• Brown Jr., P. S., 1988: The effects of filament, sheet, and disk breakup upon the drop spectrum. J. Atmos. Sci., 45 , 712718.

  • de Beauville, C. A., , R. H. Petit, , G. Marion, , and J. P. Lacaux, 1988: Evolution of peaks in the spectral distribution of raindrops from warm isolated maritime clouds. J. Atmos. Sci., 45 , 33203332.

    • Search Google Scholar
    • Export Citation
  • Feingold, G., , and Z. Levin, 1986: The lognormal fit to raindrop spectra from frontal convective clouds in Israel. J. Climate Appl. Meteor., 25 , 13461363.

    • Search Google Scholar
    • Export Citation
  • Feingold, G., , S. Tzivion, , and Z. Levin, 1988: Evolution of raindrop spectra. Part I: Solution to the stochastic collection/breakup equation using the method of moments. J. Atmos. Sci., 45 , 33873399.

    • Search Google Scholar
    • Export Citation
  • Garcia-Garcia, F., , and J. E. Gonzalez, 2000: Raindrop spectra observations from convective showers in the valley of Mexico. Proc. 13th Int. Conf. on Clouds and Precipitation, Reno, NV, International Commission on Cloud Physics and International Association for Meteorology and Precipitation, 398–401.

    • Search Google Scholar
    • Export Citation
  • Gossard, E. E., , R. G. Strauch, , and R. R. Rogers, 1990: Evolution of dropsize distribution in liquid precipitation observed by ground-based Doppler radar. J. Atmos. Oceanic Technol., 7 , 815828.

    • Search Google Scholar
    • Export Citation
  • Hallett, J., , and S. C. Mossop, 1974: Production of secondary ice particles during the riming process. Nature, 249 , 2628.

  • Hu, Z., , and R. C. Srivastava, 1995: Evolution of raindrop size distribution by coalescence, breakup, and evaporation: Theory and observations. J. Atmos. Sci., 52 , 17611783.

    • Search Google Scholar
    • Export Citation
  • Jameson, A. R., , and A. B. Kostinski, 1998: Fluctuation properties of precipitation. Part II: Reconsideration of the meaning and measurement of raindrop size distributions. J. Atmos. Sci., 55 , 283294.

    • Search Google Scholar
    • Export Citation
  • Joss, J., , and A. Waldvogel, 1967: Ein Spektrograph für Niederschlagstropfen mit automatischer Auswertung. Pure Appl. Geophys., 68 , 240246.

    • Search Google Scholar
    • Export Citation
  • Joss, J., , and A. Waldvogel, 1969: Raindrop size distribution and sampling size errors. J. Atmos. Sci., 26 , 566569.

  • Kozu, T., , K. K. Reddy, , S. Mori, , M. Thurai, , J. T. Ong, , and D. N. Rao, 2006: Seasonal and diurnal variations of raindrop size distribution in Asian monsoon region. J. Meteor. Soc. Japan, 84A , 195209.

    • Search Google Scholar
    • Export Citation
  • List, R., , and M. McFarquhar, 1990: The role of breakup and coalescence in the three-peak equilibrium distribution of raindrops. J. Atmos. Sci., 47 , 22742292.

    • Search Google Scholar
    • Export Citation
  • List, R., , N. R. Donaldson, , and R. E. Stewart, 1987: Temporal evolution of drop spectra to collisional equilibrium in steady and pulsating rain. J. Atmos. Sci., 44 , 362372.

    • Search Google Scholar
    • Export Citation
  • Low, T. B., , and R. List, 1982: Collision, coalescence and breakup of raindrops. Part I: Experimentally established coalescence efficiencies and fragment size distributions in breakup. J. Atmos. Sci., 39 , 15911606.

    • Search Google Scholar
    • Export Citation
  • Lucas, C., , A. D. MacKinnon, , R. A. Vincent, , and P. T. May, 2004: Raindrop size distribution retrievals from a VHF boundary layer profiler. J. Atmos. Oceanic Technol., 21 , 4560.

    • Search Google Scholar
    • Export Citation
  • Marshall, J. S., , and W. M. Palmer, 1948: The distribution of raindrops with size. J. Meteor., 5 , 165166.

  • McFarquhar, G. M., 2004: A new representation of collision-induced breakup of raindrops and its implications for the shapes of raindrop size distributions. J. Atmos. Sci., 61 , 777794.

    • Search Google Scholar
    • Export Citation
  • McFarquhar, G. M., , and R. List, 1993: The effect of curve fits for the disdrometer calibration on raindrop spectra, rainfall rate, and radar reflectivity. J. Appl. Meteor., 32 , 774782.

    • Search Google Scholar
    • Export Citation
  • Mossop, S. C., 1976: Production of secondary ice particles during the growth of graupel by riming. Quart. J. Roy. Meteor. Soc., 102 , 4557.

    • Search Google Scholar
    • Export Citation
  • Nissen, R., , R. List, , D. Hudak, , G. M. McFarquhar, , R. P. Lawson, , N. P. Tung, , S. K. Soo, , and T. S. Kang, 2005: Constant raindrop fall speed profiles derived from Doppler radar data analyses for steady nonconvective precipitation. J. Atmos. Sci., 62 , 220230.

    • Search Google Scholar
    • Export Citation
  • Prat, O. P., , and A. P. Barros, 2007: A robust numerical solution of the stochastic collection–breakup equation for warm rain. J. Appl. Meteor. Climatol., 46 , 14801497.

    • Search Google Scholar
    • Export Citation
  • Radhakrishna, B., , and T. N. Rao, 2009: Multipeak raindrop size distribution observed by UHF/VHF wind profilers during the passage of a mesoscale convective system. Mon. Wea. Rev., 137 , 978992.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , D. N. Rao, , K. Mohan, , and S. Raghavan, 2001: Classification of tropical precipitating systems and associated ZR relationships. J. Geophys. Res., 106 , 1769917711.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , N. V. P. Kirankumar, , B. Radhakrishna, , and D. N. Rao, 2006: On the variability of the shape-slope parameter relations of the gamma raindrop size distribution model. Geophys. Res. Lett., 33 , L22809. doi:10.1029/2006GL028440.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , N. V. P. Kirankumar, , B. Radhakrishna, , D. N. Rao, , and K. Nakamura, 2008: Classification of tropical precipitating systems using wind profiler spectral moments. Part I: Algorithm description and validation. J. Atmos. Oceanic Technol., 25 , 884897.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , B. Radhakrishna, , K. Nakamura, , and N. P. Rao, 2009: Differences in raindrop size distribution from southwest monsoon to northeast monsoon at Gadanki. Quart. J. Roy. Meteor. Soc., 135 , 16301637.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., , and C. W. Ulbrich, 2003: Cloud microphysical properties, processes, and rainfall estimation opportunities. Radar and Atmospheric Science: A Collection of Essays in Honor of David Atlas, Meteor. Monogr., No. 52, Amer. Meteor. Soc., 237–258.

    • Search Google Scholar
    • Export Citation
  • Sauvageot, H., , and M. Koffi, 2000: Multimodal raindrop size distributions. J. Atmos. Sci., 57 , 24802492.

  • Sheppard, B. E., 1990: Effect of irregularities in the diameter classification of raindrops by the Joss–Waldvogel disdrometer. J. Atmos. Oceanic Technol., 7 , 180183.

    • Search Google Scholar
    • Export Citation
  • Steiner, M., , and A. Waldvogel, 1987: Peaks in raindrop size distributions. J. Atmos. Sci., 44 , 31273133.

  • Testik, F. Y., , and A. P. Barros, 2007: Toward elucidating the microstructure of rainfall: A survey. Rev. Geophys., 45 , RG2003. doi:10.1029/2005RG000182.

    • Search Google Scholar
    • Export Citation
  • Tokay, A., , and D. A. Short, 1996: Evidence from tropical raindrop spectra of the origin of rain from stratiform versus convective clouds. J. Appl. Meteor., 35 , 355371.

    • Search Google Scholar
    • Export Citation
  • Tokay, A., , A. Kruger, , and W. F. Krajewski, 2001: Comparison of drop size distribution measurements by impact and optical disdrometers. J. Appl. Meteor., 40 , 20832097.

    • Search Google Scholar
    • Export Citation
  • Ulbrich, C. W., 1983: Natural variations in the analytical form of the raindrop size distribution. J. Climate Appl. Meteor., 22 , 17641775.

    • Search Google Scholar
    • Export Citation
  • Uma, K. N., , and T. N. Rao, 2009: Characteristics of vertical velocity cores in different convective systems observed over Gadanki, India. Mon. Wea. Rev., 137 , 953977.

    • Search Google Scholar
    • Export Citation
  • Valdez, M. P., , and K. C. Young, 1985: Number fluxes in equilibrium raindrop populations: A Markov chain analysis. J. Atmos. Sci., 42 , 10241036.

    • Search Google Scholar
    • Export Citation
  • Williams, C. R., , A. Kruger, , K. S. Gage, , A. Tokay, , R. Cifelli, , W. F. Krajewski, , and C. Kummerow, 2000: Comparison of simultaneous raindrop size distributions estimated from two surface disdrometers and a UHF profiler. Geophys. Res. Lett., 27 , 17631766.

    • Search Google Scholar
    • Export Citation
  • Willis, P. T., 1984: Functional fits to some observed drop size distributions and parameterization of rain. J. Atmos. Sci., 41 , 16481661.

    • Search Google Scholar
    • Export Citation
  • Zawadzki, I., , and M. De Agostinho Antonio, 1988: Equilibrium raindrop size distributions in tropical rain. J. Atmos. Sci., 45 , 34523459.

    • Search Google Scholar
    • Export Citation
  • Zawadzki, I., , F. Fabry, , and W. Szyrmer, 2001: Observations of supercooled water and secondary ice generation by a vertically pointing X-band Doppler radar. Atmos. Res., 59–60 , 343359.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    An illustration of the peak identification scheme from the Doppler power spectrum of LAWP at (a) 2.4 and (b) 5.7 km (in convection), and (c) the disdrometer-derived surface DSD. The arrows indicate the peaks identified by the algorithms.

  • View in gallery

    Two typical examples demonstrating the schemes employed here to classify precipitating systems. Time–height contours of reflectivity and Doppler velocity on (a),(b) 26 Aug 1999 and (d),(e) 2 Jun 2000, retrieved from LAWP. The open symbols represent cold rain, while the solid symbols represent warm rain. Different symbols represent different types of rain (square, convection; circle, stratiform; hexagon, transition inclusive; diamond, transition exclusive). (c),(f) The temporal variation of the rain integral parameters (Z, R and D0) and D0/R retrieved from collocated JWD measurements. The definition of symbols the is the same as for (a),(b).

  • View in gallery

    (a),(b) As in Figs. 2a and 2b, but for 5 Aug 2000 showing the passage of a convective storm. Sequences of Doppler power spectra showing MPs during (c) 1629–1639 and (d) 1705–1719 LT. The shading indicates the backscattered power at each spectra bin. The line spectrum at each altitude is normalized by the maximum value in that particular altitude.

  • View in gallery

    (a),(b) As in Figs. 2a and 2b, but for 10 Aug 2000 during the passage of a stratiform rain system. (c) Sequence of Doppler power spectra showing MPs during 0008–0053 LT.

  • View in gallery

    As in Fig. 3, but for 6 Aug 2000, showing the occurrence of MPs during the passage of an MCS.

  • View in gallery

    As in Fig. 3, but for 14 Jul 1999, showing the occurrence of MPs during the passage of a warm rain system.

  • View in gallery

    (left) Altitude distribution of the percentages of occurrence of single (dash–dot line), two (solid line), and multiple (more than two) (dotted line) peaks. The scale for single (two and more than two) peak(s) occurrence is provided below (above). (right) Altitude variations of the numbers of rain data points.

  • View in gallery

    As in Fig. 7 but only for two and more than two peaks in (a) cold convection, (b) transition-inclusive, (c) transition-exclusive, and (d) cold stratiform rain.

  • View in gallery

    As in Fig. 8 but only for warm rain (a) convection and (d) stratus rain.

  • View in gallery

    Percentage occurrence of peaks (solid line with circle) as a function of drop diameter in (top) stratiform, (middle) transition, and (bottom) convective precipitation. The numbers of data points (right axis) in each channel number used to calculate the percentage of occurrence are also included.

  • View in gallery

    As in Fig. 7, but for the SWM and NEM.

  • View in gallery

    As in Fig. 10, but for the SWM and NEM.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 29 29 4
PDF Downloads 17 17 2

Statistical Characteristics of Multipeak Raindrop Size Distributions at the Surface and Aloft in Different Rain Regimes

View More View Less
  • 1 National Atmospheric Research Laboratory, Gadanki, India
© Get Permissions
Full access

Abstract

Two years (∼672 h) of lower-atmospheric wind profiler (LAWP) and 4 yr (∼733 h) of Joss–Waldvogel disdrometer measurements are utilized to study the multipeak (MP) occurrence statistics at the surface and aloft. For the first time, an attempt has been made to address several key questions regarding MPs: their occurrence statistics and their dependency on height, season, and type of precipitation. MPs are not exceptional; rather, they are observed at all altitudes, albeit with different occurrence percentages. The occurrence of MPs seems to be height dependent, and this dependency varies with the type of rain system. The occurrence percentage of bimodal echo (two peaks) is high above (below) the melting level (ML) in convection (in other types of rain). The percentage occurrence of bimodal echo in warm rain is similar to that in cold rain, but only below the ML. The spectrum with more than two peaks appears to be predominantly in convection, particularly above 4 km. The MP statistics on the surface DSD derived from disdrometer data also support the profiler statistics qualitatively (occurrence is more likely in convection); however, the magnitudes of the percentage of occurrence are different at the surface and aloft. The peaks in the raindrop size distribution (DSD) spectra exist predominantly in drop diameter ranges of 0.45–0.65 and 0.9–1.3 mm in all types of rain, consistent with earlier numerical and observational studies. The MP occurrence does not have seasonal dependence aloft, but shows some variation at the surface with a larger percentage of the occurrences in the southwest monsoon. However, peaks in the surface DSD exist at same diameters in both monsoon seasons.

Corresponding author address: Dr. T. Narayana Rao, National Atmospheric Research Laboratory, P.O. Box 123, SVU Campus, Prakasham Nagar, Tirupati, 517 502 AP, India. Email: tnrao@narl.gov.in

Abstract

Two years (∼672 h) of lower-atmospheric wind profiler (LAWP) and 4 yr (∼733 h) of Joss–Waldvogel disdrometer measurements are utilized to study the multipeak (MP) occurrence statistics at the surface and aloft. For the first time, an attempt has been made to address several key questions regarding MPs: their occurrence statistics and their dependency on height, season, and type of precipitation. MPs are not exceptional; rather, they are observed at all altitudes, albeit with different occurrence percentages. The occurrence of MPs seems to be height dependent, and this dependency varies with the type of rain system. The occurrence percentage of bimodal echo (two peaks) is high above (below) the melting level (ML) in convection (in other types of rain). The percentage occurrence of bimodal echo in warm rain is similar to that in cold rain, but only below the ML. The spectrum with more than two peaks appears to be predominantly in convection, particularly above 4 km. The MP statistics on the surface DSD derived from disdrometer data also support the profiler statistics qualitatively (occurrence is more likely in convection); however, the magnitudes of the percentage of occurrence are different at the surface and aloft. The peaks in the raindrop size distribution (DSD) spectra exist predominantly in drop diameter ranges of 0.45–0.65 and 0.9–1.3 mm in all types of rain, consistent with earlier numerical and observational studies. The MP occurrence does not have seasonal dependence aloft, but shows some variation at the surface with a larger percentage of the occurrences in the southwest monsoon. However, peaks in the surface DSD exist at same diameters in both monsoon seasons.

Corresponding author address: Dr. T. Narayana Rao, National Atmospheric Research Laboratory, P.O. Box 123, SVU Campus, Prakasham Nagar, Tirupati, 517 502 AP, India. Email: tnrao@narl.gov.in

1. Introduction

Knowledge of raindrop size distribution (DSD) as a function of time and height not only advances our understanding of microphysical processes occurring within the precipitating system, but is also required for several application-oriented studies, such as soil erosion, microwave attenuation, atmospheric pollution, and satellite retrieval of precipitation (Testik and Barros 2007 and references therein). The DSD evolves with height as a result of several complex microphysical and dynamical processes, like collision–coalescence, breakup, evaporation, drop sorting due to updrafts, and horizontal wind [see a review of these processes by Rosenfeld and Ulbrich (2003)]. Which of these processes dominate depends on the rain type and ambient environment. Nevertheless, the dominant process dictates the shape of the DSD. For many practical purposes, the DSD is assumed to follow some known functional form, either exponential (Marshall and Palmer 1948), or lognormal (Feingold and Levin 1986), or gamma (Ulbrich 1983). A few observational and modeling studies have shown that the DSD deviates considerably from the above functional forms and on occasion exhibits multiple peaks (MPs) in the distribution (multipeak raindrop size distribution, MRDSD) (List et al. 1987; Steiner and Waldvogel 1987; Sauvageot and Koffi 2000; McFarquhar 2004; Prat and Barros 2007; Radhakrishna and Rao 2009, hereinafter RR09, and references therein).

Modeling efforts beginning with Low and List (1982), continuing to McFarquhar (2004), and to Prat and Barros (2007) have shown that the equilibrium DSD contains MPs (two or three depending on the parameterization scheme employed in their study). They have shown that the MPs are a result of the fragmentation of filaments, sheets, and disks. The observational evidence for MPs came from early measurements of DSD with Joss–Waldvogel disdrometers (JWDs) (Steiner and Waldvogel 1987; Zawadzki and De Agostinho Antonio 1988; de Beauville et al. 1988; Sauvageot and Koffi 2000). These studies observed peaks at drop diameters predicted by numerical models (Valdez and Young 1985; List et al. 1987; Brown 1988; Feingold et al. 1988; List and McFarquhar 1990; Hu and Srivastava 1995). McFarquhar and List (1993), however, argued that the peaks in DSD are the result of instrumental artifacts, imperfect calibration, and misplacement of bin diameter limits (Sheppard 1990). Subsequent studies with corrected DSDs and also by optical probes confirm the existence of MPs in the DSD at diameters predicted by numerical models (Willis 1984; Garcia-Garcia and Gonzalez 2000; Sauvageot and Koffi 2000). However, some of these studies attributed the occurrence of MPs to the overlapping of rain shafts (Sauvageot and Koffi 2000) and inhomogeneous rain patches (McFarquhar 2004).

As seen above and also noted by several scientists, the limitations in the JWD restrict its usage to study MRDSD comprehensively. For instance, Sauvageot and Koffi (2000) considered MPs in the DSD only for drop diameters >2 mm. Further, the sampling volume and time (1 min.) may not be adequate to get a stable and smooth DSD. It may artificially generate peaks in the distribution. Addition of consecutive spectra will reduce the variance in DSD, but the addition of data over longer periods, particularly inhomogeneous rain, will generate artificial peaks (Jameson and Kostinski 1998).

On the other hand, vertically pointing radars sample a large volume (depending on the beamwidth and pulse length) and, therefore, the Doppler spectra and the retrieved DSD are robust. If a vertically pointing Doppler radar observes two groups of precipitation particles with distinct fall velocities, then the observed spectra show a bimodal distribution. The source for bimodality in the precipitation spectra has been studied with wind profilers (Gossard et al. 1990; RR09) and X-band radars (Zawadzki et al. 2001). Gossard et al. (1990) noticed MPs in the UHF radar spectra during precipitation and attributed them to melting–breakup processes at the melting level (ML) and the coalescent growth of cloud droplets. Zawadzki et al. (2001) ascribed the presence of MPs to microphysical processes (secondary ice generation and supercooled drizzle) occurring near the ML. Recently, RR09 utilized VHF and UHF wind profiler observations during the passage of two mesoscale convective systems to understand the preferential phase and heights at which MRDSD occurs. They have shown that both dynamical [overlapping of rain shafts; Sauvageot and Koffi (2000); McFarquhar (2004)] and microphysical (riming) processes can cause the MRDSD.

The above studies, modeling as well as observational, are based on either equilibrium DSDs or some case studies. To the authors’ knowledge, the vertical distribution of MPs in the DSD has not been reported until today. Also, it is not clear whether the MRDSD is a common phenomena or exceptional. Regarding this issue, several questions are yet to be answered. What is the percentage occurrence of MRDSD at the surface and aloft? Is there any preferential height region in which MRDSD occurs? What kind of precipitating system is favorable for the occurrence of MRDSD? An attempt has been made in this paper to address the above issues with the help of UHF profiler and disdrometer measurements made at Gadanki, a tropical station in the southern peninsular region of India. The present article is the first of its kind, in which the statistics on the altitude distribution of multipeaked spectra in different types of precipitating systems are provided. Gadanki receives significant rainfall during the southwest monsoon (SWM: June–September) and northeast monsoon (NEM: October–December) seasons. The nature of rainfall in the SWM is continental, while in the NEM it is maritime (Rao et al. 2009). Therefore, it is also possible to study the occurrence probability of MPs in continental and maritime precipitating systems.

The paper is organized as follows. The details on the data and method of analysis (identification of MPs in radar spectra and disdrometer-derived DSDs, and the classification of precipitating systems using profiler and disdrometer data, etc.) are described in section 2. The results are discussed in section 3, which also contains the statistics on the seasonal variation of MRDSD, if any. The results are summarized in section 4.

2. Data and methodology

The present study utilizes 2 yr (1999–2000) of lower-atmospheric wind profiler (LAWP) and 4 yr (1999, 2000, 2006 and 2007) of collocated JWD (RD-69) measurements made at Gadanki (13.5°N, 79.2°E). The system description of the LAWP is given in Rao et al. (2001). Important specifications of the LAWP can be found in Table 1. The LAWP operates continuously, switching between low and high modes with three beams oriented along the east (15°), north (15°), and zenith (Zx) directions. The range resolution for both modes is the same (150 m), but the time taken for the completion of one scan (three beams) is different for low and high modes (7 and 4 min, respectively). A total of 7334 vertical Doppler spectra collected in ∼672 h are used for this study. The JWD disdrometer samples the raindrops in 20 channels with different diameter intervals ranging from 0.3 to 5.3 mm at every 1-min interval. The impact of the raindrop striking the Styrofoam body with its surface area of 50 cm2 is converted into an electric pulse, whose amplitude is proportional to the drop size. A detailed system description of JWD can be found in Joss and Waldvogel (1967; 1969). The limitations of the system, like underestimation of smaller drops in heavy rain due to the ringing of the receiving cone and the increase of acoustic noise, are enumerated in several studies (Tokay et al. 2001 and references therein) and therefore are not repeated here. The 1-min samples are corrected for dead time using the formulas supplied by the manufacturer. To minimize sampling errors due to insufficient raw drop counts, 1-min samples with fewer than 10 drops or R < 0.1 mm h−1 are removed (Tokay and Short, 1996). The disdrometer-recorded drop counts are converted into the number of drops per unit volume [N(D)] using the standard formula.

The peaks in the LAWP Doppler power spectra are identified using an automated algorithm, similar to the one outlined in Lucas et al. (2004). First, interference (generated by external sources or systems), if any, is removed from the observed Doppler spectra. The Doppler spectrum is smoothed twice with a three-point running mean (first from one end of the spectra and then from the other end to avoid spectral peak shifting) to minimize noise fluctuations. Figures 1a and 1b broadly depict the procedure adopted here. The thin solid line represents the observed spectrum at 2.4 km, while the thick solid line represents the spectrum after smoothing. Gradients of spectral power (P) with respect to velocity (υ)(ΔPυ) are calculated and are also smoothed twice (shown as a dashed line). The peaks in the spectrum are identified by looking at the zero crossings from the negative to the positive sides of the estimated gradients. Note that the echo from refractive index fluctuations (hereinafter, clear-air echo) is also shown in Fig. 1a. It should not be considered as a peak while counting MPs. The clear-air echo will appear in the spectrum only in light rain (Fig. 1a), when the ambient vertical air motion and reflectivity from hydrometeors are generally weak. Therefore, the echo may not satisfy the thresholds set for identifying rain echo (Z > 12 dBZe and the Doppler velocity should be beyond ±1 m s−1). However, in heavy rain (convection), the situation becomes much more complex. The clear-air backscattered power and vertical air motion can exceed the thresholds of the rain echo. Also, the Doppler velocity of the hydrometeors may fall within ±1 m s−1. The later situation generally occurs above the melting level, where the updrafts are strong and the Doppler velocities of the hydrometeors (result of vertical air motion and fall velocity of the hydrometeors) is small (Rao et al. 2001; Uma and Rao 2009; RR09). Nevertheless, in both cases it is difficult to identify MPs. Therefore, whenever Z exceeds 38 dBZe, the data are manually inspected for MPs. Manual inspection reveals that the precipitation echo masks the clear-air echo in such instances (Fig. 1b). Therefore, the clear-air echo will not appear in the spectrum. The peaks in the DSD (retrieved from the disdrometer) are identified following Steiner and Waldvogel (1987); that is, if the N(D) at any diameter is larger than its adjacent diameters, then a peak is considered to exist at that diameter. An example is shown in Fig. 1c, in which the identified peaks are denoted with arrows.

To obtain the statistics on MRDSD in different types of precipitating systems, the LAWP and disdrometer measurements are classified, following Rao et al. (2008) and Rao et al. (2001), respectively. The profiler classification scheme utilizes wind profiler spectral moments (Rao et al. 2008), while the disdrometer classification scheme uses rain integral parameters retrieved from the disdrometer (Rao et al. 2001). The profiler classification algorithm first identifies warm and cold rain based on whether the top of the rain area (defined by 12 dBZe) is below or above a threshold height, 5.5 km [=the climatological 0°C isotherm level +2σ level; Rao et al. (2008)]. Note that Rao et al. (2008) used a threshold reflectivity of 20 dBZe to delineate warm and cold rains. In the present study, this threshold is relaxed to allow weak rain and also backscatter from cloud drops. The warm rain is further divided into two types based on the reflectivity and Doppler velocity. The rain is considered to be warm convection if the reflectivity exceeds 38 dBZe or the Doppler velocity is ≤−10 m s−1 or ≥1 m s−1, below 3.5-km height, or else the rain is treated as warm stratus. The cold rain is classified into cold convection, cold stratiform, or transition-inclusive and transition-exclusive types, based on the presence or absence of a radar bright band, large reflectivities, and Doppler velocities. Cold convection is characterized by the presence of large reflectivity (>38 dBZe) or Doppler velocity (≤−10 m s−1 or ≥1 m s−1) and the absence of a radar bright band. The cold stratiform rain is identified by the presence of the radar bright band and the absence of intense rain (large reflectivity and Doppler velocity). The presence of both convection and a bright band indicates transition-inclusive precipitation, while the absence of both convection and a bright band indicates that the rain is of the transition-exclusive type.

To overcome sampling problems (i.e., MPs formed artificially due to 1-min sampling), the disdrometer data are integrated for 5 min. We arrived at a 5-min integration time after considering several factors. Earlier studies have shown that the DSD spectra are coherent for ∼5 min (Sauvageot and Koffi 2000; RR09). Further, the 5-min integration nearly matches the temporal resolution of the LAWP high mode (4 min). We constructed the MRDSD statistics with different integration intervals [1 (no integration), 3, and 5]. For intervals 3 and 5, the statistics are more or less similar. Integrating for longer periods (>5 min) may not be sensible due to the large temporal variability of the DSD (Jameson and Kostinski 1998).

The 5-min-averaged disdrometer data are classified into three principle rain regimes (convection, stratiform, and transition) based on the variations of the median volume diameter (D0) with rain rate (R) (Rao et al. 2001). If the value of D0/R < 0.1 and R > 5 mm h−1, the rain is classified as convection. If the D0/R value lies between 0.1 and 0.5, then the rain is considered to be transition, and the rain is categorized as stratiform if the D0/R value is >0.5. To illustrate how well these classification schemes are working, two case studies are considered (26 August 1999 and 2 June 2000). The time–height maps of reflectivity (dBZe) and the Doppler velocity of the hydrometeors from LAWP observations along with R, the reflectivity factor (Z), and D0 from disdrometer data for 26 August 1999 and 2 June 2000 are shown in Figs. 2a and 2b, respectively. The rain type, as identified by the above methods, is also indicated in Figs. 2a and 2b with different symbols for different types of rain. Note that the disdrometer cannot distinguish between the warm and cold rain and also the transition-inclusive and transition-exclusive types of rain. The disdrometer stratifies the data into the categories and is represented with open symbols (square, convection; hexagon, transition; and circle, stratiform).

It is evident from Fig. 2a that the rain top varies between 3 and 6.9 km, indicating the presence of both warm and cold precipitation. The profiler classification scheme (Rao et al. 2008) delineated the rain event into different types of rain regimes. Initially, the rain is of the warm convective type (1940–1950 LT), followed by transition-inclusive (1950–2015 LT), stratiform (2015–2130 LT) with clear brightband structure, and stratus again (2130–2330 LT). The disdrometer classification scheme is also largely consistent with the profiler scheme. According to the disdrometer classification, the rain is predominantly convection in the initial stages (1945–2005 LT), followed by transition and stratiform rain. Note that the profiler classification scheme employs spectral moments aloft, while the disdrometer measurements are taken at the surface. Some time shift (depending on the type of rain, fall velocity, and vertical air velocity), therefore, can be expected, as seen in Fig. 2a. Further, as discussed above, the disdrometer classification scheme mostly considers profiler-derived stratus, transition-exclusive, and stratiform rain as stratiform rain. Figure 2b also shows a cumulonimbus type of rain system, with intense convection in the initial stage, followed by transition and stratiform rain. Both radar and disdrometer results seem to distinguish these regions properly. Given the differences in the sampling volumes of these instruments, the evolutions of DSDs during their descents to the ground, and that they employ different types of classifications schemes, the agreement between the types of rain systems is quite encouraging. Utilizing these classification schemes, the occurrence percentage of MPs in different rain regimes is examined in the next section.

3. Results and discussion

a. Typical Doppler spectra showing MPs in different types of precipitation

Prior to studying the variations of the MP distributions as a function of altitude and precipitation type, typical Doppler spectra showing MPs at different altitudes and in different precipitating systems are illustrated with a series of examples. Figures 3 –6 depict the time–height maps of the reflectivity, Doppler velocity and temporal sequence of spectra during the passage of a convective storm, a stratiform rain event, a mesoscale convective system (MCS), and a warm rain event, respectively, over the profiler site. The reflectivity and Doppler velocity contours give an idea of the stage of the system at which the MPs were observed. The rain type (following Rao et al. 2008) is also indicated in Figs. 3 –6.

Figure 3 shows intense convection during 1610–1745 LT, with large reflectivities and Doppler velocity values at all altitudes. A series of spectra presented in Fig. 3c show MPs in the initial (1629–1639 LT) and decaying (to transition inclusive; 1705–1719 LT) stages of convection. The spectra show MPs at different altitudes at different times. On many occasions, the MPs are not confined to few range gates but are observed continuously over a greater depth. Looking at the slope of the traces of the MPs (lines joining MPs at different altitudes) in the spectra collected during 1629–1639 LT, the overlapping of the inhomogeneous rain patches (because the system is intensifying) seems to be responsible for the multimodality in the spectra. In the decaying stage, in particular during 1708–1715 LT, the spectra show three peaks in the vicinity of the ML. All these traces show enhancements in their Doppler velocity and reflectivity (to some extent), indicating that they are associated with solid hydrometeors. The source for this multimodality is unclear. The large Doppler velocities above the ML suggest that riming is occurring predominantly. Earlier studies have shown that the secondary ice generation (Hallett–Mossop mechanism) during riming becomes important and can produce MPs (Zawadzki et al. 2001).

The precipitation on 10 August 2000 is predominantly stratiform rain, with a clear brightband structure (enhancement in reflectivity and also a large gradient in Doppler velocity) near the ML. MPs are seen during 0008–0053 LT, in which the profiler moments have shown some variations, below the melting region. One trace of the MPs shows only slight variation with altitude near the ML, while the other shows enhancements in the reflectivity and gradient in Doppler velocity, indicating that the former is associated with liquid hydrometeors and the latter with melted solid hydrometeors (Zawadzki et al. 2001; RR09).

Figure 5 shows profiler reflectivity and Doppler velocity contours during the passage of a mesoscale convective system. Convection prevailed predominantly until 1930 LT, followed by transition-inclusive and stratiform rain. The multimodal spectra are observed during the transition period (Fig. 5c), in particular at 1943 and 1946 LT. These spectra show three peaks below 3.3 km, which could be related to collision–coalescence and breakup processes. The MPs are also observed during 2314–2321 LT, when the rain is intensifying into another storm. RR09 also observed similar types of spectra, with one trace of echoes showing large Doppler velocity gradients and reflectivities within the melting region and the other trace with small or no gradients in the Doppler velocities near the ML. They attributed these traces to solid hydrometeors (the former case) and supercooled droplets (the latter case). Figure 5 is also consistent with the result of RR09, confirming that convection and transition periods (between the convection and stratiform rain, and also transitioning from rain to no rain or vice versa) in an MCS are favorable stages for MPs to occur.

The MPs observed during a warm rain event are shown in Fig. 6. The rainfall is meager and it reached the surface only for a short duration. The rain is mostly of stratus type. MPs are observed for about 20 min during 1008–1028 LT. These peaks are noticed from the top of the cloud (∼5 km altitude). It is difficult to guess the source for this bimodality. Nevertheless, this example shows that MPs can occur even in light rain (reflectivity <20 dBZe).

b. Altitude distribution of MPs

The rain profiles are first separated from the total database to study the statistical characteristics of the MPs. The profile is considered to be a rain profile if it satisfies the reflectivity and Doppler velocity thresholds (12 dBZe, <−1 m s−1, >1 m s−1) (Rao et al. 2008) continuously for at least four range bins (600 m). A total of 7334 profiles qualified under the above conditions and are used for further analysis. The spectra of these profiles are checked for MPs by employing the MP detection algorithm discussed in section 2. The occurrence of two peaks and more than two peaks is counted and tabulated. The MPs are observed in 1078 spectra (14.7%).

The altitude distribution of the spectra with single peak, two peaks, and more than two peaks is shown in Fig. 7. It also contains the vertical profiles of a number of rain data points. Note that the number of rain data points varies with altitude depending on the height at which rain/cloud exists. The number of data points at any altitude is larger than 3000. It is clearly evident from Fig. 7 that the precipitation spectra mostly show a single peak (∼92% and ∼97% below and above the ML, respectively). Two peaks are also observed in a considerable number of spectra (∼7% and ∼2% below and above the ML). The spectra show more than two peaks in about 1% of the observations. Overall, the occurrence of MPs is higher below the ML than above.

To study the altitude distribution of MPs in different rain regimes, the rain profiles are categorized into various rain regimes, using the Rao et al. (2008) classification scheme illustrated in section 2. The occurrence percentages of different rain regimes are not compared with Rao et al. (2008) results as a different reflectivity threshold is used in the present study. The total number of profiles, number of spectra showing MPs at some altitude, and the percentages of occurrence of MPs in each rain regime are given in Table 2. The occurrence of MPs is high in cold convection (52.5%), followed by transition inclusive (42.1%), warm convection (33.1%), cold stratiform (19.5%), warm stratiform (9.2%), and transition exclusive (2.6%). The percentage occurrence is small in stratus and transition-exclusive types of rain, partly due to the large number of profiles in these regimes [profiles during weak rain and precipitating clouds (Virga type) fall in these rain regimes].

Figure 8 shows the altitude distribution of the occurrences of MPs (in terms of %) in cold rain. The vertical profiles of the numbers of valid rain data points in each rain regime are also included in Fig. 8. For estimating the percentage of occurrence in each rain category, only multipeaked spectra (i.e., spectra showing MPs at some altitude) are used. In the cold convection category (Fig. 8a), the percentage of occurrence of MPs is higher above the ML than below. About 25% (15%) of the spectra show MPs above (below) the ML in this category. The number of spectra showing more than two peaks is also larger (∼10% above the ML) in the cold convection category. It is important to understand that the microphysical and dynamical processes produce such a height distribution. The overlapping of rain shafts or inhomogeneous rain patches is certainly one of the processes that can generate MPs. However, overlapping of rain shafts can occur at any altitude and therefore the large percentage occurrence of MPs above the ML is, perhaps, related to other processes. Zawadzki et al. (2001), via a theoretical and observational study, illustrated that ice and supercooled water can coexist above the ML and can cause bimodality. The severe updrafts in convection can carry the supercooled water to greater heights in the convective cell. At Gadanki, updrafts of the order of 15 m s−1 are observed in the middle and upper troposphere in several deep convective cells (Uma and Rao 2009). Drop sorting of raindrops can also cause multimodality below the updraft core. RR09 observed MPs at almost all altitudes above the ML and have shown that both of these peaks are associated with solid hydrometeors. These solid hydrometeors grow primarily due to riming and show large Doppler velocities. Earlier studies have shown that the generation of secondary ice (by the Hallett–Mossop mechanism) during riming is a common phenomena and this process peaks at ∼−5°C (Hallett and Mossop 1974; Mossop 1976). This process can also generate MPs in the spectra above the ML. Further, the spectra in the convection are broad above the ML, partly due to intense turbulence and partly due to the presence of hydrometeors with different velocities (or different masses). The undulations within these broad spectra can appear to be multipeaks (see the spectra in Fig. 3c, particularly in and above the ML). The large percentage occurrence of MPs more than 2 are mostly a result of the undulations in the broad spectra above the ML.

In the transition rain regime, the altitude distribution of MPs is more or less similar in the inclusive (Fig. 8b) and exclusive (Fig. 8c) types of transition precipitation. Contrary to the altitude distribution in convection, the percentage occurrence in the transition rain category is higher below the ML than above. The percentages of occurrence of bimodal echoes above and below the ML in the transition-inclusive (exclusive) regime are mostly in the ranges of 10%–20% (6%–16%) and 20%–30% (14%–33%), respectively. The percentage occurrence of MPs more than 2 is mostly within 5% in both types of transition rain; however, large values are observed in the vicinity of the ML (within 2 km) in the transition-inclusive category and below the ML (2–4 km) in the transition-exclusive category. The rainfall changes considerably during the transition rain—be it inclusive or exclusive. Therefore, the interference of inhomogeneous rain patches is one of the most important processes that can cause multimodality in this category. Some of the other processes occurring in convection (stratiform) can also generate MPs in the spectra in the early (final) stages of transition-inclusive rain.

The height dependency of MPs is strong in the cold stratiform category (Fig. 8d). The percentage occurrence of bimodal echoes increases gradually with altitude and shows a peak at 3.3 km (∼33%), then decreases drastically to 4.8% at 4.5 km. Above 4.5 km, the percentage occurrence is found to be less than 8%. The altitude distribution of more than two MPs is also similar to the altitude variation of two peaks with relatively higher percentages of occurrence below 3.6 km than above. Earlier studies have observed two traces of echoes in the Doppler spectra below the ML (Zawadzki et al. 2001; RR09). One trace of echoes shows a large Doppler velocity gradient and enhancement in reflectivity within the melting layer (melting solid hydrometeors) and the other appears without much change in Doppler velocity and reflectivity (liquid hydrometeors). Gossard et al. (1990) also observed multiple peaks below the ML and attributed them to the melting of ice hydrometeors and the coalescence of cloud drops. The above types of spectra are primarily seen just below the ML (3–4 km) (see the Doppler spectra at 0031 and 0041 LT in Fig. 4) and exist predominantly in this category. This is the main reason for the high percentage of occurrence of MPs in the height region 3.3–3.6 km.

The altitude distribution of the occurrence of MPs in warm rain is similar to that in cold rain, but below the 0°C isotherm level (Fig. 9). In warm convection (Fig. 9a), two peaks are observed in ∼14%–25% of the observations below 4.2-km altitude, except for a narrow height region (2.4–2.55 km). The percentage of occurrence of MPs suddenly jumps to ∼40% in this height region. The source for this high-percentage occurrence is not immediately obvious. The percentage occurrence for MPs more than two is high just below the ML (7%–11% in the height region 3.45–4.2 km). The warm stratus (Fig. 9b) also shows a high percentage of occurrences of MPs below 4 km. The microphysical (coalescence, breakup, etc.) and dynamical (overlapping of rain shafts and inhomogeneous rain patches, drop sorting, sudden changes in the vertical wind, etc.) processes discussed with regard to cold rain below the ML can cause multimodality in warm rain.

c. MPs in the surface DSD in different rain regimes

To study the occurrence of MPs in DSD spectra at the surface, collocated disdrometer data are used. A total of 10 250 five-minute-averaged DSD spectra are utilized for this purpose. About 30% of the observations show MPs. The DSD data are grouped into convection, transition, and stratiform rain regimes, following the procedure outlined in section 2. The occurrence statistics for two peaks and more than two peaks in each rain regime are given in Table 3. The occurrence percentage of MPs is consistent with that of the profiler, with a high percentage of occurrence in convective rain (46%, among which two peaks occur for 38.9% and more than two peaks occur for 7.1%) compared to other rain regimes. The percentage occurrence of MPs of more than two is also higher in convection, consistent with profiler statistics. However, the magnitude of the percentage of occurrence in each rain regime is different at the surface and aloft. Given the differences in the outcomes of the classification algorithms [three (six) rain regimes with disdrometer (profiler) classification] and the differences in the microphysical and dynamical processes aloft and at the surface, the difference in magnitude is not really surprising.

Table 3 provides the statistics for the MPs in different rain regimes; however, it does not show at what diameters the peaks generally occur. As discussed in the introduction, the bin diameter limits set for the JWD can generate peaks artificially. As they are permanent, they are clearly observed in the averaged DSD (Sauvageot and Koffi 2000). Therefore, averaging several DSD spectra (even the stratified DSD based on either Z or R) to determine the peak occurrence is not pragmatic. Note that these artificial peaks generally have small amplitudes; therefore, they may be weak in comparison with real MPs produced by microphysical and dynamical processes. In the present study, to find the diameters at which the peaks occur, the percentage of occurrence of a peak is estimated from the 5-min-averaged individual spectra and is plotted as a function of drop diameter (similar to Figs. 2–4 in Steiner and Waldvogel 1987). The percentage occurrence of peaks at each diameter for different rain regimes is shown in Fig. 10. It is clear from Fig. 10 that the peaks can occur at all diameters in all types of rain regimes. However, the magnitudes of the percentages of occurrence differ for different types of rain systems. The distribution shows two modes in convection and transition and two clear modes and a few more weak modes (more like undulations) in stratiform rain regimes. Note that the relatively high percentage of occurrence of peaks at the large drop end in stratiform rain is primarily due to an insufficient number of samples. The first mode appears in the same drop diameter range (0.45–0.65 mm) in all of the rain regimes, albeit with varying magnitudes of percentage of occurrence. The mode at 0.55 mm is prominent in convection in comparison with other rain regimes. The second mode appears at a drop diameter of 0.91 mm in stratiform rain, and in the diameter range 0.91–1.1 mm in transition and 1.1 mm in convection rain regimes. The percentage occurrence of the mode at ∼1.2 mm is highest for convection regimes (∼43%). An underestimation of smaller drops (<1.5 mm) in heavy rain (convection) by the JWD has been noticed by several earlier investigators (Tokay et al. 2001; Williams et al. 2000). The underestimation of smaller drops means it automatically generates a peak at diameters less than 1.5 mm. Also, the surface and lower-tropospheric temperatures are high in SWM and, therefore, the efficacy of evaporation and vigorous convection is also significant in this season (Rao et al. 2006). As a result, the number of smaller drops is generally small in SWM. All these factors contribute for the high percentage of occurrence of a peak at 1.2 mm in convection. A third mode, albeit weak, is observed in stratiform rain at 1.67 mm, which is not apparent in other rain regimes. The above exercise (i.e., estimating the percentage occurrence) is performed again with DSD data that show multipeaks (not shown here). The distribution resembles Fig. 10, except that the magnitude is higher by 5%–10% at diameters corresponding to the first and second modes. The occurrences of peaks in the present study at diameters 0.55, 0.91–1.12, and 1.67 mm are consistent with earlier model studies (Valdez and Young 1985; Brown 1988). However, some of the models have shown peaks at 0.26 and 2.3 mm in the equilibrium DSD (McFarquhar 2004). The former is beyond the range of diameters that JWD can measure, while the latter is not seen prominently in any of the rain regimes.

d. Seasonal variation of occurrence statistics of MPs

Gadanki receives significant rainfall during the SWM (54% of the annual rainfall) and NEM (34%). Earlier studies have shown significant differences in rain DSDs, even after stratification of the data by R (Rao et al. 2001, 2009; Kozu et al. 2006). These studies have attributed the differences in DSDs to the type of systems affecting a particular region. For instance, the nature of rain during the SWM is continental, while during the NEM it is oceanic. Given such differences in the DSDs, it will be interesting to investigate the seasonal differences in the occurrences of MPs. Figures 11a and 11b illustrate vertical profiles of the occurrences of single peaks and MPs (two peaks and more than two peaks) during the SWM and NEM, respectively. The altitude variation of the MPs, in particular for two peaks, are not only similar during the SWM and NEM, but they are also similar to the variations of the MPs in Fig. 7. The magnitudes of the percentages of occurrence of MPs at different altitudes are also nearly equal during the SWM and NEM. The occurrence of more than two peaks is higher during the SWM than during the NEM at almost all of the heights. In the present study, the seasonal differences in the occurrence of MPs in various rain regimes are not studied, because of a lack of data. Although the number of rain profiles in some rain regimes (cold convection and cold stratiform) is sufficient for such a comparison, in other rain regimes the data are inadequate. Nevertheless, such a comparison is made below at the surface using disdrometer measurements.

Table 4 depicts the occurrence statistics of the MPs (two peaks and more than two peaks) at the surface in convection, transition, and stratiform rain regimes for the SWM and NEM. The occurrence statistics in different rain regimes are different during the SWM and NEM. For instance, the percentages of occurrence of two-peaked DSD spectra in the transition and stratiform rain regimes are higher during the SWM than during the NEM. On the other hand, nearly the same convection is seen. In addition, the spectra with more than two peaks are more prevalent during the SWM than during the NEM in all types of rain. This is consistent with profiler statistics (see Fig. 11). Overall, the percentage occurrence of MPs is higher during the SWM (35.2%) than during the NEM (20.8%). In contrast, the percentage of occurrence profiles retrieved from profiler observations has not shown large seasonal differences. As shown in the examples, some of the MPs are limited in height range; they do not propagate down to the surface. Some of the MPs, of course, can be traced down to the surface, particularly during light-to-moderate rain (Zawadzki et al. 2001; Nissen et al. 2005; RR09).

The percentages of occurrence of peaks as a function of drop diameter are plotted in different rain regimes during the SWM and NEM (Fig. 12) to help us understand the seasonal differences in peak occurrences. The peak distribution is strikingly similar during the SWM and NEM for all types of rain, except for a small shift in the peak at 1.12 mm in the transition rain regime. The magnitudes of the percentages of occurrence of peaks are nearly equal at all drop diameters in both seasons, indicating negligible seasonal variation in the peak occurrence.

4. Summary

The occurrence statistics of MPs at the surface and aloft are studied with the help of LAWP and disdrometer observations. While MPs on the surface have previously been studied (Steiner and Waldvogel 1987; Sauvageot and Koffi 2000; RR09), the altitude distributions of MPs have not reported upon in the literature. This is the first observational report dealing with the altitude distribution of MPs in different rain regimes. Typical examples of Doppler spectra showing MPs in different rain regimes are shown and possible causative mechanisms for MPs are discussed. The LAWP data are classified into six rain regimes, following Rao et al. (2008), while the disdrometer data are divided into three rain regimes, following Rao et al.’s (2001) classification scheme. In addition, the seasonal variations of MPs are also studied. The major findings of this study are briefly outlined below.

  1. The radar observations clearly demonstrate that the MPs are prevalent at all altitudes, albeit with different magnitudes of percentage of occurrence. In general, bimodal echoes appeared more commonly at lower altitudes (7% and 2% below and above the ML, respectively). The spectra with more than two echoes also appeared for 1%–2% of the total observations at all altitudes. Five-minute-integrated disdrometer measurements show that the occurrence percentage of MPs at the surface is ∼30%, much higher than that obtained with the profiler (14.7%).
  2. Contrasting vertical profiles of MPs in different rain regimes reveals interesting differences. The altitude distribution of the MPs in convection is distinctly different from that of in the other rain regimes. Bimodal echoes are found predominantly above the ML in convection (25%), while in the other rain regimes, they are more common below the ML. In the stratiform rain regime, the altitude distribution of MPs shows a peak below the melting region. The altitude distribution of the percentage of occurrence of MPs in warm rain is similar to that in cold rain, but occurs only below the ML. The spectra with more than two echoes are seen often in convection, where the percentage of occurrence of these echoes is a factor 2 larger than that in the other rain regimes, particularly above the ML. These echoes are mostly found in broad Doppler spectra and are often seen during convection in and above the ML. The broadening of the spectra is mainly due to the prevalence of intense turbulence and a spectrum of hydrometeors during convection.
  3. Convection seems to be the preferred rain regime for the generation of MPs aloft and on the ground (Tables 2 and 3). The percentage occurrence of MPs more than 2 in surface DSDs is also high in convection, consistent with the profiler observations.
  4. Peaks in the DSD spectra occurred at preferred drop diameters (or ranges of drop diameters) in all types of rain. The first mode appears in the range of 0.45–0.65 mm in all rain regimes, while the second mode appears at different diameters in different rain regimes (at 0.91 mm in stratiform rain, in the range of 1.1–1.3 mm in convection, and in between in the transition category). A weak third mode is apparent at 1.67 mm in stratiform rain, which is not clearly visible in other rain regimes. The occurrence of peaks in the present study is consistent with some of the previous numerical modeling (Valdez and Young 1985; Brown 1988) and experimental studies (Steiner and Waldvogel 1987). Nevertheless, the peaks at the extreme ends of the DSD, that is, 0.2 mm (obtained by numerical models), 2.3 [modeled by McFarquhar (2004)], and 3.2 mm [disdrometer observations in heavy rain by Steiner and Waldvogel (1987)], are not observed in the present study.
  5. The radar observations show similar MP occurrence profiles during the SWM and NEM, indicating only slight seasonal variation in MP occurrence. The peak occurrence distribution at the surface is also strikingly similar during the SWM and NEM, except for a small shift in diameter in the transition category.

REFERENCES

  • Brown Jr., P. S., 1988: The effects of filament, sheet, and disk breakup upon the drop spectrum. J. Atmos. Sci., 45 , 712718.

  • de Beauville, C. A., , R. H. Petit, , G. Marion, , and J. P. Lacaux, 1988: Evolution of peaks in the spectral distribution of raindrops from warm isolated maritime clouds. J. Atmos. Sci., 45 , 33203332.

    • Search Google Scholar
    • Export Citation
  • Feingold, G., , and Z. Levin, 1986: The lognormal fit to raindrop spectra from frontal convective clouds in Israel. J. Climate Appl. Meteor., 25 , 13461363.

    • Search Google Scholar
    • Export Citation
  • Feingold, G., , S. Tzivion, , and Z. Levin, 1988: Evolution of raindrop spectra. Part I: Solution to the stochastic collection/breakup equation using the method of moments. J. Atmos. Sci., 45 , 33873399.

    • Search Google Scholar
    • Export Citation
  • Garcia-Garcia, F., , and J. E. Gonzalez, 2000: Raindrop spectra observations from convective showers in the valley of Mexico. Proc. 13th Int. Conf. on Clouds and Precipitation, Reno, NV, International Commission on Cloud Physics and International Association for Meteorology and Precipitation, 398–401.

    • Search Google Scholar
    • Export Citation
  • Gossard, E. E., , R. G. Strauch, , and R. R. Rogers, 1990: Evolution of dropsize distribution in liquid precipitation observed by ground-based Doppler radar. J. Atmos. Oceanic Technol., 7 , 815828.

    • Search Google Scholar
    • Export Citation
  • Hallett, J., , and S. C. Mossop, 1974: Production of secondary ice particles during the riming process. Nature, 249 , 2628.

  • Hu, Z., , and R. C. Srivastava, 1995: Evolution of raindrop size distribution by coalescence, breakup, and evaporation: Theory and observations. J. Atmos. Sci., 52 , 17611783.

    • Search Google Scholar
    • Export Citation
  • Jameson, A. R., , and A. B. Kostinski, 1998: Fluctuation properties of precipitation. Part II: Reconsideration of the meaning and measurement of raindrop size distributions. J. Atmos. Sci., 55 , 283294.

    • Search Google Scholar
    • Export Citation
  • Joss, J., , and A. Waldvogel, 1967: Ein Spektrograph für Niederschlagstropfen mit automatischer Auswertung. Pure Appl. Geophys., 68 , 240246.

    • Search Google Scholar
    • Export Citation
  • Joss, J., , and A. Waldvogel, 1969: Raindrop size distribution and sampling size errors. J. Atmos. Sci., 26 , 566569.

  • Kozu, T., , K. K. Reddy, , S. Mori, , M. Thurai, , J. T. Ong, , and D. N. Rao, 2006: Seasonal and diurnal variations of raindrop size distribution in Asian monsoon region. J. Meteor. Soc. Japan, 84A , 195209.

    • Search Google Scholar
    • Export Citation
  • List, R., , and M. McFarquhar, 1990: The role of breakup and coalescence in the three-peak equilibrium distribution of raindrops. J. Atmos. Sci., 47 , 22742292.

    • Search Google Scholar
    • Export Citation
  • List, R., , N. R. Donaldson, , and R. E. Stewart, 1987: Temporal evolution of drop spectra to collisional equilibrium in steady and pulsating rain. J. Atmos. Sci., 44 , 362372.

    • Search Google Scholar
    • Export Citation
  • Low, T. B., , and R. List, 1982: Collision, coalescence and breakup of raindrops. Part I: Experimentally established coalescence efficiencies and fragment size distributions in breakup. J. Atmos. Sci., 39 , 15911606.

    • Search Google Scholar
    • Export Citation
  • Lucas, C., , A. D. MacKinnon, , R. A. Vincent, , and P. T. May, 2004: Raindrop size distribution retrievals from a VHF boundary layer profiler. J. Atmos. Oceanic Technol., 21 , 4560.

    • Search Google Scholar
    • Export Citation
  • Marshall, J. S., , and W. M. Palmer, 1948: The distribution of raindrops with size. J. Meteor., 5 , 165166.

  • McFarquhar, G. M., 2004: A new representation of collision-induced breakup of raindrops and its implications for the shapes of raindrop size distributions. J. Atmos. Sci., 61 , 777794.

    • Search Google Scholar
    • Export Citation
  • McFarquhar, G. M., , and R. List, 1993: The effect of curve fits for the disdrometer calibration on raindrop spectra, rainfall rate, and radar reflectivity. J. Appl. Meteor., 32 , 774782.

    • Search Google Scholar
    • Export Citation
  • Mossop, S. C., 1976: Production of secondary ice particles during the growth of graupel by riming. Quart. J. Roy. Meteor. Soc., 102 , 4557.

    • Search Google Scholar
    • Export Citation
  • Nissen, R., , R. List, , D. Hudak, , G. M. McFarquhar, , R. P. Lawson, , N. P. Tung, , S. K. Soo, , and T. S. Kang, 2005: Constant raindrop fall speed profiles derived from Doppler radar data analyses for steady nonconvective precipitation. J. Atmos. Sci., 62 , 220230.

    • Search Google Scholar
    • Export Citation
  • Prat, O. P., , and A. P. Barros, 2007: A robust numerical solution of the stochastic collection–breakup equation for warm rain. J. Appl. Meteor. Climatol., 46 , 14801497.

    • Search Google Scholar
    • Export Citation
  • Radhakrishna, B., , and T. N. Rao, 2009: Multipeak raindrop size distribution observed by UHF/VHF wind profilers during the passage of a mesoscale convective system. Mon. Wea. Rev., 137 , 978992.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , D. N. Rao, , K. Mohan, , and S. Raghavan, 2001: Classification of tropical precipitating systems and associated ZR relationships. J. Geophys. Res., 106 , 1769917711.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , N. V. P. Kirankumar, , B. Radhakrishna, , and D. N. Rao, 2006: On the variability of the shape-slope parameter relations of the gamma raindrop size distribution model. Geophys. Res. Lett., 33 , L22809. doi:10.1029/2006GL028440.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , N. V. P. Kirankumar, , B. Radhakrishna, , D. N. Rao, , and K. Nakamura, 2008: Classification of tropical precipitating systems using wind profiler spectral moments. Part I: Algorithm description and validation. J. Atmos. Oceanic Technol., 25 , 884897.

    • Search Google Scholar
    • Export Citation
  • Rao, T. N., , B. Radhakrishna, , K. Nakamura, , and N. P. Rao, 2009: Differences in raindrop size distribution from southwest monsoon to northeast monsoon at Gadanki. Quart. J. Roy. Meteor. Soc., 135 , 16301637.

    • Search Google Scholar
    • Export Citation
  • Rosenfeld, D., , and C. W. Ulbrich, 2003: Cloud microphysical properties, processes, and rainfall estimation opportunities. Radar and Atmospheric Science: A Collection of Essays in Honor of David Atlas, Meteor. Monogr., No. 52, Amer. Meteor. Soc., 237–258.

    • Search Google Scholar
    • Export Citation
  • Sauvageot, H., , and M. Koffi, 2000: Multimodal raindrop size distributions. J. Atmos. Sci., 57 , 24802492.

  • Sheppard, B. E., 1990: Effect of irregularities in the diameter classification of raindrops by the Joss–Waldvogel disdrometer. J. Atmos. Oceanic Technol., 7 , 180183.

    • Search Google Scholar
    • Export Citation
  • Steiner, M., , and A. Waldvogel, 1987: Peaks in raindrop size distributions. J. Atmos. Sci., 44 , 31273133.

  • Testik, F. Y., , and A. P. Barros, 2007: Toward elucidating the microstructure of rainfall: A survey. Rev. Geophys., 45 , RG2003. doi:10.1029/2005RG000182.

    • Search Google Scholar
    • Export Citation
  • Tokay, A., , and D. A. Short, 1996: Evidence from tropical raindrop spectra of the origin of rain from stratiform versus convective clouds. J. Appl. Meteor., 35 , 355371.

    • Search Google Scholar
    • Export Citation
  • Tokay, A., , A. Kruger, , and W. F. Krajewski, 2001: Comparison of drop size distribution measurements by impact and optical disdrometers. J. Appl. Meteor., 40 , 20832097.

    • Search Google Scholar
    • Export Citation
  • Ulbrich, C. W., 1983: Natural variations in the analytical form of the raindrop size distribution. J. Climate Appl. Meteor., 22 , 17641775.

    • Search Google Scholar
    • Export Citation
  • Uma, K. N., , and T. N. Rao, 2009: Characteristics of vertical velocity cores in different convective systems observed over Gadanki, India. Mon. Wea. Rev., 137 , 953977.

    • Search Google Scholar
    • Export Citation
  • Valdez, M. P., , and K. C. Young, 1985: Number fluxes in equilibrium raindrop populations: A Markov chain analysis. J. Atmos. Sci., 42 , 10241036.

    • Search Google Scholar
    • Export Citation
  • Williams, C. R., , A. Kruger, , K. S. Gage, , A. Tokay, , R. Cifelli, , W. F. Krajewski, , and C. Kummerow, 2000: Comparison of simultaneous raindrop size distributions estimated from two surface disdrometers and a UHF profiler. Geophys. Res. Lett., 27 , 17631766.

    • Search Google Scholar
    • Export Citation
  • Willis, P. T., 1984: Functional fits to some observed drop size distributions and parameterization of rain. J. Atmos. Sci., 41 , 16481661.

    • Search Google Scholar
    • Export Citation
  • Zawadzki, I., , and M. De Agostinho Antonio, 1988: Equilibrium raindrop size distributions in tropical rain. J. Atmos. Sci., 45 , 34523459.

    • Search Google Scholar
    • Export Citation
  • Zawadzki, I., , F. Fabry, , and W. Szyrmer, 2001: Observations of supercooled water and secondary ice generation by a vertically pointing X-band Doppler radar. Atmos. Res., 59–60 , 343359.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

An illustration of the peak identification scheme from the Doppler power spectrum of LAWP at (a) 2.4 and (b) 5.7 km (in convection), and (c) the disdrometer-derived surface DSD. The arrows indicate the peaks identified by the algorithms.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 2.
Fig. 2.

Two typical examples demonstrating the schemes employed here to classify precipitating systems. Time–height contours of reflectivity and Doppler velocity on (a),(b) 26 Aug 1999 and (d),(e) 2 Jun 2000, retrieved from LAWP. The open symbols represent cold rain, while the solid symbols represent warm rain. Different symbols represent different types of rain (square, convection; circle, stratiform; hexagon, transition inclusive; diamond, transition exclusive). (c),(f) The temporal variation of the rain integral parameters (Z, R and D0) and D0/R retrieved from collocated JWD measurements. The definition of symbols the is the same as for (a),(b).

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 3.
Fig. 3.

(a),(b) As in Figs. 2a and 2b, but for 5 Aug 2000 showing the passage of a convective storm. Sequences of Doppler power spectra showing MPs during (c) 1629–1639 and (d) 1705–1719 LT. The shading indicates the backscattered power at each spectra bin. The line spectrum at each altitude is normalized by the maximum value in that particular altitude.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 4.
Fig. 4.

(a),(b) As in Figs. 2a and 2b, but for 10 Aug 2000 during the passage of a stratiform rain system. (c) Sequence of Doppler power spectra showing MPs during 0008–0053 LT.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 5.
Fig. 5.

As in Fig. 3, but for 6 Aug 2000, showing the occurrence of MPs during the passage of an MCS.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 6.
Fig. 6.

As in Fig. 3, but for 14 Jul 1999, showing the occurrence of MPs during the passage of a warm rain system.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 7.
Fig. 7.

(left) Altitude distribution of the percentages of occurrence of single (dash–dot line), two (solid line), and multiple (more than two) (dotted line) peaks. The scale for single (two and more than two) peak(s) occurrence is provided below (above). (right) Altitude variations of the numbers of rain data points.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 8.
Fig. 8.

As in Fig. 7 but only for two and more than two peaks in (a) cold convection, (b) transition-inclusive, (c) transition-exclusive, and (d) cold stratiform rain.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 9.
Fig. 9.

As in Fig. 8 but only for warm rain (a) convection and (d) stratus rain.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 10.
Fig. 10.

Percentage occurrence of peaks (solid line with circle) as a function of drop diameter in (top) stratiform, (middle) transition, and (bottom) convective precipitation. The numbers of data points (right axis) in each channel number used to calculate the percentage of occurrence are also included.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 11.
Fig. 11.

As in Fig. 7, but for the SWM and NEM.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Fig. 12.
Fig. 12.

As in Fig. 10, but for the SWM and NEM.

Citation: Monthly Weather Review 137, 10; 10.1175/2009MWR2967.1

Table 1.

Important specifications and parameters of the LAWP.

Table 1.
Table 2.

Number of rain and multipeaked profiles and corresponding percentages of occurrence of MPs in different types of precipitating systems from the LAWP data.

Table 2.
Table 3.

Number of rain and multipeaked (two and more than two) spectra and the corresponding percentages of occurrence of MPs in different types of precipitating systems from JWD data.

Table 3.
Table 4.

As in Table 3 but for the SWM and NEM.

Table 4.
Save