• Atlas, D., , K. R. Hardy, , R. Wexler, , and R. Boucher, 1963: On the origin of hurricane spiral bands. Geofis. Int., 3, 123132.

  • Awaka, J., , T. Iguchi, , and K. Okamoto, 2009: TRMM PR standard algorithm 2A23 and its performance on bright band detection. J. Meteor. Soc. Japan, 87A, 3152.

    • Search Google Scholar
    • Export Citation
  • Barnes, G. M., , E. J. Zipser, , D. Jorgensen, , and F. Marks Jr., 1983: Mesoscale and convective structure of a hurricane rainband. J. Atmos. Sci., 40, 21252137.

    • Search Google Scholar
    • Export Citation
  • Black, M. L., , J. F. Gamache, , F. D. Marks, , C. E. Samsury, , and H. E. Willoughby, 2002: Eastern Pacific Hurricanes Jimena of 1991 and Olivia of 1994: The effect of vertical shear on structure and intensity. Mon. Wea. Rev., 130, 22912312.

    • Search Google Scholar
    • Export Citation
  • Black, R. A., , and J. Hallett, 1986: Observations of the distribution of ice in hurricanes. J. Atmos. Sci., 43, 802822.

  • Black, R. A., , and J. Hallett, 1999: Electrification of the hurricane. J. Atmos. Sci., 56, 20042028.

  • Braun, S. A., , M. T. Montgomery, , and Z. Pu, 2006: High-resolution simulation of Hurricane Bonnie (1998). Part I: The organization of eyewall vertical motion. J. Atmos. Sci., 63, 1942.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., , E. J. Zipser, , and S. W. Nesbitt, 2002: Reflectivity, ice scattering, and lightning characteristics of hurricane eyewalls and rainbands. Part I: Quantitative description. Mon. Wea. Rev., 130, 769784.

    • Search Google Scholar
    • Export Citation
  • Chen, S. S., , J. A. Knaff, , and F. D. Marks, 2006: Effects of vertical wind shear and storm motion on tropical cyclone rainfall asymmetries deduced from TRMM. Mon. Wea. Rev., 134, 31903208.

    • Search Google Scholar
    • Export Citation
  • Chen, Y., , and M. K. Yau, 2001: Spiral bands in a simulated hurricane. Part I: Vortex Rossby wave verification. J. Atmos. Sci., 58, 21282145.

    • Search Google Scholar
    • Export Citation
  • Corbet, J., , C. Mueller, , C. Burghart, , K. Gould, , and G. Granger, 1994: Zeb: Software for geophysical data integration, display, and management of diverse environmental datasets. Bull. Amer. Meteor. Soc., 75, 783792.

    • Search Google Scholar
    • Export Citation
  • Corbosiero, K. L., , and J. Molinari, 2002: The effects of vertical wind shear on the distribution of convection in tropical cyclones. Mon. Wea. Rev., 130, 21102123.

    • Search Google Scholar
    • Export Citation
  • Corbosiero, K. L., , J. Molinari, , A. R. Aiyyer, , and M. L. Black, 2006: The structure and evolution of Hurricane Elena (1985). Part II: Convective asymmetries and evidence for vortex Rossby waves. Mon. Wea. Rev., 134, 30733091.

    • Search Google Scholar
    • Export Citation
  • DeCarlo, L. T., 1997: On the meaning and use of kurtosis. Psychol. Methods, 2, 292307.

  • Didlake, A. C. Jr., , and R. A. Houze Jr., 2009: Convective-scale downdrafts in the principal rainband of Hurricane Katrina (2005). Mon. Wea. Rev., 137, 32693293.

    • Search Google Scholar
    • Export Citation
  • Franklin, C. N., , G. J. Holland, , and P. T. May, 2006: Mechanisms for the generation of mesoscale vorticity features in tropical cyclone rainbands. Mon. Wea. Rev., 134, 26492669.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., , and R. A. Houze Jr., 2008: Kinematic structure of convective-scale elements in the rainbands of Hurricanes Katrina and Rita (2005). J. Geophys. Res., 113, D15108, doi:10.1029/2007JD009429.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., , and R. A. Houze Jr., 2011: Vertical structure of hurricane eyewalls as seen by the TRMM Precipitation Radar. J. Atmos. Sci., 68, 16371652.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., , and R. A. Houze Jr., 2012: Vertical structure of tropical cyclones with concentric eyewalls as seen by the TRMM Precipitation Radar. J. Atmos. Sci., 69, 10211036.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., 1997: Stratiform precipitation in regions of convection: A meteorological paradox? Bull. Amer. Meteor. Soc., 78, 21792196.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., 2004: Mesoscale convective systems. Rev. Geophys., 42, RG4003, doi:10.1029/2004RG000150.

  • Houze, R. A., Jr., 2010: Clouds in tropical cyclones. Mon. Wea. Rev., 138, 293344.

  • Houze, R. A., Jr., , D. C. Wilton, , and B. F. Smull, 2007: Monsoon convection in the Himalayan region as seen by the TRMM Precipitation Radar. Quart. J. Roy. Meteor. Soc., 133, 13891411.

    • Search Google Scholar
    • Export Citation
  • James, C. N., , S. R. Brodzik, , H. Edmon, , R. A. Houze Jr., , and S. E. Yuter, 2000: Radar data processing and visualization over complex terrain. Wea. Forecasting, 15, 327338.

    • Search Google Scholar
    • Export Citation
  • Jordan, C. L., 1958: Mean soundings for the West Indies area. J. Meteor., 15, 9197.

  • Judt, F., , and S. S. Chen, 2010: Convectively generated potential vorticity in rainbands and formation of the secondary eyewall in Hurricane Rita of 2005. J. Atmos. Sci., 67, 35813599.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471.

  • Knapp, K. R., , M. C. Kruk, , D. H. Levinson, , H. J. Diamond, , and C. J. Neumann, 2010: The International Best Track Archive for Climate Stewardship (IBTrACS). Bull. Amer. Meteor. Soc., 91, 363376.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., , W. Barnes, , T. Kozu, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 809817.

    • Search Google Scholar
    • Export Citation
  • Marks, F. D., , and R. A. Houze Jr., 1987: Inner core structure of Hurricane Alicia from airborne Doppler radar observations. J. Atmos. Sci., 44, 12961317.

    • Search Google Scholar
    • Export Citation
  • May, P. T., , and G. J. Holland, 1999: The role of potential vorticity generation in tropical cyclone rainbands. J. Atmos. Sci., 56, 12241228.

    • Search Google Scholar
    • Export Citation
  • Molinari, J., , P. Moore, , and V. Idone, 1999: Convective structure of hurricanes as revealed by lightning locations. Mon. Wea. Rev., 127, 520534.

    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., , and R. J. Kallenbach, 1997: A theory for vortex Rossby-waves and its application to spiral bands and intensity changes in hurricanes. Quart. J. Roy. Meteor. Soc., 123, 435465.

    • Search Google Scholar
    • Export Citation
  • Powell, M. D., 1990a: Boundary layer structure and dynamics in outer hurricane rainbands. Part I: Mesoscale rainfall and kinematic structure. Mon. Wea. Rev., 118, 891917.

    • Search Google Scholar
    • Export Citation
  • Powell, M. D., 1990b: Boundary layer structure and dynamics in outer hurricane rainbands. Part II: Downdraft modification and mixed layer recovery. Mon. Wea. Rev., 118, 918938.

    • Search Google Scholar
    • Export Citation
  • Qiu, X., , Z.-M. Tan, , and Q. Xiao, 2010: The roles of vortex Rossby waves in hurricane secondary eyewall formation. Mon. Wea. Rev., 138, 20922109.

    • Search Google Scholar
    • Export Citation
  • Rogers, R., , S. Chen, , J. Tenerelli, , and H. Willoughby, 2003: A numerical study of the impact of vertical shear on the distribution of rainfall in Hurricane Bonnie (1998). Mon. Wea. Rev., 131, 15771599.

    • Search Google Scholar
    • Export Citation
  • Saffir, H. S., 2003: Communicating damage potentials and minimizing hurricane damage. Hurricane! Coping with Disaster, R. Simpson, Ed., Amer. Geophys. Union, 155–164.

  • Samsury, C. E., , and E. J. Zipser, 1995: Secondary wind maxima in hurricanes: Airflow and relationship to rainbands. Mon. Wea. Rev., 123, 35023517.

    • Search Google Scholar
    • Export Citation
  • Schultz, L. A., , and D. J. Cecil, 2009: Tropical cyclone tornadoes, 1950–2007. Mon. Wea. Rev., 137, 34713484.

  • Terwey, W. D., , and M. T. Montgomery, 2008: Secondary eyewall formation in two idealized, full-physics modeled hurricanes. J. Geophys. Res., 113, D12112, doi:10.1029/2007JD008897.

    • Search Google Scholar
    • Export Citation
  • TSDIS, 2007: File specifications for TRMM products levels 2 and 3. Vol. 4, Interface Control Specification between the Tropical Rainfall Measuring Mission Science Data and Information System (TSDIS) and the TSDIS Science User (TSU). NASA GSFC, TSDIS-P907, 102 pp. [Available online at http://pps.gsfc.nasa.gov/tsdis/Documents/ICSVol4.pdf.]

  • Willoughby, H. E., 1988: The dynamics of the tropical cyclone core. Aust. Meteor. Mag., 36, 183191.

  • Willoughby, H. E., , F. D. Marks, , and R. J. Feinberg, 1984: Stationary and moving convective bands in hurricanes. J. Atmos. Sci., 41, 31893211.

    • Search Google Scholar
    • Export Citation
  • Yuter, S. E., , and R. A. Houze Jr., 1995: Three-dimensional kinematic and microphysical evolution of Florida cumulonimbus. Part II: Frequency distributions of vertical velocity, reflectivity, and differential reflectivity. Mon. Wea. Rev., 123, 19411963.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Idealized plan-view radar signature of the inner and distant rainbands of a tropical cyclone. The shading is for threshold values of 30, 35, 37.5, 40, and 45 dBZ. The small hurricane symbol represents the center of the cyclone. The first reflectivity ring represents the primary eyewall (region 1). The size of the convective cells indicates the level of maturity, with the dashed border indicating collapsing cells. The dotted circle represents the boundary between the inner- and distant-rainband regions. The dashed arrow represents the environmental shear vector. Line AB is a cross section idealized in Fig. 2.

  • View in gallery

    Idealized vertical cross section along line AB of Fig. 1. Scalloped region represents the cloud boundary of the convective features. The shading is for threshold values of 25, 30, 35, 37.5, 40, and 45 dBZ. The open arrows represent the flow of ice outward from the eyewall region.

  • View in gallery

    CFADs of TRMM PR reflectivity data for all of the 1998–2007 overpasses of storms that ultimately reached category 4 or 5 intensity (CAT12345; see Table 1), for the total points of the (a) eyewall (region 1), (b) inner rainbands (regions 3–5), and (c) distant rainbands (regions 6–9). Contours represent the frequency of occurrence relative to the maximum absolute frequency in the data sample represented in the CFAD, contoured every 5%. Altitudes are geopotential height (km MSL) relative to the ellipsoidal surface of the earth. The ordinate of the CFAD is altitude (250-m increments or bins) and the abscissa is reflectivity (dBZ; 1-dB bins). The 8-km and 25-dBZ levels are indicated by the black solid lines, and the 5-km and 30-dBZ levels are indicated by the black dotted lines for ease of reference. The 20%, 50%, and 80% contours are black for reference.

  • View in gallery

    (a) Percentage of areal coverage of precipitating pixels as a function of height for all overpasses of regions 3–9 (CAT12345; see Table 1). The inner rainbands are red and the distant rainbands blue for reference. (b) The total region fraction of convectively classified precipitating pixels as a function of distance. The total sample grouping is black; the groupings are defined in Table 1.

  • View in gallery

    (a) Mean reflectivity as a function of height for CFADs of all overpasses of regions 3–9 (CAT12345; see Table 1). The inner rainbands are red and the distant rainbands are blue for reference. (b)–(d) As in (a), but for standard deviation, skewness, and kurtosis, respectively. The mean reflectivity is the conditional reflectivity (i.e., it is the mean reflectivity for pixels at which a detectable reflectivity exists). Zero reflectivities do not enter the mean.

  • View in gallery

    (a) Mean reflectivity as a function of height of 1000 randomly chosen subsamples of the total region 3 (red lines) and region 9 (blue lines) data. (b) As in (a), but for standard deviation.

  • View in gallery

    (a) Percentage of areal coverage of precipitating pixels as a function of height for all overpasses of the inner- and distant-rainband quadrants (CAT12345; see Table 1). The quadrants are aligned along the average environmental wind shear vector of the sample. The inner rainbands are red and the distant rainbands are blue for reference. (b) The quadrant fraction of convectively classified precipitating pixels as a function of distance. The quadrants are aligned along the average environmental wind shear vector of the sample.

  • View in gallery

    Region 3–5 inner rainband CFADs of TRMM PR reflectivity data for all of the 1998–2007 overpasses of storms that ultimately reached category 4 or 5 intensity (CAT12345 grouping; Table 1), arranged by quadrants oriented to the average shear vector. Quadrants are labeled downshear left (DL), downshear right (DR), upshear left (UL), and upshear right (UR). All other details are as in Fig. 3.

  • View in gallery

    (a) Mean reflectivity as a function of height of the inner-rainband-quadrant CFADs from Fig. 8 (CAT12345; Table 1). The mean reflectivity is the conditional reflectivity (i.e., it is the mean reflectivity for pixels at which a detectable reflectivity exists). Zero reflectivities do not enter the mean. (b) As in (a), but for standard deviation. (c) As in (a), but for kurtosis. Quadrants are labeled as in Fig. 8.

  • View in gallery

    Region 6–9 outer rainband CFADs of TRMM PR reflectivity data for all of the 1998–2007 overpasses of storms that ultimately reached category 4 or 5 intensity (CAT12345 grouping; Table 1), arranged by quadrants oriented to the average shear vector. Quadrants are labeled as in Fig. 8. All other details are as in Fig. 3.

  • View in gallery

    (a) Mean reflectivity as a function of height of the outer-rainband-quadrant CFADs from Fig. 10 (CAT12345; see Table 1). (b) As in (a), but for standard deviation. (c) As in (a), but for kurtosis. Quadrants are labeled as in Fig. 8.

  • View in gallery

    (a) Mean reflectivity as a function of height of the total region CFADs for the inner- and distant-rainband regions of the CAT12 and CAT45 groupings. (b) Kurtosis as a function of height of the CFADs for the inner- and distant-rainband regions of the CAT12 and CAT45 groupings. (c) As in (a), but for marginal- and high-SST groupings. (d) As in (b), but for the marginal- and high-SST groupings. See Table 1 for overpass information of groupings.

  • View in gallery

    (a) Mean reflectivity as a function of height of the quadrant CFADs for the inner-rainband regions of the low-shear grouping. (b) Kurtosis as a function of height of the quadrant CFADs for the inner-rainband regions of the low-shear grouping. (c) As in (a), but for the high-shear grouping. (d) As in (b), but for the high-shear grouping. See Table 1 for overpass information of groupings.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 219 219 41
PDF Downloads 150 150 22

Vertical Structure of Tropical Cyclone Rainbands as Seen by the TRMM Precipitation Radar

View More View Less
  • 1 University of Washington, Seattle, Washington
© Get Permissions
Full access

Abstract

Ten years of data from the Tropical Rainfall Measurement Mission satellite’s Precipitation Radar (TRMM PR) show the vertical structure of tropical cyclone rainbands. Radar-echo statistics show that rainbands have a two-layered structure, with distinct modes separated by the melting layer. The ice layer is a combination of particles imported from the eyewall and ice left aloft as convective cells collapse. This layering is most pronounced in the inner region of the storm, and the layering is enhanced by storm strength. The inner-region rainbands are vertically confined by outflow from the eyewall but nevertheless are a combination of strong embedded convective cells and robust stratiform precipitation, both of which become more pronounced in stronger cyclones.

Changes in rainband coverage, vertical structure, and the amount of active convection indicate a change in the nature of rainbands between the regions inward and outward of a radius of approximately 200 km. Beyond this radius, rainbands consist of more sparsely distributed precipitation that is more convective in nature than that of the inner-region rainbands, and the outer-region rainband structures are relatively insensitive to changes in storm intensity. The rainbands in both inner and outer regions are organized with respect to the environmental wind shear vector. The right-of-shear quadrants contain newer convection while in the left-of-shear quadrants the radar echoes are predominantly stratiform. This asymmetric distribution of rainband structures strengthens with environmental wind shear. Cool sea surfaces discourage rainband convection uniformly.

Corresponding author address: Deanna Hence, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195. E-mail: dhence@atmos.washington.edu

Abstract

Ten years of data from the Tropical Rainfall Measurement Mission satellite’s Precipitation Radar (TRMM PR) show the vertical structure of tropical cyclone rainbands. Radar-echo statistics show that rainbands have a two-layered structure, with distinct modes separated by the melting layer. The ice layer is a combination of particles imported from the eyewall and ice left aloft as convective cells collapse. This layering is most pronounced in the inner region of the storm, and the layering is enhanced by storm strength. The inner-region rainbands are vertically confined by outflow from the eyewall but nevertheless are a combination of strong embedded convective cells and robust stratiform precipitation, both of which become more pronounced in stronger cyclones.

Changes in rainband coverage, vertical structure, and the amount of active convection indicate a change in the nature of rainbands between the regions inward and outward of a radius of approximately 200 km. Beyond this radius, rainbands consist of more sparsely distributed precipitation that is more convective in nature than that of the inner-region rainbands, and the outer-region rainband structures are relatively insensitive to changes in storm intensity. The rainbands in both inner and outer regions are organized with respect to the environmental wind shear vector. The right-of-shear quadrants contain newer convection while in the left-of-shear quadrants the radar echoes are predominantly stratiform. This asymmetric distribution of rainband structures strengthens with environmental wind shear. Cool sea surfaces discourage rainband convection uniformly.

Corresponding author address: Deanna Hence, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195. E-mail: dhence@atmos.washington.edu

1. Introduction

Precipitation bands of varying size and extent typically spiral inward toward the center of a mature tropical cyclone. The most prominent of these rainbands, known as the principal rainband, makes up most of the “stationary band complex” (Willoughby et al. 1984; Willoughby 1988; Houze 2010), named for the complex’s tendency to remain in the same place relative to the translating storm center. The remainder of the stationary band complex consists of smaller secondary bands in the inner core that are often found radially inward of the principal band, and loosely organized convection that forms the distant rainbands in the storm’s environment (Houze 2010). These studies emphasize the horizontal distributions of rainbands, but information is needed on their vertical structures to understand their dynamical roles.

Case studies have noted certain features of the vertical structure and dynamics of rainbands. Specifically, their upwind tips are more convective and their downwind ends more stratiform (Atlas et al. 1963; Barnes et al. 1983; Hence and Houze 2008), convective cells have an outward-leaning tilt similar to that of the eyewall (Barnes et al. 1983; Powell 1990a,b; Hence and Houze 2008), and distinctly different microphysical and electrification properties occur in the inner and distant rainbands (Black and Hallett 1986; Molinari et al. 1999). The aggregate kinematic, dynamic, and thermodynamic impact of the rainband convection may affect the overall storm evolution (Powell 1990a,b; Samsury and Zipser 1995; May and Holland 1999; Franklin et al. 2006; Hence and Houze 2008; Didlake and Houze 2009).

Theory suggests that the secondary rainbands likely originate from propagating vortex Rossby waves (Montgomery and Kallenbach 1997; Corbosiero et al. 2006). Although Willoughby et al. (1984) proposed that the principal rainband marks and creates a boundary between the storm and outside environment, the dynamical origin of the principal rainband remains unclear. Nevertheless, recent theoretical and observational studies suggest that once the rainband complex forms, rainbands play an important role in changing the dynamic structure of the storm (Montgomery and Kallenbach 1997; Chen and Yau 2001), especially in the formation of concentric eyewalls (Terwey and Montgomery 2008; Judt and Chen 2010; Qiu et al. 2010).

Given the important role rainbands play in the overall dynamics of tropical cyclones, as indicated by these case studies, the objective of this study is to know statistically how the vertical structures of rainbands are distributed around a typical tropical cyclone. We accomplish this objective by statistically analyzing data collected by the Tropical Rainfall Measurement Mission satellite’s Precipitation Radar (TRMM PR; Kummerow et al. 1998) in the Atlantic and northwestern (NW) Pacific basins from 1997 to 2007. Although TRMM obtains only snapshot samples of data about 2–3 times per day over any given region, the 10-yr data sample provides long-term statistics on the structures of storms sampled. The vertical precipitation distribution retrieved by the active TRMM PR reveals vertical structure information unavailable in other satellite data, and we have used this dataset to reveal the structures of eyewalls (Hence and Houze 2011) and of the secondary eyewalls in storms undergoing eyewall replacement (Hence and Houze 2012). Combined with these two previous papers, the present study completes a trilogy by examining the rainbands that lie outside of the eyewall region.

After describing our dataset and methods of analysis in section 2, we will present in section 3 a schematic view of the typical rainband structure revealed by our analysis. This schematic will provide a point of reference for visualizing the detailed statistical results presented in subsequent sections. Section 4 details the general statistical characteristics of the rainband complex in contrast to the eyewall region of the storm. Sections 5 and 6 will present detailed statistics on the vertical structures of rainband radar echoes as a function of radial distance and azimuthal quadrant relative to the environmental shear vector. Section 7 examines the impact of the storm and its environment on the general rainband structure. In section 8 we synthesize the results and present our conclusions.

2. Data and methods of analysis

We used the TRMM PR version 6 2A25 radar reflectivity data (TSDIS 2007) to obtain a three-dimensional view of reflectivity structure. The ~250-m resolution (at nadir) of the PR makes it ideal for evaluating changes in the vertical distribution of precipitation. The ~215-/247-km swath width (before/after the boost in orbital altitude that occurred in August 2001) and the roughly twice daily sampling (for a given location) provide numerous overpasses of tropical cyclones. The horizontal resolution is 4.3/5 km (pre-/postboost). Since the focus of this study is on the vertical structure of precipitation, the change in horizontal resolution does not affect the results significantly; however, the larger swath width advantageously provides a somewhat more complete view of an individual storm.

Following Houze et al. (2007), we remapped the PR reflectivity data onto a Cartesian grid after applying small corrections to the geolocation of the upper-level data. Cartesian gridding allows for visualization in the National Center for Atmospheric Research (NCAR) Zebra software (Corbet et al. 1994; James et al. 2000) and facilitates computation of contoured frequency by altitude diagrams (CFADs; Yuter and Houze 1995). Individual overpass CFADs were then sorted into the groupings shown in Table 1. The data were also subdivided into convective or stratiform categories according to the TRMM version 6 2A23 product (TSDIS 2007; Awaka et al. 2009). Note that “convective” and “stratiform” precipitation carry specific dynamical definitions (Houze 1997) that are more appropriate to mesoscale convective systems (Houze 2004) than to tropical cyclones. Nevertheless, we use these terms to identify changes in the characteristics of the radar echo as we have done previously (Hence and Houze 2011).

Table 1.

Groupings of individual overpasses and the number of overpasses within each grouping.

Table 1.

To generate the CFADs, reflectivity values in the Cartesian grid were counted in 1-dB bins every 0.25 km in height, using only reflectivity data above the minimal detection of the radar (~17–18 dBZ). CFADs are joint probability distributions that allow for the accumulation of data from numerous overpasses in a single plot while taking advantage of the PR’s high resolution in the vertical. The mean, standard deviation, skewness, and excess kurtosis (kurtosis − 3; DeCarlo 1997) are calculated at every altitude on these raw CFADs. For plotting purposes, we normalized the CFADs by the maximum frequency in the sample to remove the effect of the bulk amount of radar echo in a given subset of the data, as well as to bring the maximum in the profile to the same magnitude for each computed CFAD while not changing the shape of the distribution (Houze et al. 2007; Hence and Houze 2011, 2012). A more complete discussion of this normalization is included in Hence and Houze (2011).

This study uses the storm center, intensity, track direction, storm translation speed, and eye diameter (when available) from the International Best Track Archive for Climate Stewardship (Knapp et al. 2010). For this study, we focused on storms within the Atlantic and NW Pacific basins between 1997 and 2007 that reached category 4 (59–69 m s−1; Saffir 2003) or 5 (>69 m s−1) sometime during their lifetimes. Our analysis included TRMM overpasses occurring when the storm intensity was at least category 1 (maximum sustained wind > 32 m s−1). Sea surface temperature (SST) data are from the National Oceanic and Atmospheric Administration (NOAA) Comprehensive Large Array-Data Stewardship System (CLASS), which are global SST data gridded at 50-km resolution twice weekly from 8-km global infrared satellite SST observations. The SST data are averaged within a 3° radius from the storm center. We estimated conditions and storm position at TRMM overpass times via linear interpolation between observations bracketing the time of overpass. The overpass samples are sorted by the status of the storm at the time of the overpass. All of the overpasses included in this study have the storm center contained within the TRMM Microwave Imager’s (TMI’s) swath width (Kummerow et al. 1998; 760 km preboost, 878 km postboost). The TMI 37- and 85-GHz data and the PR data (when available) were used together to determine eye diameters when necessary. In the case of an eyewall with broken echo coverage, estimates of the eye diameter were made based on the inner edge of the feature exhibiting the geometry of the arc of a circle or ellipse. For an elliptical eyewall, the major and minor axis average was used to define eyewall size. We visually checked the storm center from the best-track data against the PR and TMI data; if necessary, we manually shifted the center of the analysis to better align with the precipitation features.

To isolate purely rainband features, we examined the 371 overpasses of single eyewall cases discussed in Hence and Houze (2012) and did not include any concentric eyewall cases (to avoid mistakenly identifying secondary eyewalls as rainbands). The storm center and eye diameter reports determined the eye radius Re, which marks the inner boundary for the eyewall region 1. Following Hence and Houze (2011), the storms were divided into a series of annuli of increasing diameter. We defined the distant edge of the first annulus R1 by assuming a 45° slope of the eyewall (Marks and Houze 1987) with flow up to a tropopause height of 17 km (Jordan 1958), defining R1 as R1 = Re + 17 km. Eight subsequent annuli were defined by their distant boundaries, which are multiples of R1, such that R2 = 2R1, R3 = 3R1, . . . , R9 = 9R1. This study focused on the statistics of the annular regions 3–9. Analyses of the inner three regions are discussed in Hence and Houze (2011, 2012).

This study also uses a quadrant-by-quadrant analysis to study how features vary around the storm. The quadrants are oriented with respect to the 850–200-hPa shear vector, calculated from the National Centers for Environmental Prediction (NCEP) reanalysis zonal u and meridional υ winds (Kalnay et al. 1996). The 850-hPa wind vectors were subtracted from the 200-hPa wind vectors at every point within a ring of wind data 500–750 km from the storm center to avoid the influence of the storm’s circulation. These individual shear vectors were then averaged to create the mean shear vector and interpolated (using the interpolation method described above) to estimate the shear at the time of the overpass. The quadrants are then defined counterclockwise from the direction of this mean shear vector.

Further description of this technique, as well as an illustrated example of the region and quadrant definitions, is in Hence and Houze (2011). Similar to the results of Hence and Houze (2011), the variation of vertical structure relative to the storm translation was found to be small compared with variations relative to the shear vector. For the sake of brevity, this study focuses solely on the larger shear-relative signal. The shear-relative rainfall asymmetry reproduced with this technique is consistent with that seen in Chen et al. (2006) and many other previous studies. Thus we are confident that the corresponding vertical precipitation structure described in this study is consistent with that which makes the rainfall asymmetry possible.

3. Schematic model of tropical cyclone rainband structure

Figures 1 and 2 present a schematic framework of the horizontal and vertical structures of a typical storm’s rainband complex that is consistent with the statistics of the TRMM PR data. These figures serve as points of reference for visualizing the results presented in the following sections of this paper with respect to the storm. The figure indicates the radii bounding the annular regions 3–9; regions 3–5 will be referred to as the “inner” region, and regions 6–9 will comprise the “outer” region. Consistent with the idealized rainband complex presented by Houze (2010, see his Fig. 30), we suggest that the inner region is strongly influenced by the vortex dynamics, and the outer region is the ambient region of the vortex. In the schematic, a dotted line marks the boundary between the inner and outer regions. The schematic also indicates the orientation of the environmental shear vector, which is a major factor determining the distribution of features around the storm. The quadrants of the storm are defined relative to the direction of this shear vector.

Fig. 1.
Fig. 1.

Idealized plan-view radar signature of the inner and distant rainbands of a tropical cyclone. The shading is for threshold values of 30, 35, 37.5, 40, and 45 dBZ. The small hurricane symbol represents the center of the cyclone. The first reflectivity ring represents the primary eyewall (region 1). The size of the convective cells indicates the level of maturity, with the dashed border indicating collapsing cells. The dotted circle represents the boundary between the inner- and distant-rainband regions. The dashed arrow represents the environmental shear vector. Line AB is a cross section idealized in Fig. 2.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

Fig. 2.
Fig. 2.

Idealized vertical cross section along line AB of Fig. 1. Scalloped region represents the cloud boundary of the convective features. The shading is for threshold values of 25, 30, 35, 37.5, 40, and 45 dBZ. The open arrows represent the flow of ice outward from the eyewall region.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

The schematic models in Figs. 1 and 2 do not represent an individual realization of the rainband pattern or of individual finescale convective features, since it is impossible to isolate individual features in the statistics. Rather, the bulk rainband complex conceptualized here is implied by the statistical analysis to be presented below. The large shaded feature represents an ensemble of principal and secondary rainbands. In the outer region, the rainband complex is depicted as a broken line of convective cells spiraling around the tropical cyclone, as is typical for “distant rainbands” (Houze 2010). Closer to the storm center, the rainband complex becomes more intense in reflectivity and the echo coverage becomes more contiguous, with an increasing proportion of stratiform-like precipitation, especially downwind. At the border between the inner and outer regions, the convection is at its most frequent and intense. The radar echoes quickly die away upon reaching the innermost region. The rainbands closest to the eyewall are dominated by heavy stratiform-like precipitation, with broad uniform coverage and very little convective precipitation. The eyewall (region 1) is shown as a contiguous ring of intense reflectivity with an asymmetry consistent with Hence and Houze (2011, 2012), and region 2 is not represented.

In Fig. 1, the rainband complex is asymmetric with respect to the environmental shear vector, which points toward the top of the schematic. In the outer region, the right-of-shear quadrants tend to be the most intensely convective, with disorganized convective cells in the upshear quadrants organizing into the larger rainbands in the upshear right (UR). As these cells move with the low-level winds along the rainband, they grow, mature, and organize into distinct lines along the inner edge of each corresponding rainband. Once the cells reach the inner region in the downshear-right (DR) quadrant, they reach their maximum intensity and begin to collapse, as signified by the dotted lines around the reflectivity peaks. In the downshear-left (DL) quadrant, these collapsed convective cells contribute to the broad regions of intense stratiform echo, which maximizes both in intensity and coverage in this region. Some of this stratiform precipitation continues to stream over into the upshear-left (UL) quadrant.

Figure 2 shows an idealized vertical cross section along line AB of Fig. 1. It shows the cloud outline as well as the vertical distribution of radar echo as would be seen by the TRMM PR. On the left-of-shear side of the storm, the rainband echo in the inner region is predominantly stratiform with the radar echo exhibiting general uniformity and a bright band in the melting layer. In Hence and Houze (2011, 2012), we found the eyewall on the left-of-shear side to be deep and intense, producing smaller ice particles at upper levels that travel outward and swirl around the storm. The eyewall is equally intense but not as deep on the right-of-shear side, with ice cloud having less vertical and horizontal extent. The inner rainbands right-of-shear can be intensely convective, especially along a sharply defined inner edge (Barnes et al. 1983; Powell 1990a,b; Hence and Houze 2008; Didlake and Houze 2009). Stratiform-like precipitation separates this inner-edge convection from other convective cells outside of it, and stratiform precipitation dominates farther out.

Below we will see that the TRMM PR statistics are consistent with this model, and that they provide further details of the echo structures in each region of the storm.

4. General rainband structure compared to eyewalls

The rainbands of tropical cyclones differ from eyewall convection in intensity, vertical extent, and contiguity. As discussed in Hence and Houze (2011), the eyewall is characterized by frequent occurrence of high reflectivity values at all levels and is generally deep, with numerous occurrences of detectable reflectivity values at high altitudes. The eyewall further exhibits even more intense but highly intermittent echoes superimposed upon the mean structure. This structure is seen again in the overall CFAD of radar reflectivity of all the eyewalls of all the storms considered in this study (Fig. 3a).

Fig. 3.
Fig. 3.

CFADs of TRMM PR reflectivity data for all of the 1998–2007 overpasses of storms that ultimately reached category 4 or 5 intensity (CAT12345; see Table 1), for the total points of the (a) eyewall (region 1), (b) inner rainbands (regions 3–5), and (c) distant rainbands (regions 6–9). Contours represent the frequency of occurrence relative to the maximum absolute frequency in the data sample represented in the CFAD, contoured every 5%. Altitudes are geopotential height (km MSL) relative to the ellipsoidal surface of the earth. The ordinate of the CFAD is altitude (250-m increments or bins) and the abscissa is reflectivity (dBZ; 1-dB bins). The 8-km and 25-dBZ levels are indicated by the black solid lines, and the 5-km and 30-dBZ levels are indicated by the black dotted lines for ease of reference. The 20%, 50%, and 80% contours are black for reference.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

The rainband CFADs (Figs. 3b,c) have notable differences from the eyewall CFAD. First, the rainbands are shallower, with the greater-than-50%-of-maximum frequency distribution only reaching 8 km (yellow to red contours of Fig. 3; hereafter referred to as the modal distribution), and the less-than-50%-of-maximum frequency distribution (blue to green contours; hereafter referred to as the outlier distribution) only reaching 10 km. For this reason, the rainband radar echoes in the vertical schematic of Fig. 2 are indicated as not extending as high as the eyewall radar echoes.

Second, the distributions do not span as wide a range of reflectivities, with the modal distribution below 5 km being approximately 14–17 dB in width as opposed to 18–22 dB in the eyewall. There is a relative lack of intense outliers, which reach only 45 dBZ, compared with 52 dBZ in the eyewall. These CFAD results are generally consistent with the findings of Cecil et al. (2002), who compared the reflectivity cumulative density functions of the eyewall and rainband regions with oceanic and continental convection.

Third, the CFADs in the rainband regions separate sharply at the melting level (~5 km) into two regions. The upper-level region of ice particles has a mode that is abruptly lower in intensity with height compared with low-level rain mode. The melting layer and associated increased variance pinches the two frequency maxima off from each other. The upper-level mode drops off sharply in intensity with height above the melting level, and the outliers at upper levels remain concentrated close to the modal distribution. The bimodality of the rainbands, as well as the high uniformity of their upper-level mode, marks a microphysical and dynamic regime in the rainbands that is distinctly different from the eyewall. The separation of the upper- from the lower-level mode is most exaggerated in the inner-region rainbands, probably because of their proximity to the eyewall. Much of the ice at upper levels in the inner-region rainbands is expected to have been imported as exhaust and fallout from the eyewall, a process separate from the local convection producing the rainbands’ rainfall.

The TRMM PR data further indicate that the rainbands become more convective with increasing distance from the storm center, consistent with the early study of Atlas et al. (1963). Figure 4a shows the fraction of areal coverage of reflectivity for regions 3–9 (as defined in section 2). The areal coverage decreases rapidly with radius in regions 3–5 (red lines), but from region 6 outward this rate of decrease slows (blue lines). The black line in Fig. 4b shows that the fraction of precipitating pixels classified as convective by the 2A23 algorithm increases steadily with distance from the storm center, with the sharpest gradient from region 5 to region 6. These changes in coverage and the convectively classified fraction are consistent with the way we have drawn Fig. 1; each annulus from region 3 outward has successively less fractional area coverage by radar echo (consistent with Fig. 4a) and an increasing fractional coverage by convective echo (consistent with Fig. 4b). As a result, the spiral shape of the rainbands with more active convection in their upwind ends yields an increasing coverage by convective cells in region 5 of the schematic compared with region 3. These changes in coverage as a function of radius suggest that two different convective regimes govern the behavior of the rainbands as a function of distance from the storm center, and that this break between regimes occurs on average somewhere between regions 5 and 6. This regime change is one reason that we indicate the outer region as starting at about region 6 in Fig. 1. We speculate that this regime change occurs at about where the vortex dynamics begin to have less of a restrictive effect on convection associated with the storm. Given that the average eye radius of the sample is approximately 23 km (R1 = ~40 km; Hence and Houze 2012), this demarcation is on average approximately 200 km from the storm center, albeit with considerable storm-to-storm variance.

Fig. 4.
Fig. 4.

(a) Percentage of areal coverage of precipitating pixels as a function of height for all overpasses of regions 3–9 (CAT12345; see Table 1). The inner rainbands are red and the distant rainbands blue for reference. (b) The total region fraction of convectively classified precipitating pixels as a function of distance. The total sample grouping is black; the groupings are defined in Table 1.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

5. Rainbands in the inner region

In this section, we examine in more detail the manner in which the vertical structure of the radar echoes within rainbands in the inner region of the storm varies, first with respect to radius, then with respect to storm quadrant.

a. Variation of rainbands with distance from storm center

In Fig. 3b, a unique feature of the CFAD for the inner rainbands is that the center of the modal distribution is the highest reflectivity of any of these three broader zones, a distinction the eyewall would hold if the storm intensities in the sample were less variable (Hence and Houze 2011). The modal distribution is centered at about 31 dBZ and ranges from about 22 to 36 dBZ. This region also exhibits the most well-defined brightband of the three regions, which is why the schematic in Fig. 1 shows a great deal of stratiform coverage in the inner region, consistent with the low convective fraction in Fig. 4b. The steep and tightly packed decrease in reflectivity with height in the upper portion of the melting layer in Fig. 3b combined with robust stratiform radar echo is a distinctive characteristic of the inner rainbands. This combination of factors likely results from the rainbands in the inner region (i) containing previously active convective cells and (ii) being close to the eyewall, from which ice particles aloft ejected radially outward seed the rainbands below.

The combination of stratiform uniformity and convective activity within rainbands in the inner regions varies with distance from the storm center. In region 3, the echo is mostly stratiform, as indicated by the low convective fraction (Fig. 4b). Figure 5 shows further details of how the rainbands in the inner region vary by distance from the storm center. Vertical profiles of the mean, standard deviation, skewness, and excess kurtosis for the CFADs of each region as a function of height show the increase in convective nature of the rainbands with increasing radius. The red curves show that region 3 has the largest mean reflectivity at low levels, indicating that its widespread stratiform rain is robust. Yet region 3 also has the smallest mean reflectivity in upper levels and lowest standard deviation, consistent with a lack of convective activity. The red curves also show that region 3 has the most negatively skewed reflectivity distribution at all levels, consistent with the generally robust stratiform precipitation dominating in this region (solid red line, Fig. 5c). Region 3 has the least negative kurtosis at low levels and the least positive kurtosis in upper levels (solid red line, Fig. 5d). A positive excess kurtosis is indicative of peakedness in the distribution and/or extended tails of the distribution; a negative excess kurtosis, in contrast, indicates flatness in the distribution and/or short tails (DeCarlo 1997).

Fig. 5.
Fig. 5.

(a) Mean reflectivity as a function of height for CFADs of all overpasses of regions 3–9 (CAT12345; see Table 1). The inner rainbands are red and the distant rainbands are blue for reference. (b)–(d) As in (a), but for standard deviation, skewness, and kurtosis, respectively. The mean reflectivity is the conditional reflectivity (i.e., it is the mean reflectivity for pixels at which a detectable reflectivity exists). Zero reflectivities do not enter the mean.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

Region 3’s combination of an intense low-level mean, low standard deviation, negative skew, and less negative kurtosis together indicate that of all the regions it has the greatest volumetric (area-integrated) rain rate and echo uniformity at low levels, especially in comparison to the rainbands in the outer region of the storm. With increasing distance from the storm center, the low-level distributions weaken, become more variable, and broaden. Emphasizing the two-layer structure of the rainbands, the upper levels exhibit the reverse of this behavior; the fact of region 3’s distributions being the weakest and least variable of the inner regions but having the least positive kurtosis suggests a slightly flat but narrow peak at low reflectivities with lighter tails. The upper levels strengthen, become more variable, and become more peaked with fatter tails in the distribution (solid red line, Fig. 5). The curves of kurtosis in Fig. 5d all show a sharp minimum at the melting level, distinctly showing the break and reversal of statistics between the lower rain layer and the upper ice layer. These characteristics are consistent with the proportion of active convective precipitation increasing with distance from the storm center, as shown schematically in Fig. 1.

To test the robustness of these results, we subsampled the overpasses’ region 3 data into 1000 randomly chosen sets of 50 overpasses and compared them with 1000 randomly chosen 50-overpass sets of region 9 data. The mean and standard deviation profiles of region 3 are consistently more intense and less varied than the envelope containing all 1000 means and standard deviation profiles of the subsampled region 9 groups in the manner discussed above (Figs. 6a,b). The upper levels of region 3 are less intense and varied than all of the 1000 region 9 subsamples. We executed similar tests for the other five regions (not shown), and they have intermediate profiles whose characteristics gradually transition from those of the region 3 profile to the region 9 profile.

Fig. 6.
Fig. 6.

(a) Mean reflectivity as a function of height of 1000 randomly chosen subsamples of the total region 3 (red lines) and region 9 (blue lines) data. (b) As in (a), but for standard deviation.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

The behavior at the upper levels of the rainbands in the inner region of the storm may be related to both the radial and azimuthal circulations and associated redistribution of ice particles occurring at these levels (Marks and Houze 1987), which lessens with distance from the storm center. Black and Hallett (1986) noted that the entire inner core of a tropical cyclone is highly glaciated with ice from the eyewall, which, combined with the azimuthal particle transfer, leads to the lack of electrification of the inner rainbands (Black and Hallett 1999; Molinari et al. 1999). The narrowing of the modal distribution at 5 km (Fig. 3b), corresponding to the sudden dip in kurtosis at the same level (red lines, Fig. 5d), clearly indicates disconnection of the particle regimes between the upper and lower levels of the inner rainbands. We suggest that this behavior occurs because the source of the particles is remote from their fallout region. This tendency is present in all of the rainband regions, but is most extreme in region 3 and lessens with distance.

All of these results suggest that the inner rainbands consist of two relatively independent layers. Convective activity producing precipitation in the inner bands is strongly constrained to occur at lower levels, possibly by the locally enhanced vertical shear from the layer of outflow from the eyewall (as suggested by the inner-rainband radar echoes being located below the eyewall outflow in Fig. 2). At the same time, the levels between 6 and 8 km are filled with slowly falling ice particles. These particles ejected from the eyewall zone swirl around the storm with the azimuthal wind and seed the lower-level clouds. The innermost rainbands are close enough to receive precipitating ice particles from the eyewall circulation; thus, it is likely that the intensity and uniformity of the inner rainbands in intense storms, especially in the upper levels, is partially a result of ice fallout from the eyewall outflow. Since the less dense particles can travel significant distances (Marks and Houze 1987; Hence and Houze 2012), they would be spread rather evenly about the storm. The constraint on the depth of the inner rainband convection appears to gradually relax with distance from the storm center.

The fact that the stratiform echo is of a combination of two origins (dying cells and outflow of ice from other regions of the storm) has dynamic implications. It is within the inner region’s rainbands, specifically region 4 and inward (average radius of 120 km from storm center in this study) that these dynamic implications of the rainbands for the primary storm vortex would be most strongly felt (Judt and Chen 2010). Several studies have shown that the convection within the principal band is organized and persistent enough to alter the distribution of vorticity around the storm (Powell 1990a,b; Samsury and Zipser 1995; Hence and Houze 2008; Didlake and Houze 2009). Particular care must be used, however, when associating the vast amount of stratiform-like radar echo in the inner rainbands with the vertical potential vorticity generation thought to be associated with stratiform precipitation formed from dying convection. Some portions of the inner rainband stratiform precipitation are indeed a result of dying convection and thus have such an associated locally enhanced mass transport profile; however, in other portions of the inner region this profile may be absent because the stratiform-like echo is the result of significantly displaced ice from either the eyewall or other convectively active zones rather than collapsing convection (Hence and Houze 2008). Thus not all “stratiform” regions within the rainbands may be strong generators of potential vorticity. This cautionary note has the further implication that highly convective rainbands may have a more dynamic impact than rainbands that are not convectively active, not only in terms of the convective contributions (Hence and Houze 2008; Didlake and Houze 2009) but also for the convectively generated stratiform precipitation that this convection produces (May and Holland 1999; Franklin et al. 2006). These ideas require testing with further Doppler radar and/or model analyses that are beyond the scope of this study.

b. Variations among shear-relative quadrants

Several studies have documented the impact of vertical wind shear on tropical cyclone rainfall and eyewall structure (Black et al. 2002; Rogers et al. 2003; Chen et al. 2006; Braun et al. 2006; Hence and Houze 2011, 2012). Few studies have investigated how rainbands vary relative to the environmental shear. Figure 1 illustrates schematically how we have found rainbands structures to vary from one quadrant of the storm to another in relation to the mean environmental shear.

The precipitation coverage shown in Fig. 1 is consistent with the statistics in Fig. 7a, which show that the DL has the greatest overall precipitation coverage, followed by the UL, DR, and UR quadrants respectively, indicating a left-of-shear/right-of-shear rainfall asymmetry. However, the precipitation tends to fall out far away from the active convection that creates it. In the annulus of region 3, the DL has only about 10% coverage by convectively classified pixels, and the amount of convection in this quadrant remains fairly flat with distance (purple line, Fig. 7b). In the DR, however, the convective activity increases with distance from the storm center and is the most convective quadrant out to region 5 and beyond (black line, Fig. 7b). The UR has very little convection in regions 3 and 4, but the fraction jumps upward in region 5 to become the second most convectively active quadrant (green line, Fig. 7b). The UL has very few convective cells throughout the inner rainband region (pink line, Fig. 7b). These results suggest that rainband convection in the inner region is favored in the downshear quadrants, but this tendency changes at region 5, where the right-of-shear rainbands become convectively favored. These results are consistent with the spiral nature of the rainbands, as drawn in Fig. 1. Also consistent with this depiction, Corbosiero and Molinari (2002) found a slight preference for lightning (i.e., active convection) in the DL of the innermost rainbands (<100 km, corresponding to region 3 in this study), and a DR preference for the rest of the rainbands.

Fig. 7.
Fig. 7.

(a) Percentage of areal coverage of precipitating pixels as a function of height for all overpasses of the inner- and distant-rainband quadrants (CAT12345; see Table 1). The quadrants are aligned along the average environmental wind shear vector of the sample. The inner rainbands are red and the distant rainbands are blue for reference. (b) The quadrant fraction of convectively classified precipitating pixels as a function of distance. The quadrants are aligned along the average environmental wind shear vector of the sample.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

Figure 8 shows the CFADs for rainbands in the inner region by quadrant. A sharp contrast is seen between the UR and DL quadrants, with the other two quadrants showing intermediate characteristics. The UR has a relatively weak and broad distribution, with the low-level modal distribution centered below 30 dBZ and ranging from 19 to 36 dBZ. The brightband is distinct, but the melting layer signature covers a wide range of reflectivity in the CFAD, and the upper levels have a sharp peak at 6 km and 21 dBZ. The modal distribution reaches 8 km, and the outliers reach 10 km. The well-defined brightband signature in the UR CFAD indicates the presence of stratiform precipitation, while the heterogeneity of the low levels indicates convective cells. From the schematic in Fig. 1, this distribution is shown as a likely combination of new convective cells in this upstream zone, stratiform precipitation portions associated with the edges of principal rainbands located farther out, and ice particles at upper levels emanating from the from the eyewall region and seeding the clouds below.

Fig. 8.
Fig. 8.

Region 3–5 inner rainband CFADs of TRMM PR reflectivity data for all of the 1998–2007 overpasses of storms that ultimately reached category 4 or 5 intensity (CAT12345 grouping; Table 1), arranged by quadrants oriented to the average shear vector. Quadrants are labeled downshear left (DL), downshear right (DR), upshear left (UL), and upshear right (UR). All other details are as in Fig. 3.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

In the DL region CFAD in Fig. 8, the low-level modal distribution is much narrower and stronger than that of the UR, with the modal distribution ranging across only 12 dB and centered at about 33 dBZ. Below the melting layer, low reflectivity values (light rain rates) are very rare in the DL, whereas they are common in the UR distribution. The dominant stratiform rain in the DL is thus uniformly stronger than in other quadrants. The melting layer distribution is also more tightly packed, further indicating a robust stratiform echo population that produces a relatively large volumetric rain rate in this sector. At the same time, the embedded convection (mostly in the upstream portion of the DL region; see Fig. 1) is indicated by the outlier values to also be more intense than outlier convection in the UR region and other quadrants. The outlier reflectivities at low levels reach up to 45 dBZ, exceeding those in other quadrants. The overall upper-level peak frequency remains at about the same height in the DL as in other quadrants, but the outliers reach slightly higher heights, again indicating that intermittent convection is more intense in this region.

Figure 9 shows that the DL has the greatest low-level mean reflectivity, while the UR has the weakest mean reflectivity and is about 2 dB less than in the DL (Fig. 9a). The UL and DR are intermediate and similar. The DL and DR have the greatest mean upper-level reflectivity, followed by the UR and UL. The low-level distribution is the flattest and least tailed in the UR, becoming more peaked in the DR, UL, and DL. The DR likely has more extremes in its tails (Figs. 9b,c). In the upper levels, the DL is the broadest, and the UR the most peaked but likely with extreme tails given the similar variability (Figs. 9b,c). The UL is peaked but with low variability, signaling a weak and highly uniform distribution. The kurtosis profiles again emphasize the break between the lower-level rain mode and the upper-level ice modes, and the reversal of the statistics across the melting level. The extreme of this reversal is between the UR and DL quadrants.

Fig. 9.
Fig. 9.

(a) Mean reflectivity as a function of height of the inner-rainband-quadrant CFADs from Fig. 8 (CAT12345; Table 1). The mean reflectivity is the conditional reflectivity (i.e., it is the mean reflectivity for pixels at which a detectable reflectivity exists). Zero reflectivities do not enter the mean. (b) As in (a), but for standard deviation. (c) As in (a), but for kurtosis. Quadrants are labeled as in Fig. 8.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

These statistics are all consistent with inner-region convection being most numerous in the DR quadrant (hence, the numerous embedded convective cells in the DR in Fig. 1). Many ice particles from this convection travel downwind to fall out as primarily stratiform rain in the DL (the dominant stratiform echo in the DL of Fig. 1), in combination with ice particles extruded from the eyewall seeding the rainbands. Some of the smaller particles generated in the DR probably continue to the UL. The greater low-level variability of the DR as well as the intensification, variation, and raising of the upper-level echo from the DR to the DL followed by the lowering, weakening, and homogenizing of the UL upper-level reflectivities suggest a maturation and dissipation of rainband convection through these three quadrants (Fig. 8).

The UR has a distribution wholly unlike the other three quadrants. Since this quadrant is as sparse as a far-distant band (red dotted line, Fig. 7a), has a broad distribution unlike the uniform distribution of the other three quadrants (Fig. 8), and has a sudden jump in the convectively classified fraction in region 5 (green line, Fig. 7a), it is likely that this quadrant’s distribution contains a mixture of broken precipitation and a small portion of the principal band.

6. Rainbands in the outer region

We now investigate how the rainbands in the outer region, which are less affected by the vortex dynamics of the cyclone, differ from the rainbands in the inner region of the storm.

a. General features

In the principal band, convective cells are typically mature or dying by the time they reach the inner core (Barnes et al. 1983; Hence and Houze 2008). These cells often begin their lives in the upwind portions of the rainbands, which is in the distant environment, where the vortex is thought not to exert as much influence (Houze 2010). These convective cells are shown in the bulk rainband schematic of Fig. 1 in two groups—one entirely in the outer region, the other as a band crossing the boundary between the outer and inner regions, such that the newer cells are in the outer region and older cells in the inner region.

Figure 3c is the CFAD for outer rainband region (regions 6–9). The low-level distribution is centered at about 28 dBZ, but the modal distribution ranges from 19 to 36 dBZ; that is, compared with the inner region rainbands, there are fewer intermediate intensity echoes and more low intensities. This comparison indicated that the rainbands in the outer region are more heterogeneous, or more cellular in nature. The upper-level distribution has a relatively broad peak at 5.5–7 km. The statistics show that the mean low-level reflectivity of the rainbands in the outer region is lower than that of the inner-region rainbands, but the upper-level reflectivity is higher, bending toward greater intensities (blue lines, Fig. 5a). This difference is consistent with the distant rainbands having deep convective cores. The variability in reflectivity is greater at all levels and bends toward higher values above 7 km (blue lines, Fig. 5b), and the upper levels are less positively skewed toward intense reflectivities (blue lines, Fig. 5c). The rainbands in the outer region have the most negative low-level kurtosis profile, and although their kurtosis profile becomes positive above the melting level they decrease with height above about 8 km (blue lines, Fig. 5d).

The bump in convectively classified fraction in regions 5 and 6 (black line, Fig. 4b), as well as the associated changes in the statistics (Fig. 5), suggests that regions 5 and 6 mark the common location of the convective portion of the principal band. This result guided our decision to draw the boundary between the inner and outer regions of the storm at the dotted circle in Fig. 1. Outside of this border region, the statistics show the makeup of the rainbands to be sparse, convective, and widely varied. At radial distances outward of about 240 km, it is unlikely that the distant rainbands are being seeded by many precipitation-sized ice particles from the eyewall. Therefore, the upper-level distribution of the distant rainbands must be mostly locally generated, either within the principal rainband or the individual convective cells outside of it. The higher intensity and variability in the upper-level profiles suggest that a variety of particle sizes and/or amounts exist within the distant rainbands, instead of having the strong uniformity of the inner bands (Figs. 5a,b). These statistical differences all point to the rainbands in the outer regions being less contiguous, generally more convective, and less stratiform than in the inner regions of storms, consistent with Fig. 1.

Distant rainbands often generate tornadoes upon landfall since the vertical wind shear in the rainbands provides an ideal environment for both supercells and tornadoes to form (Schultz and Cecil 2009). Distant rainbands are also more electrically active than any other region of the tropical cyclone (Black and Hallett 1999; Molinari et al. 1999). Therefore, while the rainbands in the outer region may not directly contribute to the storm dynamically, they extend the impact of the storm well away from its inner core.

b. Variations among shear-relative quadrants

Figure 7b shows that outward of region 5, the right-of-shear quadrants have the greatest amount of active convection. In the outermost regions, the UR has the most convective activity, followed by the DR, UL, and DL, respectively. Conversely, the DL is the quadrant with the highest stratiform content. This result is further supported by the fact that the DL has the greatest overall coverage of the outer regions, followed by the DR, UL, and UR, respectively (Fig. 7b). These azimuthal variations are consistent with the schematic in Fig. 1.

Figure 10 contains the outer-region CFADs for all quadrants. They show that in the UL, the distribution is flat, weak, and highly variable at all levels, with a broad peak centered at about 25 dBZ. The UR modal distribution tightens somewhat around these low reflectivity values, but the distribution gains some intensity and height in the outliers. The DR’s modal distribution is more intense, centered at about 27–28 dBZ, but remains relatively flat. As in the inner regions, the greatest difference in CFADs is between the UR and DL regions. In the DL region, low reflectivities are rare and the distribution is the most intense, peaked, and uniform, with a center near 31 dBZ. These general characteristics are consistent with the way that we have drawn Fig. 1, which shows widespread stratiform precipitation in DL with a few intense embedded cells on its upwind side. The DL is the only quadrant in which the CFAD pinches in the melting level to clearly separate the rain and ice modes of the CFAD; the other quadrant distributions are more continuous through the melting layer.

Fig. 10.
Fig. 10.

Region 6–9 outer rainband CFADs of TRMM PR reflectivity data for all of the 1998–2007 overpasses of storms that ultimately reached category 4 or 5 intensity (CAT12345 grouping; Table 1), arranged by quadrants oriented to the average shear vector. Quadrants are labeled as in Fig. 8. All other details are as in Fig. 3.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

At low levels, the statistics differ most between the downshear and upshear quadrants; the downshear quadrants are more intense (blue and purple lines, Fig. 11a), with the DR being the most variable and the DL the most peaked and narrow (blue and purple lines, Figs. 11b,c). These differences are consistent with Fig. 1, in which the DR and DL show extensive intense echo, but with the DR containing more embedded convective cells and the DL being predominantly stratiform.

Fig. 11.
Fig. 11.

(a) Mean reflectivity as a function of height of the outer-rainband-quadrant CFADs from Fig. 10 (CAT12345; see Table 1). (b) As in (a), but for standard deviation. (c) As in (a), but for kurtosis. Quadrants are labeled as in Fig. 8.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

A small dip in the melting-layer kurtosis occurs in all of the quadrants, but the extent of the dip is most extreme in the DL (Fig. 11c), indicating a more profound separation of lower- and upper-level regimes. Above the dip, the upper-level profiles differ most between the right-of-shear and left-of-shear quadrants, with the right-of-shear side being the most intense and most variable (pink and green lines, Figs. 11a,b)—that is, more convective. The upper-level kurtosis is the most positive in the DL, followed by the DR, UL, and UR respectively (Fig. 11c).

These several factors suggest that in the outer region of the storm the UR quadrant is the preferred genesis zone for organized rainband convection: the quadrant’s sparseness of echoes (dotted blue line, Fig. 7a); its high convectively classified fraction (green line, Fig. 7b); the weakness of its low-level distribution (Fig. 10); and the intensity, variability, and flatness of its upper-level distribution (green lines, Fig. 11). This convection likely grows and matures in the DR, as evidenced by the continued upper-level intensity and variability, the increased low-level intensity and variability, and the increased kurtosis of the upper-level distribution, suggesting a more narrowly peaked distribution with extreme tails. The DL has a large amount of stratiform echo, likely from collapsed convection and possibly combined with small ice particles blowing outward from the inner region of the storm. These traits are evidenced by the DL’s high mean reflectivity but low variability; furthermore, the DL exhibits the least negative/most positive kurtosis, signifying a more narrowly peaked distribution, especially in the upper levels. The UL falls in the middle of these profiles, being relatively weak and highly variable but relatively well covered. These traits suggest that there is a mix of stratiform and convective precipitation in the UL, possibly from the continued advection of ice particles combined with other less-organized convection (Fig. 7).

7. Modifications to the vertical structure of rainbands by internal storm dynamics and ambient environmental conditions

a. Storm intensity effects on the inner and outer regions of the storm

We might expect the impact of storm intensity on rainband structures to differ between the inner and outer regions, since the above observations suggest that in the inner regions of storms the eyewall outflow vertically constrains the rainbands and seeds them with a deep layer of precipitating ice particles. Now that we have determined the general characteristics of the rainbands, we investigate how differing internal storm dynamics (as indicated by storm intensity category) might affect this typical structure. The black line in Fig. 4b shows that the average convectively classified fraction increases almost linearly with distance, but with an upward bump around region 6. Storm intensity changes the shape of this curve. Category 1 and 2 storms (CAT12, Table 1; dark purple line, Fig. 4b) are more convective than average inside of region 6 but less outward. In contrast, category 4 and 5 storms (CAT45; Table 1; light purple line, Fig. 4b) are convectively suppressed in regions 3–5, but from region 6 outward the rainbands become steadily more convective, reaching the average around regions 8 and 9.

Figures 12a and 12b show that for CAT12 storms, rainbands in the inner and outer regions are statistically similar except in the uppermost levels, where the inner-rainband distribution is flatter, less skewed, and less variable than the distant rainbands (blue lines, Figs. 12a,b; skewness and standard deviation not shown). The CAT45 rainbands, however, behave differently in the inner and outer regions. The lower-level echoes are more intense in the rainbands of the inner regions, but there is a stronger convective suppression in their upper levels (pink lines, Fig. 12a). The CAT45 rainbands also tend to have a much more peaked and uniform distribution in the inner region than in the outer region (pink lines, Fig. 12b; standard deviation not shown). These observations are consistent with Willoughby et al. (1984), who noted that the stationary band complex tended to be closer to the storm center in weaker storms and farther away in stronger storms. Thus these results indicate that storm intensity has the greatest influence upon the radial variations in rainband convection.

Fig. 12.
Fig. 12.

(a) Mean reflectivity as a function of height of the total region CFADs for the inner- and distant-rainband regions of the CAT12 and CAT45 groupings. (b) Kurtosis as a function of height of the CFADs for the inner- and distant-rainband regions of the CAT12 and CAT45 groupings. (c) As in (a), but for marginal- and high-SST groupings. (d) As in (b), but for the marginal- and high-SST groupings. See Table 1 for overpass information of groupings.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

b. How SST affects rainbands in the inner and outer regions of the storm

SST is the only parameter that influences rainband reflectivity features across the storm uniformly. Changes in SST shift the curve in Fig. 4b as a unit to higher convective fraction—high SSTs (hiSST; >28°C; dark green line) shift the curves upward, indicating that more convection occurs in all radial regions when the SSTs are warm. Less convection occurs overall in marginal SST conditions (margSST; 26° ≤ SST ≤ 28°C; light green line), although the bump at region 6 becomes more pronounced, suggesting that the convective forcing in that region may be more of a factor than in the other regions. The mean low-level reflectivity also shows a change in convective intensity, with both the inner- and distant-band mean reflectivities shifting to greater values in high SST cases (pink lines, Fig. 12c). The melting layer signature also shifts to higher altitudes (pink lines, Figs. 12c,d).

These results are consistent with the idea that warmer water is better able to support the rainband convection overall, but based on these results, the buoyancy needed to reach upper levels is better utilized in the rainbands far from the storm center. However, these results also suggest that the principal rainband convection is minimally impacted by changes in SST in the range tested, suggesting that this convection is forced dynamically. Thus, although SST seems to modulate the existing rainband signature, within the temperature range investigated, SST does not fundamentally change the overall rainband structure.

c. How shear affects the quadrant-by-quadrant distribution of inner region rainbands

When the environmental shear increases, the rainbands in the inner region of the storm are affected. Figure 13 shows that in cases of low shear, the differences between the quadrants are fairly small, but as the shear increases, the intensity variation among quadrants increases. The differences between the mean low-level reflectivity profiles become exaggerated, with the rainbands in the DL region becoming stronger and those in the UR becoming weaker, especially at low levels (Figs. 13a,c). The excess kurtosis (Figs. 13b,d) becomes less negative in all quadrants, reaching nearly zero in the DL. This change implies that as the shear increases, the distribution becomes more sharply peaked, indicating a more robust stratiform rain regime. At upper levels, the kurtosis in the DL becomes less positive, indicating a less-peaked distribution, suggesting that the embedded convection becomes stronger. Thus, the increased shear makes both the stratiform and convective components of the rainbands in the inner DL region stronger, resulting in the rainbands being strongly asymmetric.

Fig. 13.
Fig. 13.

(a) Mean reflectivity as a function of height of the quadrant CFADs for the inner-rainband regions of the low-shear grouping. (b) Kurtosis as a function of height of the quadrant CFADs for the inner-rainband regions of the low-shear grouping. (c) As in (a), but for the high-shear grouping. (d) As in (b), but for the high-shear grouping. See Table 1 for overpass information of groupings.

Citation: Journal of the Atmospheric Sciences 69, 9; 10.1175/JAS-D-11-0323.1

These observations suggest that shear likely determines the azimuthal placement of the rainband convection and the resulting precipitation. Highly sheared storms tend to have a strong rainband asymmetry, suggesting that only one principal band forms with its convection favoring the downshear side. The evenness in the convection of weakly sheared storms may signal a tighter wrapping of the spiral bands, a greater tendency for convective secondary bands, or a tendency for these storms to have multiple principal bands.

8. Conclusions

A statistical analysis of 10 yr of TRMM PR three-dimensional reflectivity data reveals variations in the vertical structures of tropical cyclone rainbands. Variations in vertical structure are indicated by CFADs of the reflectivity in different portions of the rainbands and across different portions of the storm. Using this approach, we have been able to discern rainband structure variations as a function of both radius and azimuth relative to the storm. Radial variations are controlled primarily by the intensity of the tropical cyclone vortex. The azimuthal variations are controlled primarily by the shear in the large-scale environment of the storm.

CFAD analysis shows that the vertical structure of rainbands is two-layered. The statistical characteristics of the rainbands are all characterized by a separation, or pinching off, of the frequency distribution of reflectivity at the melting level. The CFADs show distinctly separate modes of reflectivity above and below the melting level, which indicate different degrees and types of convective and stratiform structure within the rainbands. Another distinct qualitative difference in rainband structures occurs at an average distance of about 200 km from the storm center, marking the likely boundary of the dynamical influence of the storm’s vortex dynamics upon rainband convection.

Rainbands inward of this boundary consist of a mixture of vertically limited convection and robust stratiform precipitation. The vertical compression of the maximum height of echoes is expected because the outflow layer of the eyewall and associated shearing effects dominate the clouds in the upper levels of the inner region of the storm. This outflow has both a dynamic effect in limiting the height of the convection in the inner region and a microphysical effect via seeding of the rainbands with ice generated in the eyewall and upwind portions of rainbands. The CFADs drawn from rainbands in the inner region show a rather uniformly distributed layer of ice above the melting layer that is distinctly pinched off from the intense and highly uniform low-level precipitation. The ice-layer statistics suggest that the robust stratiform precipitation seen in the downwind portions of rainbands in the inner region of the storm is likely a strong combination of precipitation seeded from other parts of the storm as well as stratiform precipitation resulting from the collapse of convective cells traveling along the rainbands. In contrast, the reflectivity statistics indicate that rainbands in the outer region of the cyclone are more discretely cellular and convective in nature. The CFADs of rainbands of the outer region show a less-sharp separation at the melting level, suggesting that they mostly contain precipitation that is more locally connected to its source, vertically and horizontally, instead of containing particles at upper levels that have been significantly displaced from their sources.

The rainbands generally become more convective with distance from the storm center, but not in an azimuthally uniform manner. Rainband convection in the outer regions typically forms in the upshear quadrants before organizing into lines on the right-of-shear side. As the rainbands spiral inwards toward the storm center and cross into the inner regions, the convective cells collapse on the downshear side, leaving most of the resulting stratiform precipitation on the left-of-shear side. As the environmental shear increases, these asymmetries in the rainband ensemble intensify, pushing more of the rainband convection and precipitation downshear.

The vertical structure of rainbands in the inner regions of the storm changes with storm intensity. In the inner region, the differences in the distributions of reflectivity in the upper and lower portions of the rainband echoes become more exaggerated as the intensity category increases. By contrast, in the outer region, the vertical structures of rainbands do not change much with storm intensity. Changes in storm intensity mainly impact the convection in the inner region, likely because the vertical constraint upon convection in the inner region is in the zone affected by the vortex dynamics. Cool SSTs uniformly discourage rainband convection around the storm, although principal rainband convection is minimally affected.

Doppler analysis and/or numerical modeling of numerous tropical cyclones will be required to statistically investigate the kinematic, dynamic, and microphysical ideas outlined here. Possible future projects include expanding this analysis to the remaining tropical cyclone basins and using this technique to examine the features of tropical cyclone precipitation generated in numerical models. This technique could also potentially be used to examine structure changes in tropical cyclones undergoing rapid intensification.

Acknowledgments

We are grateful to Anthony Didlake for editing and comments, Stacy Brodzik for database and software support, and Beth Tully for editing and graphics. Tyler Burns and Tia Lerud assisted with case identification and database creation. NCEP reanalysis data are provided by the NOAA/OAR/ESRL PSD (at http://www.esrl.noaa.gov/psd/). This research was sponsored by NSF RAINEX Grant ATM-0743180, NASA PMM Grant NNX10AH70G, NASA Earth and Space Science Fellowship NNX06AF17H, and a NASA Space Grant Fellowship.

REFERENCES

  • Atlas, D., , K. R. Hardy, , R. Wexler, , and R. Boucher, 1963: On the origin of hurricane spiral bands. Geofis. Int., 3, 123132.

  • Awaka, J., , T. Iguchi, , and K. Okamoto, 2009: TRMM PR standard algorithm 2A23 and its performance on bright band detection. J. Meteor. Soc. Japan, 87A, 3152.

    • Search Google Scholar
    • Export Citation
  • Barnes, G. M., , E. J. Zipser, , D. Jorgensen, , and F. Marks Jr., 1983: Mesoscale and convective structure of a hurricane rainband. J. Atmos. Sci., 40, 21252137.

    • Search Google Scholar
    • Export Citation
  • Black, M. L., , J. F. Gamache, , F. D. Marks, , C. E. Samsury, , and H. E. Willoughby, 2002: Eastern Pacific Hurricanes Jimena of 1991 and Olivia of 1994: The effect of vertical shear on structure and intensity. Mon. Wea. Rev., 130, 22912312.

    • Search Google Scholar
    • Export Citation
  • Black, R. A., , and J. Hallett, 1986: Observations of the distribution of ice in hurricanes. J. Atmos. Sci., 43, 802822.

  • Black, R. A., , and J. Hallett, 1999: Electrification of the hurricane. J. Atmos. Sci., 56, 20042028.

  • Braun, S. A., , M. T. Montgomery, , and Z. Pu, 2006: High-resolution simulation of Hurricane Bonnie (1998). Part I: The organization of eyewall vertical motion. J. Atmos. Sci., 63, 1942.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., , E. J. Zipser, , and S. W. Nesbitt, 2002: Reflectivity, ice scattering, and lightning characteristics of hurricane eyewalls and rainbands. Part I: Quantitative description. Mon. Wea. Rev., 130, 769784.

    • Search Google Scholar
    • Export Citation
  • Chen, S. S., , J. A. Knaff, , and F. D. Marks, 2006: Effects of vertical wind shear and storm motion on tropical cyclone rainfall asymmetries deduced from TRMM. Mon. Wea. Rev., 134, 31903208.

    • Search Google Scholar
    • Export Citation
  • Chen, Y., , and M. K. Yau, 2001: Spiral bands in a simulated hurricane. Part I: Vortex Rossby wave verification. J. Atmos. Sci., 58, 21282145.

    • Search Google Scholar
    • Export Citation
  • Corbet, J., , C. Mueller, , C. Burghart, , K. Gould, , and G. Granger, 1994: Zeb: Software for geophysical data integration, display, and management of diverse environmental datasets. Bull. Amer. Meteor. Soc., 75, 783792.

    • Search Google Scholar
    • Export Citation
  • Corbosiero, K. L., , and J. Molinari, 2002: The effects of vertical wind shear on the distribution of convection in tropical cyclones. Mon. Wea. Rev., 130, 21102123.

    • Search Google Scholar
    • Export Citation
  • Corbosiero, K. L., , J. Molinari, , A. R. Aiyyer, , and M. L. Black, 2006: The structure and evolution of Hurricane Elena (1985). Part II: Convective asymmetries and evidence for vortex Rossby waves. Mon. Wea. Rev., 134, 30733091.

    • Search Google Scholar
    • Export Citation
  • DeCarlo, L. T., 1997: On the meaning and use of kurtosis. Psychol. Methods, 2, 292307.

  • Didlake, A. C. Jr., , and R. A. Houze Jr., 2009: Convective-scale downdrafts in the principal rainband of Hurricane Katrina (2005). Mon. Wea. Rev., 137, 32693293.

    • Search Google Scholar
    • Export Citation
  • Franklin, C. N., , G. J. Holland, , and P. T. May, 2006: Mechanisms for the generation of mesoscale vorticity features in tropical cyclone rainbands. Mon. Wea. Rev., 134, 26492669.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., , and R. A. Houze Jr., 2008: Kinematic structure of convective-scale elements in the rainbands of Hurricanes Katrina and Rita (2005). J. Geophys. Res., 113, D15108, doi:10.1029/2007JD009429.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., , and R. A. Houze Jr., 2011: Vertical structure of hurricane eyewalls as seen by the TRMM Precipitation Radar. J. Atmos. Sci., 68, 16371652.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., , and R. A. Houze Jr., 2012: Vertical structure of tropical cyclones with concentric eyewalls as seen by the TRMM Precipitation Radar. J. Atmos. Sci., 69, 10211036.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., 1997: Stratiform precipitation in regions of convection: A meteorological paradox? Bull. Amer. Meteor. Soc., 78, 21792196.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., 2004: Mesoscale convective systems. Rev. Geophys., 42, RG4003, doi:10.1029/2004RG000150.

  • Houze, R. A., Jr., 2010: Clouds in tropical cyclones. Mon. Wea. Rev., 138, 293344.

  • Houze, R. A., Jr., , D. C. Wilton, , and B. F. Smull, 2007: Monsoon convection in the Himalayan region as seen by the TRMM Precipitation Radar. Quart. J. Roy. Meteor. Soc., 133, 13891411.

    • Search Google Scholar
    • Export Citation
  • James, C. N., , S. R. Brodzik, , H. Edmon, , R. A. Houze Jr., , and S. E. Yuter, 2000: Radar data processing and visualization over complex terrain. Wea. Forecasting, 15, 327338.

    • Search Google Scholar
    • Export Citation
  • Jordan, C. L., 1958: Mean soundings for the West Indies area. J. Meteor., 15, 9197.

  • Judt, F., , and S. S. Chen, 2010: Convectively generated potential vorticity in rainbands and formation of the secondary eyewall in Hurricane Rita of 2005. J. Atmos. Sci., 67, 35813599.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471.

  • Knapp, K. R., , M. C. Kruk, , D. H. Levinson, , H. J. Diamond, , and C. J. Neumann, 2010: The International Best Track Archive for Climate Stewardship (IBTrACS). Bull. Amer. Meteor. Soc., 91, 363376.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., , W. Barnes, , T. Kozu, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15, 809817.

    • Search Google Scholar
    • Export Citation
  • Marks, F. D., , and R. A. Houze Jr., 1987: Inner core structure of Hurricane Alicia from airborne Doppler radar observations. J. Atmos. Sci., 44, 12961317.

    • Search Google Scholar
    • Export Citation
  • May, P. T., , and G. J. Holland, 1999: The role of potential vorticity generation in tropical cyclone rainbands. J. Atmos. Sci., 56, 12241228.

    • Search Google Scholar
    • Export Citation
  • Molinari, J., , P. Moore, , and V. Idone, 1999: Convective structure of hurricanes as revealed by lightning locations. Mon. Wea. Rev., 127, 520534.

    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., , and R. J. Kallenbach, 1997: A theory for vortex Rossby-waves and its application to spiral bands and intensity changes in hurricanes. Quart. J. Roy. Meteor. Soc., 123, 435465.

    • Search Google Scholar
    • Export Citation
  • Powell, M. D., 1990a: Boundary layer structure and dynamics in outer hurricane rainbands. Part I: Mesoscale rainfall and kinematic structure. Mon. Wea. Rev., 118, 891917.

    • Search Google Scholar
    • Export Citation
  • Powell, M. D., 1990b: Boundary layer structure and dynamics in outer hurricane rainbands. Part II: Downdraft modification and mixed layer recovery. Mon. Wea. Rev., 118, 918938.

    • Search Google Scholar
    • Export Citation
  • Qiu, X., , Z.-M. Tan, , and Q. Xiao, 2010: The roles of vortex Rossby waves in hurricane secondary eyewall formation. Mon. Wea. Rev., 138, 20922109.

    • Search Google Scholar
    • Export Citation
  • Rogers, R., , S. Chen, , J. Tenerelli, , and H. Willoughby, 2003: A numerical study of the impact of vertical shear on the distribution of rainfall in Hurricane Bonnie (1998). Mon. Wea. Rev., 131, 15771599.

    • Search Google Scholar
    • Export Citation
  • Saffir, H. S., 2003: Communicating damage potentials and minimizing hurricane damage. Hurricane! Coping with Disaster, R. Simpson, Ed., Amer. Geophys. Union, 155–164.

  • Samsury, C. E., , and E. J. Zipser, 1995: Secondary wind maxima in hurricanes: Airflow and relationship to rainbands. Mon. Wea. Rev., 123, 35023517.

    • Search Google Scholar
    • Export Citation
  • Schultz, L. A., , and D. J. Cecil, 2009: Tropical cyclone tornadoes, 1950–2007. Mon. Wea. Rev., 137, 34713484.

  • Terwey, W. D., , and M. T. Montgomery, 2008: Secondary eyewall formation in two idealized, full-physics modeled hurricanes. J. Geophys. Res., 113, D12112, doi:10.1029/2007JD008897.

    • Search Google Scholar
    • Export Citation
  • TSDIS, 2007: File specifications for TRMM products levels 2 and 3. Vol. 4, Interface Control Specification between the Tropical Rainfall Measuring Mission Science Data and Information System (TSDIS) and the TSDIS Science User (TSU). NASA GSFC, TSDIS-P907, 102 pp. [Available online at http://pps.gsfc.nasa.gov/tsdis/Documents/ICSVol4.pdf.]

  • Willoughby, H. E., 1988: The dynamics of the tropical cyclone core. Aust. Meteor. Mag., 36, 183191.

  • Willoughby, H. E., , F. D. Marks, , and R. J. Feinberg, 1984: Stationary and moving convective bands in hurricanes. J. Atmos. Sci., 41, 31893211.

    • Search Google Scholar
    • Export Citation
  • Yuter, S. E., , and R. A. Houze Jr., 1995: Three-dimensional kinematic and microphysical evolution of Florida cumulonimbus. Part II: Frequency distributions of vertical velocity, reflectivity, and differential reflectivity. Mon. Wea. Rev., 123, 19411963.

    • Search Google Scholar
    • Export Citation
Save