifth Aviation, Technology, Integration, and Operations Conf., Reston, VA, AIAA, 15 pp.

    • Crossref
    • Export Citation
  • Foote, G. B., and J. C. Fankhauser, 1973: Airflow and moisture budget beneath a northeast Colorado hailstorm. J. Appl. Meteor., 12, 13301353, https://doi.org/10.1175/1520-0450(1973)012<1330:AAMBBA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gal-Chen, T., 1978: A method for the initialization of the anelastic equations: Implications for matching models with observations. Mon. Wea. Rev., 106, 587606, https://doi.org/10.1175/1520-0493(1978)106<0587:AMFTIO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gallice, A., F. G. Wienhold, C. R. Hoyle, F. Immler, and T. Peter, 2011: Modeling the ascent of sounding balloons: Derivation of the vertical air motion. Atmos. Meas. Tech., 4, 22352253, https://doi.org/10.5194/amt-4-2235-2011.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gao, J., M. Xue, A. Shapiro, and K. K. Droegemeier, 1999: A variational method for the analysis of three-dimensional wind fields from two Doppler radars. Mon. Wea. Rev., 127, 21282142, https://doi.org/10.1175/1520-0493(1999)127<2128:AVMFTA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Geerts, B., and Coauthors, 2018: Recommendations for in situ and remote sensing capabilities in atmospheric convection and turbulence. Bull. Amer. Meteor. Soc., 99, 24632470, https://doi.org/10.1175/BAMS-D-17-0310.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilman, D. L., F. J. Fuglister, and J. M. Mitchell, 1963: On the power spectrum of “red noise.” J. Atmos. Sci., 20, 182184, https://doi.org/10.1175/1520-0469(1963)020<0182:OTPSON>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Helmus, J. J., and S. M. Collis, 2016: The Python ARM Radar Toolkit (Py-ART), a library for working with weather radar data in the Python programming language. J. Open Res. Software, 4, e25, https://doi.org/10.5334/jors.119.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., and D. J. Musil, 1982: Case study of a hailstorm in Colorado. Part II: Particle growth processes at mid-levels deduced from in-situ measurements. J. Atmos. Sci., 39, 28472866, https://doi.org/10.1175/1520-0469(1982)039<2847:CSOAHI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • InterMet Systems, 2016: iMet-1-ABxn data sheet. Grand Rapids Michigan, 1 pp., https://www.intermetsystems.com/ee/pdf/202060_iMet-1-ABxn_Data_161006.pdf.

  • Kropfli, R. A., and L. J. Miller, 1976: Kinematic structure and flux quantities in a convective storm from dual-Doppler radar observations. J. Atmos. Sci., 33, 520529, https://doi.org/10.1175/1520-0469(1976)033<0520:KSAFQI>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lehmiller, G. S., H. B. Bluestein, P. J. Neiman, F. M. Ralph, and W. F. Feltz, 2001: Wind structure in a supercell thunderstorm as measured by a UHF wind profiler. Mon. Wea. Rev., 129, 19681986, https://doi.org/10.1175/1520-0493(2001)129<1968:WSIAST>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marinescu, P. J., S. C. van den Heever, S. M. Saleeby, and S. M. Kreidenweis, 2016: The microphysical contributions to and evolution of latent heating profiles in two MC3E MCSs. J. Geophys. Res. Atmos., 121, 79137935, https://doi.org/10.1002/2016JD024762.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marinescu, P. J., E. J. T. Levin, D. Collins, S. M. Kreidenweis, and S. C. van den Heever, 2019: Quantifying aerosol size distributions and their temporal variability in the Southern Great Plains, USA. Atmos. Chem. Phys., 19, 11 98512 006, https://doi.org/10.5194/acp-19-11985-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., Y. P. Richardson, S. J. Richardson, and A. Petersson, 2018: Aboveground thermodynamic observations in convective storms from balloonborne probes acting as pseudo-Lagrangian drifters. Bull. Amer. Meteor. Soc., 99, 711724, https://doi.org/10.1175/BAMS-D-17-0204.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marshall, T. C., W. D. Rust, and M. Stolzenburg, 1995: Electrical structure and updraft speeds in thunderstorms over the southern Great Plains. J. Geophys. Res., 100, 10011015, https://doi.org/10.1029/94JD02607.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marwitz, J. D., 1972: The structure and motion of severe hailstorms. Part I: Supercell storms. J. Appl. Meteor., 11, 166179, https://doi.org/10.1175/1520-0450(1972)011<0166:TSAMOS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marwitz, J. D., 1973: Trajectories within the weak echo regions of hailstorms. J. Appl. Meteor., 12, 11741182, https://doi.org/10.1175/1520-0450(1973)012<1174:TWTWER>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marwitz, J. D., and E. X. Berry, 1971: The airflow within the weak echo region of an Alberta hailstorm. J. Appl. Meteor., 10, 487492, https://doi.org/10.1175/1520-0450(1971)010<0487:TAWTWE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, L. J., 1975: Internal airflow of a convective storm from dual-Doppler radar measurements. Pure Appl. Geophys., 113, 765785, https://doi.org/10.1007/BF01592958.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Miller, L. J., and S. M. Fredrick, 1998: CEDRIC custom editing and display of reduced information in Cartesian space. NCAR, 130 pp., https://www.eol.ucar.edu/system/files/cedric_doc.pdf.

  • Mullendore, G. L., D. R. Durran, and J. R. Holton, 2005: Cross-tropopause tracer transport in midlatitude convection. J. Geophys. Res., 110, D06113, https://doi.org/10.1029/2004JD005059.

    • Search Google Scholar
    • Export Citation
  • Musil, D. J., A. J. Heymsfield, and P. L. Smith, 1986: Microphysical characteristics of a well-developed weak echo region in a high plains supercell thunderstorm. J. Climate Appl. Meteor., 25, 10371051, https://doi.org/10.1175/1520-0450(1986)025<1037:MCOAWD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • National Centers for Environmental Information (NCEI), 2016: NCDC storm events database (storm data). NOAA, accessed 13 February 2020, https://www.ncdc.noaa.gov/stormevents/.

  • National Centers for Environmental Information (NCEI), 2017: NCDC storm events database (storm data). NOAA, accessed 13 February 2020, https://www.ncdc.noaa.gov/stormevents/.

  • Nelson, S. P., and R. A. Brown, 1987: Error sources and accuracy of vertical velocities computed from multiple-Doppler radar measurements in deep convective storms. J. Atmos. Oceanic Technol., 4, 233238, https://doi.org/10.1175/1520-0426(1987)004<0233:ESAAOV>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ooyama, K. V., 2002: The cubic-spline transform method: Basic definitions and tests in a 1D single domain. Mon. Wea. Rev., 130, 23922415, https://doi.org/10.1175/1520-0493(2002)130<2392:TCSTMB>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Oue, M., P. Kollias, A. Shapiro, A. Tatarevic, and T. Matsui, 2019: Investigation of observational error sources in multi-Doppler-radar three-dimensional variational vertical air motion retrievals. Atmos. Meas. Tech., 12, 19992018, https://doi.org/10.5194/amt-12-1999-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Palmer, A. D. F., 1912: The Theory of Measurements. McGraw-Hill, 264 pp.

  • Peters, J. M., C. J. Nowotarski, and H. Morrison, 2019: The role of vertical wind shear in modulating maximum supercell updraft velocities. J. Atmos. Sci., 76, 31693189, https://doi.org/10.1175/JAS-D-19-0096.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Potvin, C. K., D. Betten, L. J. Wicker, K. L. Elmore, and M. I. Biggerstaff, 2012: 3DVAR versus traditional dual-Doppler wind retrievals of a simulated supercell thunderstorm. Mon. Wea. Rev., 140, 34873494, https://doi.org/10.1175/MWR-D-12-00063.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Purser, R. J., W. S. Wu, D. F. Parrish, and N. M. Roberts, 2003: Numerical aspects of the application of recursive filters to variational statistical analysis. Part I: Spatially homogeneous and isotropic Gaussian covariances. Mon. Wea. Rev., 131, 15241535, https://doi.org/10.1175//1520-0493(2003)131<1524:NAOTAO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., M. R. Kumjian, S. M. Ganson, and P. Zhang, 2013: Polarimetric radar characteristics of melting hail. Part II: Practical implications. J. Appl. Meteor. Climatol., 52, 28712886, https://doi.org/10.1175/JAMC-D-13-074.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Shapiro, A., K. M. Willingham, and C. K. Potvin, 2010: Spatially variable advection correction of radar data. Part II: Test results. J. Atmos. Sci., 67, 34573470, https://doi.org/10.1175/2010JAS3466.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Söder, J., M. Gerding, A. Schneider, A. Dörnbrack, H. Wilms, J. Wagner, and F. J. Lübken, 2019: Evaluation of wake influence on high-resolution balloon-sonde measurements. Atmos. Meas. Tech., 12, 41914210, https://doi.org/10.5194/amt-12-4191-2019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Son, K., J. Choi, W. P. Jeon, and H. Choi, 2010: Effect of free-stream turbulence on the flow over a sphere. Phys. Fluids, 22, 045101, https://doi.org/10.1063/1.3371804.

    • Crossref
    • 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, https://doi.org/10.1175/1520-0434(2003)018<1243:CPSWSE>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Varble, A., and Coauthors, 2014: Evaluation of cloud-resolving and limited area model intercomparison simulations using TWP-ICE observations: 1. Deep convective updraft properties. J. Geophys. Res., 119, 13 89113 918, https://doi.org/10.1002/2013JD021372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, J., J. Bian, W. O. Brown, H. Cole, V. Grubišić, and K. Young, 2009: Vertical air motion from T-REX radiosonde and dropsonde data. J. Atmos. Oceanic Technol., 26, 928942, https://doi.org/10.1175/2008JTECHA1240.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Waugh, S., and T. J. Schuur, 2018: On the use of radiosondes in freezing precipitation. J. Atmos. Oceanic Technol., 35, 459472, https://doi.org/10.1175/JTECH-D-17-0074.1.

    • Crossref
    • 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, https://doi.org/10.1175/1520-0493(1982)110<0504:TDONSC>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weisman, M. L., and J. B. Klemp, 1984: The structure and classification of numerically simulated convective storms in directionally varying wind shears. Mon. Wea. Rev., 112, 24792498, https://doi.org/10.1175/1520-0493(1984)112<2479:TSACON>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
1

The CAPE calculations in this study are based on Bryan (2008).

Save