• Adler, R. F., , H.-Y. M. Yeh, , N. Prasad, , W.-K. Tao, , and J. Simpson, 1991: Microwave simulations of a tropical rainfall system with a three-dimensional cloud model. J. Appl. Meteor., 30, 924953, doi:10.1175/1520-0450-30.7.924.

    • Search Google Scholar
    • Export Citation
  • Atlas, D., , S. Y. Matrosov, , A. J. Heymsfield, , M.-D. Chou, , and D. B. Wolff, 1995: Radar and radiation properties of ice clouds. J. Appl. Meteor., 34, 23292345, doi:10.1175/1520-0450(1995)034<2329:RARPOI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., 2009: Passive microwave brightness temperatures as proxies for hailstorms. J. Appl. Meteor. Climatol., 48, 12811286, doi:10.1175/2009JAMC2125.1.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., , and C. B. Blankenship, 2012: Toward a global climatology of severe hailstorms as estimated by satellite passive microwave imagers. J. Climate, 25, 687703, doi:10.1175/JCLI-D-11-00130.1.

    • Search Google Scholar
    • Export Citation
  • Dolan, B., , and S. A. Rutledge, 2009: A theory-based hydrometeor identification algorithm for X-band polarimetric radars. J. Atmos. Oceanic Technol., 26, 20712088, doi:10.1175/2009JTECHA1208.1.

    • Search Google Scholar
    • Export Citation
  • Ferraro, R. R., , and G. F. Marks, 1995: The development of SSM/I rain-rate retrieval algorithms using ground-based radar measurements. J. Atmos. Oceanic Technol., 12, 755770, doi:10.1175/1520-0426(1995)012<0755:TDOSRR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ferraro, R. R., , J. Beauchamp, , D. J. Cecil, , and G. M. Heymsfield, 2015: A prototype hail detection algorithm and hail climatology developed with the Advanced Microwave Sounding Unit (AMSU). Atmos. Res., doi:10.1016/j.atmosres.2014.08.010, in press.

    • Search Google Scholar
    • Export Citation
  • Gatlin, P. N., , M. Thurai, , V. N. Bringi, , W. Petersen, , D. Wolff, , A. Tokay, , L. Carey, , and M. Wingo, 2015: Searching for large raindrops: A global summary of two-dimensional video disdrometer observations. J. Appl. Meteor. Climatol., 54, 10691089, doi:10.1175/JAMC-D-14-0089.1.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., 1977: Precipitation development in stratiform ice clouds: A microphysical and dynamical study. J. Atmos. Sci., 34, 367381, doi:10.1175/1520-0469(1977)034<0367:PDISIC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, G. M., , L. Tian, , L. Li, , M. McLinden, , and J. I. Cervantes, 2013: Airborne radar observations of severe hailstorms: Implications for future spaceborne radar. J. Appl. Meteor. Climatol., 52, 18511867, doi:10.1175/JAMC-D-12-0144.1.

    • Search Google Scholar
    • Export Citation
  • Hou, A. Y., and Coauthors, 2014: The Global Precipitation Measurement (GPM) mission. Bull. Amer. Meteor. Soc., 95, 701722, doi:10.1175/BAMS-D-13-00164.1.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C. D., , W. Barnes, , T. Kozu, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 809817, doi:10.1175/1520-0426(1998)015<0809:TTRMMT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liu, C., , D. Cecil, , and E. J. Zipser, 2011: Relationships between lightning flash rates and passive microwave brightness temperatures at 85 and 37 GHz over the tropics and subtropics. J. Geophys. Res., 116, D23108, doi:10.1029/2011JD016463.

    • Search Google Scholar
    • Export Citation
  • Liu, H., , and V. Chandrasekar, 2000: Classification of hydrometeors based on polarimetric radar measurements: Development of fuzzy logic and neuro-fuzzy systems, and in situ verification. J. Atmos. Oceanic Technol., 17, 140164, doi:10.1175/1520-0426(2000)017<0140:COHBOP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • McGaughey, G., , E. J. Zipser, , R. W. Spencer, , and R. E. Hood, 1996: High-resolution passive microwave observations of convective systems over the tropical Pacific Ocean. J. Appl. Meteor., 35, 19211947, doi:10.1175/1520-0450(1996)035<1921:HRPMOO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mohr, K. I., , and E. J. Zipser, 1996a: Defining mesoscale convective systems by their 85-GHz ice-scattering signatures. Bull. Amer. Meteor. Soc., 77, 11791189, doi:10.1175/1520-0477(1996)077<1179:DMCSBT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mohr, K. I., , and E. J. Zipser, 1996b: Mesoscale convective systems defined by their 85-GHz ice scattering signature: Size and intensity comparison over tropical oceans and continents. Mon. Wea. Rev., 124, 24172437, doi:10.1175/1520-0493(1996)124<2417:MCSDBT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mugnai, A., , E. A. Smith, , and G. J. Tripoli, 1993: Foundations for statistical–physical precipitation retrieval from passive microwave satellite measurements. Part II: Emission-source and generalized weighting-function properties of a time-dependent cloud-radiation model. J. Appl. Meteor., 32, 1739, doi:10.1175/1520-0450(1993)032<0017:FFSPRF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Oye, R., , C. Mueller, , and S. Smith, 1995: Software for radar translation, visualization, editing, and interpolation. Preprints, 27th Conf. on Radar Meteorology, Vail, CO, Amer. Meteor. Soc., 359–361.

  • Ryzhkov, A. V., , and D. S. Zrnic, 1995a: Comparison of dual-polarization radar estimators of rain. J. Atmos. Oceanic Technol., 12, 249256, doi:10.1175/1520-0426(1995)012<0249:CODPRE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., , and D. S. Zrnic, 1995b: Precipitation and attenuation measurements at a 10-cm wavelength. J. Appl. Meteor., 34, 21212134, doi:10.1175/1520-0450(1995)034<2120:PAAMAA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., , and D. S. Zrnic, 1996: Rain in shallow and deep convection measured with a polarimetric radar. J. Atmos. Sci., 53, 29892995, doi:10.1175/1520-0469(1996)053<2989:RISADC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., , D. S. Zrnic, , and B. A. Gordon, 1998: Polarimetric method for ice water content determination. J. Appl. Meteor., 37, 125134, doi:10.1175/1520-0450(1998)037<0125:PMFIWC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schuur, T. J., , A. V. Ryzhkov, , D. S. Zrnic, , and M. Schonhuber, 2001: Drop size distributions measured by a 2D video disdrometer: Comparison with dual-polarization radar data. J. Appl. Meteor., 40, 10191034, doi:10.1175/1520-0450(2001)040<1019:DSDMBA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Smith, E. A., , H. J. Cooper, , X. Xiang, , A. Mugnai, , and G. J. Tripoli, 1992: Foundations for statistical–physical precipitation retrieval from passive microwave satellite measurements. Part I: Brightness temperature properties of a time-dependent cloud-radiation model. J. Appl. Meteor., 31, 506531, doi:10.1175/1520-0450(1992)031<0506:FFSPPR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , and D. A. Santek, 1985: Measuring the global distribution of intense convection over land with passive microwave radiometry. J. Climate Appl. Meteor., 24, 860864, doi:10.1175/1520-0450(1985)024<0860:MTGDOI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , W. S. Olson, , W. Rongzhang, , D. W. Martin, , J. A. Weinman, , and D. A. Santek, 1983: Heavy thunderstorms observed over land by the Nimbus 7 scanning multichannel microwave radiometer. J. Climate Appl. Meteor., 22, 10411046, doi:10.1175/1520-0450(1983)022<1041:HTOOLB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , M. R. Howland, , and D. A. Santek, 1987: Severe storm identification with satellite microwave radiometry: An initial investigation with Nimbus-7 SMMR data. J. Climate Appl. Meteor., 26, 749754, doi:10.1175/1520-0450(1987)026<0749:SSIWSM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , H. M. Goodman, , and R. E. Hood, 1989: Precipitation retrieval over land and ocean with the SSM/I: Identification and characteristics of the scattering signal. J. Atmos. Oceanic Technol., 6, 254273, doi:10.1175/1520-0426(1989)006<0254:PROLAO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , R. E. Hood, , F. J. LaFontaine, , E. A. Smith, , R. Platt, , J. Galliano, , V. L. Griffin, , and E. Lobl, 1994: High-resolution imaging of rain systems with the Advanced Microwave Precipitation Radiometer. J. Atmos. Oceanic Technol., 11, 849857, doi:10.1175/1520-0426(1994)011<0849:HRIORS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Straka, J. M., , D. S. Zrnic, , and A. V. Ryzhkov, 2000: Bulk hydrometeor classification and quantification using polarimetric radar data: Synthesis of relations. J. Appl. Meteor., 39, 13411372, doi:10.1175/1520-0450(2000)039<1341:BHCAQU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Thomason, J. W. G., , A. J. Illingworth, , and V. Marecal, 1995: Density and size distribution of aggregating snow particles inferred from coincident aircraft and radar observations. Preprints, 27th Conf. on Radar Meteorology, Vail, CO, Amer. Meteor. Soc., 127–129.

  • Toracinta, E. R., , D. J. Cecil, , E. J. Zipser, , and S. W. Nesbitt, 2002: Radar, passive microwave, and lightning characteristics of precipitating systems in the tropics. Mon. Wea. Rev., 130, 802824, doi:10.1175/1520-0493(2002)130<0802:RPMALC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wang, J. R., , P. E. Racette, , and J. R. Piepmeier, 2008: A comparison of near-concurrent measurements from the SSMIS and CoSMIR for some selected channels over the frequency range of 50–183 GHz. IEEE Trans. Geosci. Remote Sens., 46, 923933, doi:10.1109/TGRS.2007.904038.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , M. S. V. Rao, , E. B. Rodgers, , and J. S. Theon, 1977: A satellite technique for quantitatively mapping rainfall rates over the oceans. J. Appl. Meteor., 16, 551560, doi:10.1175/1520-0450(1977)016<0551:ASTFQM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , and L. S. Chiu, 1991: Retrieval of monthly rainfall indices from microwave radiometric measurements using probability distribution functions. J. Atmos. Oceanic Technol., 8, 118136, doi:10.1175/1520-0426(1991)008<0118:ROMRIF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Composite gridded KVNX reflectivity valid at 2144 UTC 24 May 2011 and (b) corresponding hydrometeor identification showing the hydrometeor with the greatest mass in each vertical column. The hydrometeors, as defined in the text, are DZ, RN, IC, AG, WS, VI, LG, HG, HL, and BD. The black line in each panel indicates the location of the cross section shown in Fig. 2, and the plus symbol shows the start position of each cross section. The KVNX radar is located near the center of each panel.

  • View in gallery

    Cross section of (a) gridded KVNX reflectivity along a 177-km ER-2 flight segment beginning at 2137 UTC 24 May 2011 and (b) corresponding hydrometeor identification. The location of the cross section is shown in Fig. 1 by the black line, and the plus symbol corresponds to the left side of each cross section. The lines plotted over (a) show the collocated AMPR BTs and over (b) show the collocated CoSMIR BTs. Hydrometeor categories are the same as in Fig. 1, and the triangle in each panel represents the lack of radar coverage associated with the radar’s cone of silence.

  • View in gallery

    As in Fig. 1, but valid at 2145 UTC 23 May 2011. The black line in each panel corresponds to the cross sections shown in Fig. 4, and the asterisk symbol in each panel corresponds to the location of the KVNX radar.

  • View in gallery

    Cross section of (a) gridded KVNX reflectivity along a 229-km ER-2 flight segment beginning at 2135 UTC 23 May 2011 and (b) corresponding hydrometeor identification. The location of the cross section is shown in Fig. 3 by the black line, and the plus symbol in Fig. 3 corresponds to the left side of each cross section. The lines plotted over (a) show the collocated CoSMIR BTs, and stippling in (b) indicates hydrometeor mass shown by the key to the right of (b). Hydrometeor categories are the same as in Fig. 1.

  • View in gallery

    (a) Composite gridded KVNX reflectivity valid at 2251 UTC 22 Apr 2011 and (b) ER-2 flight track on 22 Apr 2011. The asterisk in (a) indicates the location of the KVNX radar and in (b) indicates the start of the flight track. Portions of the flight track in (b) colored red (blue) correspond to collocated vertical profiles not associated with any hydrometeors and a 10-GHz BT > (≤) 285 K.

  • View in gallery

    BT PDFs for vertical profiles that contain HL, HG, and LG valid at (a) 10, (b) 19, (c) 37, and (d) 85 GHz. The BTs were collected by AMPR on 22 Apr 2011 or 24 May 2011. The “none” category includes profiles without any valid hydrometeor identification. For (a) and (b), the BT bin size is 10 (5) K to the left (right) of the thin vertical line. For (c) and (d) the bin size is 20 (10) K to the left (right) of the thin vertical line. The minimum BT measured at each frequency for the shown hydrometeor species is indicated by the number in the top-left corner of each panel.

  • View in gallery

    BT PDFs for vertical profiles that contain HL, HG, LG, and no valid hydrometeor identification (none) valid at (a) 89, (b) 165, and (c) 183 GHz. The BTs were measured with CoSMIR on 23–24 May 2011. The BT bin size is 10 (5) K to the left (right) of the thin vertical line in each panel. The values along the left (right) side of the plot correspond to the hail and graupel (none) categories. The minimum BT measured at each frequency for the shown hydrometeor species is indicated by the number in the top-left corner of each panel.

  • View in gallery

    As in Fig. 6, but for profiles containing WS, AG, small IC, and no valid hydrometeor identification.

  • View in gallery

    As in Fig. 7, but for profiles containing WS, AG, small IC, and no valid hydrometeor identification.

  • View in gallery

    BT PDFs for vertical profiles that contain WS, AG, IC, and no valid hydrometeor identification (none) valid at (a) 85 and (b) 89 GHz on 24 May 2011. The BT bin size is 20 (10) K to the left (right) of the thin vertical line in each panel.

  • View in gallery

    As in Fig. 6, but for profiles containing BD, RN, and profiles containing no valid hydrometeor identification.

  • View in gallery

    As in Fig. 7, but for profiles containing BD, RN, and profiles containing no valid hydrometeor identification.

  • View in gallery

    Probability of BD, HL, HG, LG, RN, WS, AG, IC, or DZ being the dominant hydrometeor species from our hierarchy as a function of collocated (a) 10-, (b) 19-, (c) 37-, and (d) 85-GHz BT measured with AMPR on 22 Apr or 24 May 2011. The BT bin sizes are as in Fig. 6.

  • View in gallery

    Probability of BD, HL, HG, LG, RN, WS, AG, IC, or DZ being the dominant hydrometeor species from our hierarchy as a function of collocated (a) 89-, (b) 165-, and (c) 183-GHz BT measured with CoSMIR on 23–24 May 2011. The BT bin sizes are as in Fig. 7.

  • View in gallery

    Probability of a vertically integrated nonzero mass of HL, HG, LG, and BD as a function of collocated AMPR BT at (a) 10, (b) 19, (c) 37, and (d) 85 GHz for data collected on 22 Apr or 24 May 2011. The thin gray line indicates the number of profiles in each BT bin as indicated by the axis along the right side of each panel. The thin vertical lines and BT bin sizes relative to those lines are the same as in Fig. 6.

  • View in gallery

    Probability of a vertically integrated nonzero mass of HL, HG, LG, and BD as a function of collocated CoSMIR BT at (a) 89, (b) 165, and (c) 183 GHz for data collected on 23–24 May 2011. The thin gray line indicates the number of profiles in each BT bin as indicated by the axis along the right side of each panel. The thin vertical lines and BT bin sizes relative to those lines are the same as in Fig. 7.

  • View in gallery

    Probability of a nonzero mass of any hydrometeor type at low levels (≤2.5-km height) as a function of collocated AMPR BT at various frequencies for data collected on 22 Apr or 24 May 2011 within 180 km of the radar. The blue dotted line shows the number of profiles in each 85-GHz BT bin as indicated by the axis along the right side of the figure (counts for other frequencies are not shown for clarity). The colored asterisks at the bottom of the figure indicate the BT at which the bin sizes change for each frequency, and bin sizes relative to each symbol are the same as in Fig. 6.

  • View in gallery

    Probability of a nonzero mass of any hydrometeor type at low levels (≤2.5-km height) as a function of collocated CoSMIR BT at various frequencies for data collected on 23–24 May 2011 within 180 km from the radar. The green dotted line shows the number of profiles in each 183-GHz BT bin as indicated by the axis along the right side of the figure (counts for other frequencies are not shown for clarity). To the left (right) of the thin vertical line, the BT bin size is 10 (5) K.

  • View in gallery

    Probability of the occurrence of HL at low levels (i.e., ≤2.5 km) or at upper levels (i.e., >2.5 km) in a vertical column as a function of collocated AMPR BT at (a) 10, (b) 19, (c) 37, and (d) 85 GHz for data collected on 22 Apr or 24 May 2011 within 180 km of the radar. The BT bin size in (a) is 10 K. In (b)–(d) the BT bin size is 20 (10) K to the left (right) of the thin vertical line.

  • View in gallery

    Probability of the occurrence of HL at low levels (i.e., ≤2.5 km) or at upper levels (i.e., >2.5 km) in a vertical column as a function of collocated CoSMIR BT at (a) 89, (b) 165, and (c) 183 GHz for data collected on 23–24 May 2011 within 180 km of the radar. The BT bin size is 20 (10) K to the left (right) of the thin vertical line in each panel.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 38 38 11
PDF Downloads 28 28 14

Signatures of Hydrometeor Species from Airborne Passive Microwave Data for Frequencies 10–183 GHz

View More View Less
  • 1 Earth System Science Center, University of Alabama in Huntsville, Huntsville, Alabama
  • 2 NASA Marshall Space Flight Center, Huntsville, Alabama
© Get Permissions
Full access

Abstract

Passive microwave brightness temperatures (BTs) collected above severe thunderstorms using the Advanced Microwave Precipitation Radiometer and Conical Scanning Millimeter-Wave Imaging Radiometer during the Midlatitude Continental Convective Clouds Experiment are compared with a hydrometeor identification applied to dual-polarimetric Weather Surveillance Radar-1988 Doppler radar data collected at Vance Air Force Base, Oklahoma (KVNX). The goal of this work is to determine the signatures of various hydrometeor species in terms of BTs measured at frequencies used by the Global Precipitation Measurement mission Microwave Imager. Results indicate that hail is associated with an ice-scattering signature at all frequencies examined, including 10.7 GHz. However, it appears that frequencies ≤ 37.1 GHz are most useful for identifying hail. Low-level (below 2.5 km) hail becomes probable for a BT below 240 K at 19.4 GHz, 170 K at 37.1 GHz, 90 K at 85.5 GHz, 80 K at 89.0 GHz, 100 K at 165.5 GHz, and 100 K at 183.3 ± 7 GHz. Graupel may be distinguished from hail and profiles without any hydrometeor species by its strong scattering signature at higher frequencies (e.g., 165.5 GHz) and its relative lack of scattering at frequencies ≤ 19.4 GHz. There is a clearer distinction between profiles that contain liquid precipitation and profiles without any hydrometeors when the liquid is associated above with hail and/or graupel (i.e., a hydrometeor category with a strong scattering signature) than when the liquid is associated with smaller ice. Near-surface precipitation is much more likely for a 19.4-GHz BT < 250 K, 37.1-GHz BT < 240 K, 89.0-GHz BT < 220 K, and 165.5-GHz BT < 140 K.

Corresponding author address: Kenneth Leppert II, NSSTC, Rm. 4074, 320 Sparkman Dr. Huntsville, AL 35805. E-mail: leppert@nsstc.uah.edu

Abstract

Passive microwave brightness temperatures (BTs) collected above severe thunderstorms using the Advanced Microwave Precipitation Radiometer and Conical Scanning Millimeter-Wave Imaging Radiometer during the Midlatitude Continental Convective Clouds Experiment are compared with a hydrometeor identification applied to dual-polarimetric Weather Surveillance Radar-1988 Doppler radar data collected at Vance Air Force Base, Oklahoma (KVNX). The goal of this work is to determine the signatures of various hydrometeor species in terms of BTs measured at frequencies used by the Global Precipitation Measurement mission Microwave Imager. Results indicate that hail is associated with an ice-scattering signature at all frequencies examined, including 10.7 GHz. However, it appears that frequencies ≤ 37.1 GHz are most useful for identifying hail. Low-level (below 2.5 km) hail becomes probable for a BT below 240 K at 19.4 GHz, 170 K at 37.1 GHz, 90 K at 85.5 GHz, 80 K at 89.0 GHz, 100 K at 165.5 GHz, and 100 K at 183.3 ± 7 GHz. Graupel may be distinguished from hail and profiles without any hydrometeor species by its strong scattering signature at higher frequencies (e.g., 165.5 GHz) and its relative lack of scattering at frequencies ≤ 19.4 GHz. There is a clearer distinction between profiles that contain liquid precipitation and profiles without any hydrometeors when the liquid is associated above with hail and/or graupel (i.e., a hydrometeor category with a strong scattering signature) than when the liquid is associated with smaller ice. Near-surface precipitation is much more likely for a 19.4-GHz BT < 250 K, 37.1-GHz BT < 240 K, 89.0-GHz BT < 220 K, and 165.5-GHz BT < 140 K.

Corresponding author address: Kenneth Leppert II, NSSTC, Rm. 4074, 320 Sparkman Dr. Huntsville, AL 35805. E-mail: leppert@nsstc.uah.edu

1. Introduction

Passive microwave instruments on satellites have long been used for precipitation retrieval (e.g., Wilheit et al. 1977), with higher-resolution airborne instruments used to better understand the characteristics of precipitation within storms (e.g., McGaughey et al. 1996). This continues with the Global Precipitation Measurement (GPM) mission and related field campaigns (Hou et al. 2014). The basic physics related to passive microwave retrieval involves emission by liquid hydrometeors (e.g., Wilheit et al. 1991) and scattering by ice hydrometeors (e.g., Ferraro and Marks 1995). The emission-based methods are based on the tendency of liquid precipitation to cause an increase in brightness temperature (BT) primarily at frequencies below 22 GHz over a background that is radiometrically cold, often an ocean background (e.g., Spencer et al. 1989; Adler et al. 1991; Wilheit et al. 1991; McGaughey et al. 1996). Over a radiometrically warm land surface, the absorption and emission by cooler raindrops reduces the BT in proportion to the rain amount and the temperature difference between the land surface and the raindrops. However, this small temperature difference leaves little signal for precipitation retrieval. Thus, a different type of retrieval must be used over land. In particular, scattering rainfall retrieval algorithms take advantage of scattering by precipitation-sized ice at higher frequencies (generally ≥37 GHz) that reduce the measured BTs above a relatively warm background (e.g., Spencer et al. 1983; Spencer and Santek 1985; Spencer et al. 1989; Smith et al. 1992; Ferraro and Marks 1995) and take advantage of relationships between ice and liquid water to estimate surface rainfall. Note that while the 19-GHz channel is mostly sensitive to emission by liquid water, it can also be sensitive to scattering by frozen hydrometeor species (e.g., Adler et al. 1991; Mugnai et al. 1993). Therefore, the 19-GHz frequency may be useful in both emission and scattering retrieval algorithms typically used over the ocean and land, respectively.

In addition to rainfall retrieval, passive microwave information has also been used for the detection of intense convection. Strong scattering and a significant reduction in observed BTs at 37 GHz require the presence of millimeter-sized particles in the upper portions of clouds (Toracinta et al. 2002). These large ice hydrometeors in the upper portions of clouds are generally associated with strong updrafts and intense convection to support the formation and maintenance of such ice. Spencer and Santek (1985) used the 37-GHz scattering signature from the Scanning Multichannel Microwave Radiometer (SMMR) to map the distribution of intense convection over land. Scattering at 85 GHz has also provided clues about the intensity of convection (e.g., Mohr and Zipser 1996a,b).

Another use for the 37-GHz scattering signature is for the detection of severe weather. In particular, Spencer et al. (1987) associated cold 37-GHz BTs from the SMMR with the occurrence of large hail, strong winds, and/or tornadoes/funnel clouds. Results of that study suggested that passive microwave information could have some utility in the detection and monitoring of severe weather given measurements with adequate spatial and temporal resolutions.

Other studies have specifically examined the relation between BTs at multiple frequencies and hail (e.g., Cecil 2009; Cecil and Blankenship 2012; Ferraro et al. 2015). For example, Cecil (2009) compared hail reports in the U.S. severe-storm database with Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) measurements at 19.35, 37.0, and 85.5 GHz. At all three frequencies examined, decreasing BTs were associated with an increasing probability of large hail at the surface. Specifically, hail reports occurred more often when BTs were <70, <180, and <230 K at 85.5, 37.0, and 19.35 GHz, respectively. In addition, Cecil (2009) found that the 37-GHz channel was best suited for hail detection from TMI. Cecil and Blankenship (2012) used the 36.5- and 89.0-GHz channels on the Advanced Microwave Scanning Radiometer for Earth Observing System and the 37.0- and 85.5-GHz channels on the TMI to develop a climatology for severe hailstorms. Similarly, Ferraro et al. (2015) developed a hail detection algorithm and hail climatology based on data from the Advanced Microwave Sounding Unit (AMSU).

The GPM mission expands upon the successful TRMM program (Kummerow et al. 1998) to provide global rainfall and snowfall observations every 3 h (Hou et al. 2014). The GPM mission consists of a Core Observatory (launched in February 2014) with a 65° inclined orbit at an altitude of 407 km and a constellation of additional microwave radiometers. The Core Observatory carries the Dual-Frequency Precipitation Radar and the GPM Microwave Imager (GMI). The GMI is a conically scanning microwave radiometer with 13 channels ranging from 10 to 183 GHz. The spatial resolution of the GMI ranges from 19.4 km × 32.2 km at 10.65 GHz to 4.4 km × 7.3 km at all frequencies at and above 89.0 GHz (Hou et al. 2014).

To aid in validation for GPM, several field experiments have been conducted, including the Midlatitude Continental Convective Clouds Experiment (MC3E) during April–June 2011 in Oklahoma. The goal of this paper is to determine the signatures from various hydrometeor species in terms of passive microwave BTs measured at frequencies used by GMI. This will be accomplished by comparing high-resolution, airborne passive microwave measurements collected during three MC3E case days with a hydrometeor identification (HID) applied to ground-based dual-polarimetric radar data. The case days were chosen based on the presence of strong/severe convection overflown by an ER-2 aircraft. These hydrometeor signatures could potentially be used to improve precipitation estimates via physical validation of the underlying models. The results presented here can serve as one observational baseline for assessing to what extent cloud resolving models coupled with radiative transfer models produce BTs that are consistent with observations. Inconsistencies in such an assessment could suggest changes to how the precipitation microphysics are handled in those models. Aside from the applicability to precipitation retrievals, the results here can aid in the interpretation of satellite signatures for potential studies of hazardous weather and climate.

2. Data and methodology

For this study, data collected during MC3E using the Advanced Microwave Precipitation Radiometer (AMPR), Conical Scanning Millimeter-wave Imaging Radiometer (CoSMIR), and dual-polarimetric radar data from the WSR-88D at Vance Air Force Base in Oklahoma (KVNX) are utilized. The AMPR instrument is a scanning radiometer with channels at 10.7, 19.35, 37.1, and 85.5 GHz (Spencer et al. 1994). The instrument scans ±45° in the cross-track direction, and the polarization of the measurements varies with scan angle. In particular, measurements vary from being fully horizontally polarized to the far right of the cross-track scan to being fully vertically polarized to the far left of the scan. At the 20-km flight altitude of the ER-2, the horizontal resolution of AMPR is 2.8 km at the two lowest frequencies, 1.5 km at 37.1 GHz, and 0.64 km at 85.5 GHz. Note that some of the AMPR data (especially 85.5 GHz) tended to get noisy as each flight progressed during MC3E, but this noise has relatively little impact on the results of this study. For more information regarding AMPR, see Spencer et al. (1994).

The CoSMIR instrument (Wang et al. 2008) is a nine-channel radiometer that can scan conically up to an angle of 53.6°, across track, or perform a scan pattern that includes both conical and cross-track components. At the 20-km altitude of the ER-2, CoSMIR has a 1.4-km horizontal resolution at nadir. The CoSMIR channels most relevant for the GMI are both polarizations of the 89.0- and 165.5-GHz channels as well as the 183.3 ± 3- and 183.3 ± 7-GHz channels (Hou et al. 2014). Here, we focus only on the horizontally polarized 89.0-, 165.5-, and 183.3 ± 7-GHz channels because the vertically polarized 89.0-, 165.5-, and horizontally polarized 183.3 ± 3-GHz channels provided similar results. Note that hereinafter all frequencies will be referred to by their integer value (e.g., 165.5 GHz will be referred to as 165 GHz) for conciseness. While the GMI is a conically scanning instrument, we only use cross-track data from CoSMIR here to better match the geometry of a vertical profile of ground-based radar data.

Dual-polarimetric radar data from the S-band KVNX radar was converted from its native polar coordinates to a Cartesian grid using the REORDER software package (Oye et al. 1995). The Cartesian grid spanned an area of 600 km × 600 km centered on the radar and stretched from the surface to a height of 16 km with a horizontal (vertical) resolution of 1.0 (0.5) km. Such a large horizontal grid was used in order to maximize the sample size. Data up to 377 km from the radar were used in the analysis, and the 1° beamwidth of KVNX results in a beam 6.6 km in diameter at this distance. This beam size is too large to trust details of the vertical profiles at such large distances. However, we focus on whether or not a hydrometeor species occurs anywhere in the vertical column, and the HID used (described below) gives credible results when used in this manner for the selected cases.

Next, a fuzzy-logic HID based on Liu and Chandrasekar (2000) and Dolan and Rutledge (2009) was applied to the gridded radar data. Dolan and Rutledge (2009) used scattering simulations to develop one-dimensional membership beta functions for various hydrometeor species and observed variables, and these beta functions were used here. The beta functions provide a score from 0.0 to 1.0 that indicates how well each observation represents each hydrometeor type. A score of 0.0 indicates the observation does not correspond with a given hydrometeor type at all, while a score of 1.0 indicates that an observation matches a given hydrometeor type perfectly. Ten hydrometeor types were used here including drizzle (DZ), rain (RN), ice crystals (IC), aggregates (AG), wet snow (WS), vertically oriented ice (VI), low-density graupel (LG; density ≤0.55 g cm−3), high-density graupel (HG; density >0.55 g cm−3), hail (HL), and big drops [BD, resulting from melting hail or graupel; Gatlin et al. 2015)]. Four radar observables [reflectivity (Zh), differential reflectivity (Zdr), specific differential phase (Kdp), and correlation coefficient (ρhv)] and temperature were used in the HID. Temperature at each height of the gridded KVNX dataset was vertically interpolated from the nearest-in-time sounding taken at Lamont, Oklahoma. Next, at each grid point, a beta function value was calculated using each observed variable separately for each hydrometeor type. Following Dolan and Rutledge (2009), the beta function values for Zdr, Kdp, and ρhv were weighted by 0.8, 0.8, and 0.4, respectively, and summed for each hydrometeor species. Then, this weighted sum was multiplied by the beta function value for temperature and Zh to provide a single combined beta function value for each hydrometeor type at each grid point, and the species with the greatest combined value was assigned to each grid point. Note that this HID assigns a single dominant hydrometeor category to each grid point while additional less dominant species may still be present in the grid cell.

After applying the HID to the KVNX data, the mass of the assigned hydrometeor category at each point was calculated following the relations found in Straka et al. (2000) and references therein. The specific relations used and the conditions under which they are applied are provided in Table 1. Some variables shown in Table 1 but not previously defined include Zυ, which is the reflectivity at vertical polarization, and C, which is a constant that depends on radar wavelength [for KVNX, C equals 0.48 (Straka et al. 2000)]. In the mass relation for HG and hail, ρ is the density of hail assumed to be 900 kg m−3 from Straka et al. (2000), ρa is the air density calculated from sounding data, R is the hail rate (equivalent to rain rate except for hail), and V is the terminal fall velocity. Both R and V were calculated using the relations from Table 1 in Straka et al. (2000). After the mass of each hydrometeor category was calculated, the column-integrated mass was calculated by integrating the mass of each species in the vertical.

Table 1.

The equations used to calculate the mass (M; g m−3) of each hydrometeor species. The conditions under which each equation is applied and the applicable reference are also given. Note that the mass relations for AG, LG, HG, and HL are not conditional on the values of any of the radar observables. Logarithmic units are used for Zh and Zdr when assessing the conditions under which to use each mass relation but are converted to linear units when calculating the mass. See the text for more information.

Table 1.

To compare the HID with the passive microwave measurements from AMPR and/or CoSMIR, the passive microwave pixels closest to nadir were matched with vertical profiles of gridded KVNX data and the associated HID. The matching procedure was performed by selecting the single KVNX volume scan that began closest in time to each passive microwave observation within the domain of the gridded radar data. Then, the gridded radar profile that was closest to the latitude and longitude of the BT pixel was selected as the matched profile. Note that we also applied the HID and matching procedure to the NASA S-band dual-polarimetric (NPOL) radar, but the number of matched profiles was much greater for KVNX than NPOL on the selected case days. Therefore, to take advantage of this larger sample size and the more robust results it can provide, only results from KVNX are shown here.

Figure 1 shows a map of composite gridded Zh and associated HID valid for 2144 UTC 24 May 2011. The map of HID shows the hydrometeor category with the greatest mass in each vertical column. As can be seen from Fig. 1, the convection on this day was organized into a squall-line-like structure to the south oriented roughly north–south along 98°W at this time with trailing discrete convective cells to the northwest. For information on the environmental conditions associated with this case, see Heymsfield et al. (2013). As expected, hail and HG dominate in the convective cores while VI and aggregates dominate in the leading stratiform/anvil regions.

Fig. 1.
Fig. 1.

(a) Composite gridded KVNX reflectivity valid at 2144 UTC 24 May 2011 and (b) corresponding hydrometeor identification showing the hydrometeor with the greatest mass in each vertical column. The hydrometeors, as defined in the text, are DZ, RN, IC, AG, WS, VI, LG, HG, HL, and BD. The black line in each panel indicates the location of the cross section shown in Fig. 2, and the plus symbol shows the start position of each cross section. The KVNX radar is located near the center of each panel.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Vertical cross sections of Zh and the corresponding HID taken along the black curves in Fig. 1 are shown in Fig. 2. Note that the decrease in height of the valid Zh and HID around 2154 UTC is due to the aircraft moving within ~9 km of the radar. As expected, hail and graupel are predominantly found in the Zh convective cores with rain and/or big drops below. The curves plotted over the Zh (HID) cross section indicate corresponding AMPR (CoSMIR) BTs. The AMPR frequencies are quite noisy, especially the two highest frequencies. Even when the aircraft turned, the pixel closest to nadir was matched with each radar profile. Hence, this noise should not be due to aircraft movement. In general, the expected decrease in BT (i.e., scattering signature) of hail and graupel in the convective cores is observed. One obvious exception to this occurs near 2139 UTC where BTs at frequencies ≥ 19 GHz actually peak where the maximum Zh > 56 dBZ and a hail core extends down to a height of 2 km. As can be seen from Fig. 1, the aircraft flew over the edge of this convective core over a strong reflectivity gradient. Given the different temporal resolution of the BT and radar data, it is possible that the Zh and HID profiles indicated in the cross sections in Fig. 2 are not representative of the area sampled by AMPR and CoSMIR near 2139 UTC. To investigate this particular cross section, we tried offsetting the radar–radiometer matchups by ±1 km and ±1 radar volume scan. It turns out that using the previous volume scan (not shown) does give a much better match for this particular case. The previous volume scan sampled the 8-km altitude in the region of interest about 2 min before the radiometer observation, and the next volume scan (the one used in Fig. 2) sampled the 8-km altitude about 3 min after the radiometer observation. The volume scan used for Fig. 2 is used because it began closest in time to the radiometer measurement.

Fig. 2.
Fig. 2.

Cross section of (a) gridded KVNX reflectivity along a 177-km ER-2 flight segment beginning at 2137 UTC 24 May 2011 and (b) corresponding hydrometeor identification. The location of the cross section is shown in Fig. 1 by the black line, and the plus symbol corresponds to the left side of each cross section. The lines plotted over (a) show the collocated AMPR BTs and over (b) show the collocated CoSMIR BTs. Hydrometeor categories are the same as in Fig. 1, and the triangle in each panel represents the lack of radar coverage associated with the radar’s cone of silence.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Another case of intense convection occurred on 23 May 2011, and an example of composite gridded Zh and associated HID valid for 2145 UTC 23 May 2011 is shown in Fig. 3. No attempt was made to remove artifacts from the radar data such as ground clutter. However, most of the matched radar profiles contain valid meteorological radar echo or occur far enough from the radar where ground clutter and its associated invalid HID are not an issue. Thus, there should be little if any impact on the results from an invalid HID as a result of ground clutter.

Fig. 3.
Fig. 3.

As in Fig. 1, but valid at 2145 UTC 23 May 2011. The black line in each panel corresponds to the cross sections shown in Fig. 4, and the asterisk symbol in each panel corresponds to the location of the KVNX radar.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Figure 4 shows vertical cross sections of Zh and the corresponding HID taken along the black lines shown in Fig. 3. Similar to what is observed in Fig. 2, hail and graupel are predominantly found in the Zh convective cores with rain and/or big drops below, as expected. Not surprisingly, minima in CoSMIR BTs (AMPR BTs were not available on this day) shown by the lines plotted over the Zh cross section generally correspond with Zh cores that contain hail and/or graupel. Other noteworthy observations from Fig. 4 are that, in between the convective cores, wet snow tends to be affiliated with the melting level, aggregates dominate between the freezing level and ~11 km, and VI tends to occur above 11 km.

Fig. 4.
Fig. 4.

Cross section of (a) gridded KVNX reflectivity along a 229-km ER-2 flight segment beginning at 2135 UTC 23 May 2011 and (b) corresponding hydrometeor identification. The location of the cross section is shown in Fig. 3 by the black line, and the plus symbol in Fig. 3 corresponds to the left side of each cross section. The lines plotted over (a) show the collocated CoSMIR BTs, and stippling in (b) indicates hydrometeor mass shown by the key to the right of (b). Hydrometeor categories are the same as in Fig. 1.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

In addition to the 23 and 24 May cases, the case on 22 April was also examined, and Fig. 5a shows a map of composite Zh valid for 2251 UTC 22 April 2011. This case occurred farther from the radar than the May cases so that the low levels were not sampled as well on 22 April. This case was chosen because it provided another example of intense convection, which happened to be in the form of a squall line (Fig. 5a). Note that AMPR data were only available for the 22 April and 24 May cases, while CoSMIR data were only available for the 23–24 May cases.

Fig. 5.
Fig. 5.

(a) Composite gridded KVNX reflectivity valid at 2251 UTC 22 Apr 2011 and (b) ER-2 flight track on 22 Apr 2011. The asterisk in (a) indicates the location of the KVNX radar and in (b) indicates the start of the flight track. Portions of the flight track in (b) colored red (blue) correspond to collocated vertical profiles not associated with any hydrometeors and a 10-GHz BT > (≤) 285 K.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

3. Results

Most of the vertical profiles of radar data contain multiple hydrometeor types as shown, for example, in Fig. 2b, and some types would be expected to have a stronger signal than others at the frequencies examined. For example, the 37-GHz BT measured above a profile with small ice crystals situated above hail should be dominated by the scattering signal of hail. Therefore, to minimize the effect of the signal from one hydrometeor species dominating the signal from other species and better isolate the signal from each species separately, a subjective hierarchy of hydrometeor categories was applied. Each hydrometeor type was assigned a certain priority roughly based on its radiative impact due to particle size, density, and vertical location. The type with the greatest priority in each profile was assigned to that column. The big drops category was given the highest priority followed by hail, HG, LG, rain, wet snow, aggregates, ice crystals (which were combined with VI), and drizzle. In this way, all profiles with big drops were assigned that category regardless of the other hydrometeor species present in the column, while profiles assigned to drizzle only contained that hydrometeor type. The numbers of profiles matched with the CoSMIR and AMPR instruments and assigned to each hydrometeor type after application of the hierarchy are shown in Table 2. Note that except for the big drops category, the priorities were assigned to each hydrometeor species such that species with stronger (weaker) scattering signatures were assigned higher (lower) priorities. All profiles with big drops also contain hail and/or graupel. Thus, ranking big drops behind hail or graupel in the hierarchy would have effectively eliminated this category, with all the relevant profiles being assigned to hail or graupel instead.

Table 2.

Numbers of vertical profiles that are matched with AMPR and CoSMIR data and are labeled with each hydrometeor category after application of the hydrometeor hierarchy discussed in the text. The none category refers to profiles without any identified hydrometeor species.

Table 2.

Figure 6 shows the probability density function (PDF) of BTs associated with profiles that contain hail, HG, and LG, as well as those with no identified hydrometeor types using AMPR channels. Given that a profile contains hail or graupel, Fig. 6 shows the probability that the profile will be associated with a certain BT. The 10-GHz channel has been shown to respond primarily to emission from liquid precipitation (e.g., Adler et al. 1991; McGaughey et al. 1996). However, Fig. 6a clearly shows a scattering signature for hail at 10 GHz where the distribution for hail is shifted to colder BTs relative to the distribution for profiles without any hydrometeors. At all frequencies, the presence of hail extends to colder BTs than do the other categories. In addition, there appears to be a tendency for HG to show a greater scattering signature than LG, but there is still quite a bit of overlap between the two PDFs.

Fig. 6.
Fig. 6.

BT PDFs for vertical profiles that contain HL, HG, and LG valid at (a) 10, (b) 19, (c) 37, and (d) 85 GHz. The BTs were collected by AMPR on 22 Apr 2011 or 24 May 2011. The “none” category includes profiles without any valid hydrometeor identification. For (a) and (b), the BT bin size is 10 (5) K to the left (right) of the thin vertical line. For (c) and (d) the bin size is 20 (10) K to the left (right) of the thin vertical line. The minimum BT measured at each frequency for the shown hydrometeor species is indicated by the number in the top-left corner of each panel.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

The double-peak shown for the none category at the lower frequencies in Fig. 6 may be due to different surface emissivities as a result of differing levels of soil moisture. Most of the profiles without any hydrometeors (1000 of 1187) were observed on 22 April 2011, and Fig. 5b shows the ER-2 flight track within the KVNX domain on that day. The locations of profiles without any identified hydrometeors associated with a 10-GHz BT ≤ 285 K (the cooler peak observed for the none category in Fig. 6a) are colored blue, while the locations of profiles associated with a 10-GHz BT > 285 K (the warmer peak in Fig. 6a) are colored red. Figure 5 indicates that most of the profiles associated with the cooler peak occurred to the north of the squall line where it had recently rained. In contrast, profiles associated with the warmer peak mainly occurred to the west of the squall line where it may not have recently rained. Note that the distribution for the none category does not show a double-peak structure at 37 or 85 GHz. Mugnai et al. (1993) found that the 37- and 85-GHz channels are less sensitive to surface emission when observing the surface through precipitation. However, the profiles used for the none PDF correspond to profiles through clear air. Nevertheless, the lack of a double-peak structure in the distribution for the none category at 37 and 85 GHz suggests that these higher frequencies are less sensitive to variations in surface wetness/emissivity even in scenes without precipitation.

Table 3 lists the most common BT values associated with each hydrometeor species. All hydrometeor categories except for hail, HG, and big drops (which almost always include hail or HG) are most often associated with a 10-GHz BT in the 280s K, suggesting little sensitivity to different hydrometeor types at the low frequency. High-density graupel does typically produce a 10-GHz scattering signature, with about a 10-K depression compared to scenes without HG. The presence of hail and big drops further reduces the 10-GHz BT. Moving to higher frequencies (shorter wavelengths) in Table 3, the magnitudes of the BT depressions generally increase, and sensitivity to progressively more hydrometeor types is seen. Low-density graupel usually accomplishes some BT reduction in the 19- and 37-GHz channels (but not 10 GHz), and greater reductions in the higher-frequency channels. Rain (without graupel or hail above) corresponds with large BT reductions in the CoSMIR channels (89, 165, and 183 GHz). Wet snow (melting from aggregates above) is associated with large ice-scattering signatures in the highest-frequency (165 and 183 GHz) channels.

Table 3.

The BT bin (i.e., range; K) that occurs most frequently for each hydrometeor species as a function of passive microwave channel.

Table 3.

The BT PDFs associated with profiles containing hail and graupel using CoSMIR frequencies (Fig. 7) show little distinction between the different hydrometeor categories. However, the distributions for hail, LG, and HG all show clear separation from the distribution for profiles without any hydrometeors. In particular, all the PDFs for the hydrometeor species are shifted to much colder temperatures than the PDF for the none category due to significant scattering. The modal (i.e., most common) BTs (Table 3) for hail and graupel range from 110 to 170 K at 89 GHz, 120 to 170 K at 165 GHz, and 130 to 160 K at 183 GHz. In contrast, the modal BTs for the none category are 285–290 K at 89 and 165 GHz and 270–275 K at 183 GHz.

Fig. 7.
Fig. 7.

BT PDFs for vertical profiles that contain HL, HG, LG, and no valid hydrometeor identification (none) valid at (a) 89, (b) 165, and (c) 183 GHz. The BTs were measured with CoSMIR on 23–24 May 2011. The BT bin size is 10 (5) K to the left (right) of the thin vertical line in each panel. The values along the left (right) side of the plot correspond to the hail and graupel (none) categories. The minimum BT measured at each frequency for the shown hydrometeor species is indicated by the number in the top-left corner of each panel.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Figure 8 shows the BT PDFs associated with profiles that contain small ice (wet snow, aggregates, and ice crystals) but no graupel or hail using AMPR channels. Because of the relatively low density and/or small size, little scattering may be expected by small ice species at the frequencies measured by AMPR. Indeed, at 10 and 19 GHz, there is virtually no separation between the distributions for small ice and profiles without any hydrometeors. At 37 GHz, there is a shift in the PDFs for small ice to colder BTs (distributions for all small ice categories nearly coincide and peak near 285 K) relative to that for the none category (peak in the distribution near 305 K), but there is still considerable overlap between the hydrometeor PDFs and the PDF for the none category. At 85 GHz, the distribution for profiles with wet snow and those without any hydrometeors are very similar, while the distributions for ice crystals and especially aggregates are shifted slightly to cooler BTs relative to the none category. The peaks of the PDFs for none and snow occur for BTs of 290–300 K, while the peaks for ice crystals and aggregates occur for BTs of 280–290 and 260–270 K, respectively, at 85 GHz (Fig. 8; Table 3).

Fig. 8.
Fig. 8.

As in Fig. 6, but for profiles containing WS, AG, small IC, and no valid hydrometeor identification.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

At CoSMIR frequencies, the BT PDFs for small ice hydrometeor types tend to peak near the peak in the PDF for profiles without any hydrometeors (Fig. 9), while tails of the hydrometeor species PDFs extend to colder BTs than that for the none category. Thus, at CoSMIR frequencies, there is relatively little distinction between wet snow, aggregates, small ice crystals, and profiles without any hydrometeors. Because the AMPR 85-GHz channel is close in frequency to the CoSMIR 89-GHz channel, it may be expected to see similar results in these channels. However, at 85 GHz (Fig. 8d), the PDFs for ice crystals and aggregates are shifted to cooler BTs relative to the PDF for profiles without any hydrometeors, while no such shift is observed at 89 GHz (Fig. 9a). The differences between these two frequencies appear to be a result of different samples used for the AMPR (data from 22 April and 24 May 2011) and CoSMIR (data from 23 to 24 May 2011) figures. An analysis of the BT PDFs associated with various hydrometeor species valid at 85 and 89 GHz using only data on 24 May when both CoSMIR and AMPR data were available indicates similar patterns at the two frequencies. For example, the BT PDFs shown in Fig. 10 using data valid on 24 May only indicate that profiles with wet snow, aggregates, ice crystals, and without any hydrometeor types have peak values near 235, 275, 295, and 285 K, respectively, at 85 GHz. The corresponding 89-GHz values are 245, 285, 285, and 285 K, which are within one BT bin of the 85-GHz values.

Fig. 9.
Fig. 9.

As in Fig. 7, but for profiles containing WS, AG, small IC, and no valid hydrometeor identification.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Fig. 10.
Fig. 10.

BT PDFs for vertical profiles that contain WS, AG, IC, and no valid hydrometeor identification (none) valid at (a) 85 and (b) 89 GHz on 24 May 2011. The BT bin size is 20 (10) K to the left (right) of the thin vertical line in each panel.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

For each case day, PDFs for each hydrometeor type vary somewhat because of such factors as differences in the emission characteristics of the background land and different masses of the hydrometeor species. But, regardless of which case day is chosen, hail always shows a stronger scattering signature than both graupel categories at 37 GHz, for example. In contrast, the comparison between Figs. 8d and 9a indicates that the relative scattering signature among the small ice species does vary from day to day. Thus, the relative scattering signature among the small ice species may not be generalizable to more times and places, but the relative scattering of the other species could potentially be more broadly generalized.

The BT PDFs for profiles that contain liquid hydrometeor types, except for drizzle, are shown in Fig. 11 for AMPR frequencies. Drizzle profiles are not shown because the 27 drizzle profiles (Table 2) appear to contain ground clutter misclassified as drizzle. The PDFs for big drops in Fig. 11 are very similar to the PDFs for hail shown in Fig. 6 because profiles containing big drops also contain large ice. Big drops form from large ice that subsequently melts. Recall that big drops were assigned the highest priority in the hydrometeor hierarchy used for the PDF figures, and, therefore, profiles with big drops may contain any of the other hydrometeor types as well. Thus, the shift in the PDF for profiles with big drops to colder BTs relative to the PDF for profiles without any hydrometeors and relative to that for rain profiles at all AMPR frequencies is likely due to scattering from large ice.

Fig. 11.
Fig. 11.

As in Fig. 6, but for profiles containing BD, RN, and profiles containing no valid hydrometeor identification.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Because the data collected here are over land, it is not surprising that the emission characteristics of the lower frequencies provide little distinction between PDFs for profiles with and without rain, as is seen in Fig. 11. However, the ice-scattering signature at 85 GHz has been shown to be useful over land for rainfall retrieval (e.g., Spencer et al. 1989; Ferraro and Marks 1995). This retrieval tends to work best when there is a strong scattering signature such as from hail or graupel. Based on the methodology used here, the rain profiles contain no such large ice, but big drop profiles do contain such ice. Hence, at 85 GHz (Fig. 11d), the distribution for rain shows a slight shift to colder BTs relative to the distribution without any hydrometeors, but the big drops distribution shows a much larger shift (peak for big drops occurs near 90 K while peak for none occurs at 295 K). Therefore, BTs measured at all AMPR frequencies may be useful for identifying liquid precipitation that forms from melting hail or graupel over land, but less effective in identifying rain that is not associated with frozen hydrometeor types that exhibit a strong scattering signature.

In contrast to what is observed at AMPR frequencies, the BT PDFs for the big drop and rain profiles show clear separation from that of profiles without any hydrometeors at all CoSMIR frequencies (Fig. 12). Considering the similarities between Figs. 9 and 12, this rain signature in the CoSMIR frequencies is likely due to scattering by aggregates and wet snow, which melts and falls as rain. Figure 12 and Table 3 indicate that the PDF for big drops peaks for BTs of 140–150, 130–140, and 120–130 K at 89, 165, and 183 GHz, respectively, while the peak of the rain PDF occurs at warmer BTs (i.e., 215–220, 170–180, and 180–190 K at 89, 165, and 183 GHz, respectively). In contrast, the PDF for profiles without any hydrometeors peaks for BTs 285–290 K at the two lower CoSMIR frequencies and 270–275 K at 183 GHz.

Fig. 12.
Fig. 12.

As in Fig. 7, but for profiles containing BD, RN, and profiles containing no valid hydrometeor identification.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Figure 13 shows the probability of each hydrometeor species in a vertical column given a particular BT measured at each AMPR frequency. At the lower BTs for each frequency, the probabilities of big drops, hail, and HG dominate. At warmer BTs, rain and small ice species, especially aggregates, become more probable. The relatively high probabilities of LG and big drops for 85-GHz BTs ≥ 310 K are probably related to the noise issues that affected the higher-frequency AMPR channels, as mentioned previously. Thus, the few 85-GHz BTs ≥ 310 K are likely not accurate.

Fig. 13.
Fig. 13.

Probability of BD, HL, HG, LG, RN, WS, AG, IC, or DZ being the dominant hydrometeor species from our hierarchy as a function of collocated (a) 10-, (b) 19-, (c) 37-, and (d) 85-GHz BT measured with AMPR on 22 Apr or 24 May 2011. The BT bin sizes are as in Fig. 6.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

The probabilities of each hydrometeor species given a particular BT measured at each CoSMIR frequency are shown in Fig. 14. Similar to what is observed at AMPR frequencies in Fig. 13, at the lowest BTs, the probabilities of big drops, hail, and HG are largest. Toward the middle range of BTs at each CoSMIR frequency, LG and rain achieve their greatest probabilities. Finally, at the warmest BTs, wet snow and aggregates become most probable.

Fig. 14.
Fig. 14.

Probability of BD, HL, HG, LG, RN, WS, AG, IC, or DZ being the dominant hydrometeor species from our hierarchy as a function of collocated (a) 89-, (b) 165-, and (c) 183-GHz BT measured with CoSMIR on 23–24 May 2011. The BT bin sizes are as in Fig. 7.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Figures 13 and 14 show probabilities of each hydrometeor species after application of the hydrometeor hierarchy. In contrast, Fig. 15 shows the probability of certain hydrometeor types given a particular BT measured at each AMPR frequency without the hierarchy (note the hydrometeor hierarchy is not applied to any successive figure either). Hence, Fig. 15 shows the probabilities of hail, graupel, and big drops without regard to which other hydrometeor species may be present in the column. In general, profiles having LG are most common, and those having big drops are least common in these cases, in contrast to what is observed at the lowest BTs in Fig. 13. Figure 15 indicates that the probability of hail rapidly increases above zero below BTs of 275, 265, 265, and 225 K at 10, 19, 37, and 85 GHz, respectively. Hail becomes most likely (probability ≥ 50%) below BTs of 240, 235, 195, and 145 K at 10, 19, 37, and 85 GHz, respectively (the decrease in hail probability below 50% below a 10-GHz BT of 220 K and near a 37-GHz BT of 100 K is likely a result of small sample size).

Fig. 15.
Fig. 15.

Probability of a vertically integrated nonzero mass of HL, HG, LG, and BD as a function of collocated AMPR BT at (a) 10, (b) 19, (c) 37, and (d) 85 GHz for data collected on 22 Apr or 24 May 2011. The thin gray line indicates the number of profiles in each BT bin as indicated by the axis along the right side of each panel. The thin vertical lines and BT bin sizes relative to those lines are the same as in Fig. 6.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Figure 16 is similar to Fig. 15, but for CoSMIR channels. Over the range of BTs at each frequency, the probability of LG is greater than that of HG, which is generally greater than that of hail and big drops. The probability of hail and big drops are similar across the range of BTs except below 135 K at 165 and 183 GHz where the probability of hail is greater. Consistent with the pattern at 85 GHz (Fig. 15d), the probability of hail increases above zero below 225 K at 89 GHz (Fig. 16a). A rapid increase in the probability of hail occurs below a colder BT (i.e., 155 K) for the other two CoSMIR frequencies.

Fig. 16.
Fig. 16.

Probability of a vertically integrated nonzero mass of HL, HG, LG, and BD as a function of collocated CoSMIR BT at (a) 89, (b) 165, and (c) 183 GHz for data collected on 23–24 May 2011. The thin gray line indicates the number of profiles in each BT bin as indicated by the axis along the right side of each panel. The thin vertical lines and BT bin sizes relative to those lines are the same as in Fig. 7.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Figure 17 shows the probability that precipitation (liquid or solid) is present at low levels (<2.5-km height) as a function of BT at various AMPR frequencies. Figure 17 is somewhat noisy because of small sample sizes. However, the probability of low-level precipitation increases rapidly below 290 K at frequencies ≤37 GHz and below 280 K for 85 GHz. With decreasing BT, the probabilities of low-level precipitation initially reach 100% near 230, 250, 240, and 230 K at 10, 19, 37, and 85 GHz, respectively.

Fig. 17.
Fig. 17.

Probability of a nonzero mass of any hydrometeor type at low levels (≤2.5-km height) as a function of collocated AMPR BT at various frequencies for data collected on 22 Apr or 24 May 2011 within 180 km of the radar. The blue dotted line shows the number of profiles in each 85-GHz BT bin as indicated by the axis along the right side of the figure (counts for other frequencies are not shown for clarity). The colored asterisks at the bottom of the figure indicate the BT at which the bin sizes change for each frequency, and bin sizes relative to each symbol are the same as in Fig. 6.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

The probability of low-level precipitation as a function of CoSMIR BTs shown in Fig. 18 indicates similar patterns to those seen at each AMPR frequency (Fig. 17). In particular, probabilities rapidly increase below a BT of 280 K at 89 GHz and 270 K at the two higher frequencies. Below an 89-GHz BT of 220 K, the probability of low-level precipitation is ~100% which is comparable to the corresponding 85-GHz BT (230 K; Fig. 17), as expected. At 165 and 183 GHz, the 100% probability occurs at a colder BT (i.e., <140 K).

Fig. 18.
Fig. 18.

Probability of a nonzero mass of any hydrometeor type at low levels (≤2.5-km height) as a function of collocated CoSMIR BT at various frequencies for data collected on 23–24 May 2011 within 180 km from the radar. The green dotted line shows the number of profiles in each 183-GHz BT bin as indicated by the axis along the right side of the figure (counts for other frequencies are not shown for clarity). To the left (right) of the thin vertical line, the BT bin size is 10 (5) K.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

The probability of occurrence of hail at low levels (<2.5-km height) and upper levels (>2.5-km height) as a function of BT at each AMPR frequency is shown in Fig. 19. For a given BT at each frequency, the probability of hail at upper levels is generally larger than that of hail at low levels. Hail that reaches near the surface may be expected to be associated with a deeper column of hail and/or larger hail (less likely to melt) than hail that occurs only at upper levels, resulting in greater scattering of upwelling radiation. This may help to explain why nonzero probabilities of low-level hail begin at colder BTs than for upper-level hail at all frequencies, except 10 GHz. At 10 GHz, trends/patterns shown by the probability of upper- and low-level hail are generally similar except that the probability of low-level hail is slightly lower than the corresponding upper-level hail value. At 19, 37, and 85 GHz, the probabilities of low-level hail rapidly increase above zero with decreasing BTs below 250, 210, and 180 K, respectively. The corresponding values for upper-level hail are 270, 240, and 270 K.

Fig. 19.
Fig. 19.

Probability of the occurrence of HL at low levels (i.e., ≤2.5 km) or at upper levels (i.e., >2.5 km) in a vertical column as a function of collocated AMPR BT at (a) 10, (b) 19, (c) 37, and (d) 85 GHz for data collected on 22 Apr or 24 May 2011 within 180 km of the radar. The BT bin size in (a) is 10 K. In (b)–(d) the BT bin size is 20 (10) K to the left (right) of the thin vertical line.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

Cecil (2009) found that surface hail reports are common when TMI BTs are <230, <180, and <70 K at 19, 37, and 85 GHz, respectively. However, the TMI sampling volume is ~50–100 times larger than that of AMPR at each frequency. Beam-filling effects should artificially increase the TMI BTs over convective cores, especially for the lower-frequency channels. The probabilities found here (Fig. 19) for low-level hail vary with decreasing BT for the coldest BTs likely because of the small sample size used (27 profiles with low-level hail). This makes it difficult to determine a threshold BT at each frequency below which low-level hail is more likely. However, with decreasing BT, the first substantial (>0.3) peaks in probability of low-level hail occur near 240, 170, and 90 K at 19, 37, and 85 GHz, respectively. For a larger sample size, it may be expected that the probability of low-level hail would remain relatively high for BTs below these peaks. Surprisingly, these AMPR-derived BT thresholds for hail compare reasonably well with those from Cecil (2009). While beam-filling effects should result in higher BT thresholds for TMI than AMPR, differences in the methodologies used by Cecil (2009) and that used here may at least partially offset these effects. Specifically, Cecil (2009) compared reports of hail ≥ 0.75 in. with TMI BTs, while no size threshold was used here. Thus, smaller hail associated with less scattering may be included in the sample here.

The probability of low- and upper-level hail as a function of BT at each CoSMIR frequency given in Fig. 20 shows patterns similar to what is observed at the various AMPR frequencies. For example, the probability of low-level hail is generally less than the corresponding upper-level hail probability. In addition, probabilities of low-level hail become greater than 0.1 at cooler BTs (80, 100, and 100 K at 89, 165, and 183 GHz, respectively) than for upper-level hail (210, 160, and 160 K for 89, 165, and 183 GHz, respectively). These thresholds for low-level hail can be compared with the hail thresholds used in Ferraro et al. (2015) of 228.2, 206.9, and 204.6 K valid at 89, 150, and 183 GHz, respectively. The finer spatial resolution of the airborne CoSMIR instrument used here may be expected to resolve colder BTs due to scattering from hail than the spaceborne AMSU used in Ferraro et al. (2015). Another possible reason for the differences between hail thresholds found here and those of Ferraro et al. (2015) are due to the use of different samples. The sample used here is valid for intense convection for only 2 days over Oklahoma while the sample used in Ferraro et al. (2015) was valid over the entire continental United States for 12 yr.

Fig. 20.
Fig. 20.

Probability of the occurrence of HL at low levels (i.e., ≤2.5 km) or at upper levels (i.e., >2.5 km) in a vertical column as a function of collocated CoSMIR BT at (a) 89, (b) 165, and (c) 183 GHz for data collected on 23–24 May 2011 within 180 km of the radar. The BT bin size is 20 (10) K to the left (right) of the thin vertical line in each panel.

Citation: Journal of Applied Meteorology and Climatology 54, 6; 10.1175/JAMC-D-14-0145.1

As an initial examination of the impacts of nonuniform beam filling for coarser-resolution satellite data, we performed a simple average of AMPR and CoSMIR BTs over footprint sizes representative of GMI. For simplicity, we assumed that the high-resolution pixel of interest (i.e., pixel associated with hail) occurred at the center of the simulated GMI footprint and did not account for details such as antenna beam pattern or incidence angle. Table 4 lists the minimum BTs associated with hail for each frequency for both the high-resolution airborne and coarser-resolution simulated GMI BTs. At 85 and 37 GHz, the minimum high-resolution BTs are much lower than the corresponding satellite-resolution BTs (25 versus 52 K for the 85-GHz channel and 56 versus 122 K for the 37-GHz channel). These satellite-resolution values are consistent with a separate survey of satellite observations, but the high-resolution values are substantially lower than what has been noted from satellites. Admittedly, there are questions about the absolute calibrations of airborne radiometers at such low values.

Table 4.

Minimum and median BTs using high-resolution airborne data from the AMPR or CoSMIR instrument and using coarser-resolution-simulated GMI BTs associated with vertical profiles containing HL as a function of various frequencies. In addition, the final column refers to the median standard deviation of high-resolution BTs within each GMI footprint.

Table 4.

The median high-resolution BTs are also lower than the median coarse-resolution BTs for hail-related pixels in Table 4, except at the higher frequencies where the GMI footprint size becomes more comparable to the size of a severe convective cell. In addition, the standard deviation of the high-resolution BTs was computed for each of the coarse-resolution footprints. The median of these standard deviations is also listed for each frequency in Table 4. There are two primary influences on this value. The first is that GMI pixel size decreases with frequency from 10 to 89 GHz and then levels off with further increases in frequency (Hou et al. 2014). A larger field of view may be expected to have greater variability within that field of view. The second major contributor to variability within a GMI pixel is the dynamic range of BTs observed at each frequency. As seen in Figs. 6 and 7, for example, a smaller range of BTs is observed for the low-frequency channels than for the high-frequency channels. This yields smaller standard deviations for the lowest-frequency channels. The influence of these two factors results in an increase in median standard deviation with frequency up to 37 GHz (41 K) and a decrease in standard deviations for the higher frequencies (~10 K; Table 4).

4. Summary and conclusions

In this paper, we compared airborne passive microwave BTs collected over intense convection during the MC3E over Oklahoma during 2011 with an HID applied to ground-based dual-polarimetric radar data. The purpose of this work is to determine the signatures of various hydrometeor species in terms of passive microwave BTs measured at frequencies used by GMI.

Results indicate a strong scattering signature of hail at all frequencies examined, including 10 GHz. Specifically, the BT PDFs for vertical profiles containing hail show a clear shift to cooler BTs relative to the PDF for profiles without any hydrometeors. Previous work has emphasized the sensitivity of the 10-GHz channel to emission from liquid hydrometeors (e.g., McGaughey et al. 1996). However, results here confirm that significant ice scattering can occur at this frequency over intense hailstorms. Frequencies ≤37 GHz show the strongest distinction between hail and other hydrometeor types. Frequencies ≥37 GHz appear to have a much greater sensitivity to scattering by ice in general than to scattering by liquid or emission/absorption from liquid water, but without much distinction between particular types of large ice.

A greater volume and/or bigger hailstones may increase the chance of hail reaching the surface and, therefore, may be expected to be associated with a stronger scattering signature than hail that only occurs in the upper portions of clouds (i.e., hail that melts before reaching the surface). Results, indeed, indicate that low-level hail (below 2.5-km height) becomes probable at a colder BT than that of upper-level hail (above 2.5-km height) for frequencies ≥ 19 GHz. Low-level hail becomes probable for a BT below 240 K at 19 GHz, 170 K at 37 GHz, 90 K at 85 GHz, 80 K at 89 GHz, 100 K at 165 GHz, and 100 K at 183 GHz.

Profiles with graupel are generally associated with colder BTs than those without any hydrometeors at frequencies ≥ 37 GHz, but there is relatively little difference between the passive microwave signature of graupel and hail or between low- and high-density graupel at the higher frequencies examined. To identify graupel from passive microwave information, multiple channels may need to be used. For example, a large reduction in measured BT at 165 GHz relative to the background land coupled with little scattering signature at 10 and 19 GHz may indicate the presence of graupel as opposed to hail.

Not surprisingly, passive microwave BTs at the frequencies examined generally provide little discrimination between profiles with small and/or low-density ice (i.e., snow, aggregates, and ice crystals) and those without any hydrometeors. However, snow and aggregates appear to be associated with some ice scattering at CoSMIR frequencies (≥89 GHz).

The hydrometeors in the big drops category form from melting hail and/or graupel (e.g., Ryzhkov and Zrnic 1995b; Ryzhkov and Zrnic 1996; Schuur et al. 2001; Gatlin et al. 2015). Therefore, a strong scattering signature is observed for the big drop category at all frequencies examined. In contrast, the rain category may only be associated with smaller and/or less dense ice, which provides little if any scattering signature at the frequencies examined. Therefore, liquid precipitation can be best distinguished from no-rain profiles over land when associated above with hail and/or graupel (hydrometeor species associated with a strong ice scattering signature), as expected.

Probabilities of nonzero masses of any hydrometeor species at low levels as a function of BT were also examined. Results indicate that the probability of surface precipitation becomes very likely (~100%) for a BT below 230 K at 10 GHz, 250 K at 19 GHz, 240 K at 37 GHz, 230 K at 85 GHz, 220 K at 89 GHz, 140 K at 165 GHz, and 140 K at 183 GHz. Note that BTs measured over precipitation depend on various factors, including characteristics of the background surface and masses of various hydrometeor species, which vary across different times and geographic regions. Given the limited sample size used here, the thresholds provided likely cannot be generalized to other times/places; the thresholds are given primarily as a reference to compare with past and future studies. To derive thresholds that may be more widely applicable, future work may seek to derive and examine thresholds from BT depressions (i.e., differences in BTs between raining pixels and nearby nonraining pixels).

It should be noted that these results are valid over intense convection for a couple of days over Oklahoma in spring 2011 (the sample size is small). However, this work provides a first look at which hydrometeor species may be able to be detected and identified using data collected by the GMI, in particular over deep convection. Future work could involve further comparisons, where available, between airborne and/or spaceborne passive microwave observations and an HID applied to ground-based radar for different times and places. Additional MC3E cases of lighter rain could be included to obtain a more robust sample for liquid hydrometeors. Other future work could involve examining joint PDFs between multiple passive microwave channels [similar to the work of Liu et al. (2011), who examined the probability of lightning relative to 37- and 85-GHz BTs] in order to better understand the relationship between different channels and the unique and/or redundant information each channel provides.

Acknowledgments

Funding for this research was generously provided through the NASA Precipitation Measurement Missions Science Team. The authors thank Dr. Brenda Dolan for providing the code used for the hydrometeor identification and her help in running the code. The authors also thank Dr. Timothy Lang for his helpful suggestions for this work and three anonymous reviewers for providing thoughtful and helpful suggestions for improving the manuscript. In addition, the authors gratefully acknowledge the NASA EOSDIS Global Hydrology Resource Center DAAC for providing the AMPR, CoSMIR, and KVNX radar data.

REFERENCES

  • Adler, R. F., , H.-Y. M. Yeh, , N. Prasad, , W.-K. Tao, , and J. Simpson, 1991: Microwave simulations of a tropical rainfall system with a three-dimensional cloud model. J. Appl. Meteor., 30, 924953, doi:10.1175/1520-0450-30.7.924.

    • Search Google Scholar
    • Export Citation
  • Atlas, D., , S. Y. Matrosov, , A. J. Heymsfield, , M.-D. Chou, , and D. B. Wolff, 1995: Radar and radiation properties of ice clouds. J. Appl. Meteor., 34, 23292345, doi:10.1175/1520-0450(1995)034<2329:RARPOI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., 2009: Passive microwave brightness temperatures as proxies for hailstorms. J. Appl. Meteor. Climatol., 48, 12811286, doi:10.1175/2009JAMC2125.1.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., , and C. B. Blankenship, 2012: Toward a global climatology of severe hailstorms as estimated by satellite passive microwave imagers. J. Climate, 25, 687703, doi:10.1175/JCLI-D-11-00130.1.

    • Search Google Scholar
    • Export Citation
  • Dolan, B., , and S. A. Rutledge, 2009: A theory-based hydrometeor identification algorithm for X-band polarimetric radars. J. Atmos. Oceanic Technol., 26, 20712088, doi:10.1175/2009JTECHA1208.1.

    • Search Google Scholar
    • Export Citation
  • Ferraro, R. R., , and G. F. Marks, 1995: The development of SSM/I rain-rate retrieval algorithms using ground-based radar measurements. J. Atmos. Oceanic Technol., 12, 755770, doi:10.1175/1520-0426(1995)012<0755:TDOSRR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ferraro, R. R., , J. Beauchamp, , D. J. Cecil, , and G. M. Heymsfield, 2015: A prototype hail detection algorithm and hail climatology developed with the Advanced Microwave Sounding Unit (AMSU). Atmos. Res., doi:10.1016/j.atmosres.2014.08.010, in press.

    • Search Google Scholar
    • Export Citation
  • Gatlin, P. N., , M. Thurai, , V. N. Bringi, , W. Petersen, , D. Wolff, , A. Tokay, , L. Carey, , and M. Wingo, 2015: Searching for large raindrops: A global summary of two-dimensional video disdrometer observations. J. Appl. Meteor. Climatol., 54, 10691089, doi:10.1175/JAMC-D-14-0089.1.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, A. J., 1977: Precipitation development in stratiform ice clouds: A microphysical and dynamical study. J. Atmos. Sci., 34, 367381, doi:10.1175/1520-0469(1977)034<0367:PDISIC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Heymsfield, G. M., , L. Tian, , L. Li, , M. McLinden, , and J. I. Cervantes, 2013: Airborne radar observations of severe hailstorms: Implications for future spaceborne radar. J. Appl. Meteor. Climatol., 52, 18511867, doi:10.1175/JAMC-D-12-0144.1.

    • Search Google Scholar
    • Export Citation
  • Hou, A. Y., and Coauthors, 2014: The Global Precipitation Measurement (GPM) mission. Bull. Amer. Meteor. Soc., 95, 701722, doi:10.1175/BAMS-D-13-00164.1.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C. D., , W. Barnes, , T. Kozu, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 809817, doi:10.1175/1520-0426(1998)015<0809:TTRMMT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liu, C., , D. Cecil, , and E. J. Zipser, 2011: Relationships between lightning flash rates and passive microwave brightness temperatures at 85 and 37 GHz over the tropics and subtropics. J. Geophys. Res., 116, D23108, doi:10.1029/2011JD016463.

    • Search Google Scholar
    • Export Citation
  • Liu, H., , and V. Chandrasekar, 2000: Classification of hydrometeors based on polarimetric radar measurements: Development of fuzzy logic and neuro-fuzzy systems, and in situ verification. J. Atmos. Oceanic Technol., 17, 140164, doi:10.1175/1520-0426(2000)017<0140:COHBOP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • McGaughey, G., , E. J. Zipser, , R. W. Spencer, , and R. E. Hood, 1996: High-resolution passive microwave observations of convective systems over the tropical Pacific Ocean. J. Appl. Meteor., 35, 19211947, doi:10.1175/1520-0450(1996)035<1921:HRPMOO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mohr, K. I., , and E. J. Zipser, 1996a: Defining mesoscale convective systems by their 85-GHz ice-scattering signatures. Bull. Amer. Meteor. Soc., 77, 11791189, doi:10.1175/1520-0477(1996)077<1179:DMCSBT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mohr, K. I., , and E. J. Zipser, 1996b: Mesoscale convective systems defined by their 85-GHz ice scattering signature: Size and intensity comparison over tropical oceans and continents. Mon. Wea. Rev., 124, 24172437, doi:10.1175/1520-0493(1996)124<2417:MCSDBT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Mugnai, A., , E. A. Smith, , and G. J. Tripoli, 1993: Foundations for statistical–physical precipitation retrieval from passive microwave satellite measurements. Part II: Emission-source and generalized weighting-function properties of a time-dependent cloud-radiation model. J. Appl. Meteor., 32, 1739, doi:10.1175/1520-0450(1993)032<0017:FFSPRF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Oye, R., , C. Mueller, , and S. Smith, 1995: Software for radar translation, visualization, editing, and interpolation. Preprints, 27th Conf. on Radar Meteorology, Vail, CO, Amer. Meteor. Soc., 359–361.

  • Ryzhkov, A. V., , and D. S. Zrnic, 1995a: Comparison of dual-polarization radar estimators of rain. J. Atmos. Oceanic Technol., 12, 249256, doi:10.1175/1520-0426(1995)012<0249:CODPRE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., , and D. S. Zrnic, 1995b: Precipitation and attenuation measurements at a 10-cm wavelength. J. Appl. Meteor., 34, 21212134, doi:10.1175/1520-0450(1995)034<2120:PAAMAA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., , and D. S. Zrnic, 1996: Rain in shallow and deep convection measured with a polarimetric radar. J. Atmos. Sci., 53, 29892995, doi:10.1175/1520-0469(1996)053<2989:RISADC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ryzhkov, A. V., , D. S. Zrnic, , and B. A. Gordon, 1998: Polarimetric method for ice water content determination. J. Appl. Meteor., 37, 125134, doi:10.1175/1520-0450(1998)037<0125:PMFIWC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schuur, T. J., , A. V. Ryzhkov, , D. S. Zrnic, , and M. Schonhuber, 2001: Drop size distributions measured by a 2D video disdrometer: Comparison with dual-polarization radar data. J. Appl. Meteor., 40, 10191034, doi:10.1175/1520-0450(2001)040<1019:DSDMBA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Smith, E. A., , H. J. Cooper, , X. Xiang, , A. Mugnai, , and G. J. Tripoli, 1992: Foundations for statistical–physical precipitation retrieval from passive microwave satellite measurements. Part I: Brightness temperature properties of a time-dependent cloud-radiation model. J. Appl. Meteor., 31, 506531, doi:10.1175/1520-0450(1992)031<0506:FFSPPR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , and D. A. Santek, 1985: Measuring the global distribution of intense convection over land with passive microwave radiometry. J. Climate Appl. Meteor., 24, 860864, doi:10.1175/1520-0450(1985)024<0860:MTGDOI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , W. S. Olson, , W. Rongzhang, , D. W. Martin, , J. A. Weinman, , and D. A. Santek, 1983: Heavy thunderstorms observed over land by the Nimbus 7 scanning multichannel microwave radiometer. J. Climate Appl. Meteor., 22, 10411046, doi:10.1175/1520-0450(1983)022<1041:HTOOLB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , M. R. Howland, , and D. A. Santek, 1987: Severe storm identification with satellite microwave radiometry: An initial investigation with Nimbus-7 SMMR data. J. Climate Appl. Meteor., 26, 749754, doi:10.1175/1520-0450(1987)026<0749:SSIWSM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , H. M. Goodman, , and R. E. Hood, 1989: Precipitation retrieval over land and ocean with the SSM/I: Identification and characteristics of the scattering signal. J. Atmos. Oceanic Technol., 6, 254273, doi:10.1175/1520-0426(1989)006<0254:PROLAO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, R. W., , R. E. Hood, , F. J. LaFontaine, , E. A. Smith, , R. Platt, , J. Galliano, , V. L. Griffin, , and E. Lobl, 1994: High-resolution imaging of rain systems with the Advanced Microwave Precipitation Radiometer. J. Atmos. Oceanic Technol., 11, 849857, doi:10.1175/1520-0426(1994)011<0849:HRIORS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Straka, J. M., , D. S. Zrnic, , and A. V. Ryzhkov, 2000: Bulk hydrometeor classification and quantification using polarimetric radar data: Synthesis of relations. J. Appl. Meteor., 39, 13411372, doi:10.1175/1520-0450(2000)039<1341:BHCAQU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Thomason, J. W. G., , A. J. Illingworth, , and V. Marecal, 1995: Density and size distribution of aggregating snow particles inferred from coincident aircraft and radar observations. Preprints, 27th Conf. on Radar Meteorology, Vail, CO, Amer. Meteor. Soc., 127–129.

  • Toracinta, E. R., , D. J. Cecil, , E. J. Zipser, , and S. W. Nesbitt, 2002: Radar, passive microwave, and lightning characteristics of precipitating systems in the tropics. Mon. Wea. Rev., 130, 802824, doi:10.1175/1520-0493(2002)130<0802:RPMALC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wang, J. R., , P. E. Racette, , and J. R. Piepmeier, 2008: A comparison of near-concurrent measurements from the SSMIS and CoSMIR for some selected channels over the frequency range of 50–183 GHz. IEEE Trans. Geosci. Remote Sens., 46, 923933, doi:10.1109/TGRS.2007.904038.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , M. S. V. Rao, , E. B. Rodgers, , and J. S. Theon, 1977: A satellite technique for quantitatively mapping rainfall rates over the oceans. J. Appl. Meteor., 16, 551560, doi:10.1175/1520-0450(1977)016<0551:ASTFQM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , and L. S. Chiu, 1991: Retrieval of monthly rainfall indices from microwave radiometric measurements using probability distribution functions. J. Atmos. Oceanic Technol., 8, 118136, doi:10.1175/1520-0426(1991)008<0118:ROMRIF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
Save