• Bowman, K. P., and M. D. Fowler, 2015: The diurnal cycle of precipitation in tropical cyclones. J. Climate, 28, 53255334, doi:10.1175/JCLI-D-14-00804.1.

    • Search Google Scholar
    • Export Citation
  • Bowman, K. P., J. C. Collier, G. R. North, Q. Wu, E. Ha, and J. Hardin, 2005: Diurnal cycle of tropical precipitation in Tropical Rainfall Measuring Mission (TRMM) satellite and ocean buoy rain gauge data. J. Geophys. Res., 110, D21104, doi:10.1029/2005JD005763.

    • Search Google Scholar
    • Export Citation
  • Browner, S. P., W. L. Woodley, and C. G. Griffith, 1977: Diurnal oscillation of the area of cloudiness associated with tropical storms. Mon. Wea. Rev., 105, 856864, doi:10.1175/1520-0493(1977)105<0856:DOOTAO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., 2007: Satellite-derived rain rates in vertically sheared tropical cyclones. Geophys. Res. Lett., 34, L02811, doi:10.1029/2006GL027942.

    • 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, doi:10.1175/1520-0493(2002)130<0769:RISALC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Demuth, J. L., M. DeMaria, and J. A. Knaff, 2006: Improvement of Advanced Microwave Sounding Unit tropical cyclone intensity and size estimation algorithms. J. Appl. Meteor. Climatol., 45, 15731581, doi:10.1175/JAM2429.1.

    • Search Google Scholar
    • Export Citation
  • Dunion, J. P., C. D. Thorncroft, and C. S. Velden, 2014: The tropical cyclone diurnal cycle of mature hurricanes. Mon. Wea. Rev., 142, 39003919, doi:10.1175/MWR-D-13-00191.1.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105, 11711188, doi:10.1175/1520-0493(1977)105<1171:DVODCC>2.0.CO;2.

    • 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, doi:10.1175/2011JAS3578.1.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., and R. A. Houze Jr., 2012: Vertical structure of tropical cyclone rainbands as seen by the TRMM Precipitation Radar. J. Atmos. Sci., 69, 26442661, doi:10.1175/JAS-D-11-0323.1.

    • Search Google Scholar
    • Export Citation
  • Hobgood, J. S., 1986: A possible mechanism for the diurnal oscillations of tropical cyclones. J. Atmos. Sci., 43, 29012922, doi:10.1175/1520-0469(1986)043<2901:APMFTD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Jiang, H., C. Liu, and E. J. Zipser, 2011: A TRMM-based tropical cyclone cloud and precipitation feature database. J. Appl. Meteor. Climatol., 50, 12551274, doi:10.1175/2011JAMC2662.1.

    • Search Google Scholar
    • Export Citation
  • Kossin, J. P., 2002: Daily hurricane variability inferred from GOES infrared imagery. Mon. Wea. Rev., 130, 22602270, doi:10.1175/1520-0493(2002)130<2260:DHVIFG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kozu, T., and Coauthors, 2001: Development of precipitation radar onboard the Tropical Rainfall Measuring Mission (TRMM) satellite. IEEE Trans. Geosci. Remote Sens., 39, 102116, doi:10.1109/36.898669.

    • 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, doi:10.1175/1520-0426(1998)015<0809:TTRMMT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., S. Ringerud, J. Crook, D. Randel, and W. Berg, 2011: An observationally generated a priori database for microwave rainfall retrievals. J. Atmos. Oceanic Technol., 28, 113130, doi:10.1175/2010JTECHA1468.1.

    • Search Google Scholar
    • Export Citation
  • Lajoie, F. A., and I. J. Butterworth, 1984: Oscillation of high-level cirrus and heavy precipitation around Australian region tropical cyclones. Mon. Wea. Rev., 112, 535544, doi:10.1175/1520-0493(1984)112<0535:OOHLCA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Landsea, C. W., and J. L. Franklin, 2013: Atlantic hurricane database uncertainty and presentation of a new database format. Mon. Wea. Rev., 141, 35763592, doi:10.1175/MWR-D-12-00254.1.

    • Search Google Scholar
    • Export Citation
  • Li, Q., and Y. Wang, 2012: Formation and quasi-periodic behavior of outer spiral rainbands in a numerically simulated tropical cyclone. J. Atmos. Sci., 69, 9971020, doi:10.1175/2011JAS3690.1.

    • Search Google Scholar
    • Export Citation
  • Liu, C., E. J. Zipser, D. J. Cecil, S. W. Nesbitt, and S. Sherwood, 2008: A cloud and precipitation feature database from nine years of TRMM observations. J. Appl. Meteor. Climatol., 47, 27122728, doi:10.1175/2008JAMC1890.1.

    • Search Google Scholar
    • Export Citation
  • Muramatsu, T., 1983: Diurnal variations of satellite-measured TBB areal distribution and eye diameter of mature typhoons. J. Meteor. Soc. Japan, 61, 7790.

    • Search Google Scholar
    • Export Citation
  • NASA, 2016: Precipitation Processing System Tropical Rainfall Measuring Mission: File specification for TRMM products, version 7.004. NASA Tech. Doc., 328 pp. [Available online at http://pps.gsfc.nasa.gov/Documents/filespec.TRMM.V7.pdf.]

  • Nesbitt, S. W., and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16, 14561475, doi:10.1175/1520-0442-16.10.1456.

    • Search Google Scholar
    • Export Citation
  • Nesbitt, S. W., E. J. Zipser, and D. J. Cecil, 2000: A census of precipitating features in the tropics using TRMM: Radar, ice scattering, and lightning observations. J. Climate, 13, 40874106, doi:10.1175/1520-0442(2000)013<4087:ACOPFI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Olander, T. L., and C. S. Velden, 2007: The advanced Dvorak technique: Continued development of an objective scheme to estimate tropical cyclone intensity using geostationary infrared satellite imagery. Wea. Forecasting, 22, 287298, doi:10.1175/WAF975.1.

    • Search Google Scholar
    • Export Citation
  • Panofsky, H. A., and G. W. Brier, 1958: Some Applications of Statistics to Meteorology. The Pennsylvania State University, 224 pp.

  • Simpson, J., R. F. Adler, and G. R. North, 1988: A proposed Tropical Rainfall Measuring Mission (TRMM) satellite. Bull. Amer. Meteor. Soc., 69, 278295, doi:10.1175/1520-0477(1988)069<0278:APTRMM>2.0.CO;2.

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

    • Search Google Scholar
    • Export Citation
  • Steranka, J., E. B. Rodgers, and R. C. Gentry, 1984: The diurnal variation of Atlantic Ocean tropical cyclone cloud distribution inferred from geostationary satellite infrared measurements. Mon. Wea. Rev., 112, 23382344, doi:10.1175/1520-0493(1984)112<2338:TDVOAO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wu, Q., Z. Ruan, D. Chen, and T. Lian, 2015: Diurnal variations of tropical cyclone precipitation in the inner and outer rainbands. J. Geophys. Res. Atmos., 120, 111, doi:10.1002/2014JD022190.

    • Search Google Scholar
    • Export Citation
  • Yang, G.-Y., and J. Slingo, 2001: The diurnal cycle in the tropics. Mon. Wea. Rev., 129, 784801, doi:10.1175/1520-0493(2001)129<0784:TDCITT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Composite VIRS IR brightness temperature (BT) for (a) inner radii (<500 km) and (b) outer radii (500–1000 km) as a function of LST for Atlantic tropical cyclones of at least tropical storm strength [i.e., maximum sustained wind (VMAX) ≥34 kt] regardless of 200–850-hPa vertical wind shear magnitude. (c),(d) As in (a),(b), but using Atlantic tropical cyclones of at least Saffir–Simpson category 2 intensity (i.e., VMAX ≥ 83 kt) and wind shear ≤15 kt. Note that the scale on the ordinate is inverted.

  • View in gallery

    Composite percentage of PR pixels with reflectivity ≥20 dBZ at the 2-km height level for (a) inner radii and (b) outer radii as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (i.e., VMAX ≥ 34 kt). (c),(d) Percentage anomalies from the daily mean coverage of the composite values in (a) and (b), respectively.

  • View in gallery

    Composite percentage of PR pixels with reflectivity ≥20 dBZ at the (a) 8- and (b) 10-km height level for inner radii as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (VMAX ≥ 34 kt) displayed as percentage anomalies from the daily mean coverage.

  • View in gallery

    Composite percentage of PR pixels with reflectivity ≥20 dBZ at the (a) 2-, (b) 8-, and (c) 10-km height level for inner radii as a function of LST for Atlantic tropical cyclones of at least hurricane strength (VMAX ≥ 64 kt) displayed as percentage anomalies from the daily mean coverage.

  • View in gallery

    Composite percentage of TMI pixels with a rain rate >0 mm h−1 for (a) inner radii (<500 km) and (b) outer radii (500–1000 km) as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (i.e., VMAX ≥ 34 kt) displayed as percentage anomalies from the daily mean coverage.

  • View in gallery

    Composite unconditional TMI rain rate (includes nonraining pixels) for (a) inner radii and (b) outer radii as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (i.e., VMAX ≥ 34 kt) displayed as percentage anomalies from the daily mean rain rate.

  • View in gallery

    Comparison of results from this study and the literature. 20-dBZ PR echo coverage at 2-km altitude (black solid) and TMI rain coverage (black dashed) at 200–300-km radius for hurricanes only (percent anomaly). Mean IR brightness temperature at 300-km radius in 31 major hurricanes (red solid; DTV14) and Hurricane Beatriz (1999; red dashed; Kossin 2002). Mean TRMM precipitation rates from inner 500 km of tropical cyclones (blue solid; Bowman and Fowler 2015), tropical west Pacific (blue dashed; Bowman et al. 2005), and oceanic mesoscale convective systems (blue dash–dot; Nesbitt and Zipser 2003). Rain rates from the literature are scaled to fit on the same axis, and the colors of the lines and y axes correspond with one another.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 223 212 21
PDF Downloads 190 178 14

Tropical Cyclone Diurnal Cycle as Observed by TRMM

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

Abstract

Previous work has indicated a clear, consistent diurnal cycle in rainfall and cold cloudiness coverage around tropical cyclones. This cycle may have important implications for structure and intensity changes of these storms and the forecasting of such changes. The goal of this paper is to use passive and active microwave measurements from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) and Precipitation Radar (PR), respectively, to better understand the tropical cyclone diurnal cycle throughout a deep layer of a tropical cyclone’s clouds.

The composite coverage by PR reflectivity ≥20 dBZ at various heights as a function of local standard time (LST) and radius suggests the presence of a diurnal signal for radii <500 km through a deep layer (2–10-km height) of the troposphere using 1998–2011 Atlantic tropical cyclones of at least tropical storm strength. The area covered by reflectivity ≥20 dBZ at radii 100–500 km peaks in the morning (0130–1030 LST) and reaches a minimum 1030–1930 LST. Radii between 300 and 500 km tend to reach a minimum in coverage closer to 1200 LST before reaching another peak at 2100 LST. The inner core (0–100 km) appears to be associated with a single-peaked diurnal cycle only at upper levels (8–10 km) with a maximum at 2230–0430 LST. The TMI rainfall composites suggest a clear diurnal cycle at all radii between 200 and 1000 km with peak rainfall coverage and rain rate occurring in the morning (0130–0730 LST).

Current affiliation: University of Louisiana at Monroe, Monroe, Louisiana.

Corresponding author address: Kenneth Leppert II, School of Science, University of Louisiana at Monroe, Hanna Rm. 306, 700 University Ave., Monroe, LA 71209. E-mail: leppert@ulm.edu

Abstract

Previous work has indicated a clear, consistent diurnal cycle in rainfall and cold cloudiness coverage around tropical cyclones. This cycle may have important implications for structure and intensity changes of these storms and the forecasting of such changes. The goal of this paper is to use passive and active microwave measurements from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) and Precipitation Radar (PR), respectively, to better understand the tropical cyclone diurnal cycle throughout a deep layer of a tropical cyclone’s clouds.

The composite coverage by PR reflectivity ≥20 dBZ at various heights as a function of local standard time (LST) and radius suggests the presence of a diurnal signal for radii <500 km through a deep layer (2–10-km height) of the troposphere using 1998–2011 Atlantic tropical cyclones of at least tropical storm strength. The area covered by reflectivity ≥20 dBZ at radii 100–500 km peaks in the morning (0130–1030 LST) and reaches a minimum 1030–1930 LST. Radii between 300 and 500 km tend to reach a minimum in coverage closer to 1200 LST before reaching another peak at 2100 LST. The inner core (0–100 km) appears to be associated with a single-peaked diurnal cycle only at upper levels (8–10 km) with a maximum at 2230–0430 LST. The TMI rainfall composites suggest a clear diurnal cycle at all radii between 200 and 1000 km with peak rainfall coverage and rain rate occurring in the morning (0130–0730 LST).

Current affiliation: University of Louisiana at Monroe, Monroe, Louisiana.

Corresponding author address: Kenneth Leppert II, School of Science, University of Louisiana at Monroe, Hanna Rm. 306, 700 University Ave., Monroe, LA 71209. E-mail: leppert@ulm.edu

1. Introduction

Previous work (e.g., Gray and Jacobson 1977; Yang and Slingo 2001; Nesbitt and Zipser 2003; Bowman et al. 2005) has shown a consistent diurnal cycle in rainfall and/or deep convection over tropical oceanic regions with a maximum in the morning. In particular, Bowman et al. (2005) used data from rain gauges on buoys and the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) and Precipitation Radar (PR) to better understand the characteristics of the diurnal cycle of rainfall over the ocean as observed by different instruments. Both the gauge and the TRMM data indicated a peak in oceanic rainfall between 0300 and 0600 local time.

Using satellite IR brightness temperature data, several studies (e.g., Browner et al. 1977; Muramatsu 1983; Lajoie and Butterworth 1984; Steranka et al. 1984; Kossin 2002; Dunion et al. 2014, hereafter DTV14) have specifically examined the cloud cover around tropical cyclones and found a clear diurnal cycle in the coverage of cold cloud tops around these storms. In contrast to tropical convection not associated with tropical cyclones, the timing of peak coverage by cold cloud tops around tropical cyclones varies depending on radius from the storm center. Using 36 Atlantic tropical cyclones, DTV14 found a consistent increase in the coverage by cold cloud tops near the inner core of each storm around the time of sunset, which subsequently propagated outward to several hundred kilometers over the course of the following day. This consistent cycle in cold cloud cover/convection around tropical cyclones may be associated with structure and intensity changes of these storms and have important implications for the forecasting of such changes. For example, DTV14 found that objective Dvorak intensity estimates (Olander and Velden 2007) for Atlantic tropical cyclones were associated with a significant diurnal signal. In addition, the numerical modeling work of Hobgood (1986) indicated a diurnal fluctuation in the intensity of a tropical cyclone in the early stages of their simulations while the storm was still relatively weak.

Browner et al. (1977) and Muramatsu (1983) hypothesized that the cyclical expansion of the cold cloud shield around tropical cyclones was related to inner-core convection and its associated outflow. In particular, they proposed that a peak in coverage by inner-core convection in the early morning hours [note that DTV14 found no evidence of a diurnal cycle over the inner core (100-km radius) using mean IR brightness temperatures] is followed by the expansion of cirrus outflow from the convection. As the cirrus cloud expands, it thins and dissipates, and the whole process begins again the next morning.

Most of the storms examined in Kossin (2002) showed no evidence of a diurnal cycle within 200 km from the storm center, suggesting that something other than a diurnal cycle of inner-core convection was responsible for the changes in cloud cover observed farther from the storm center. Thus, Kossin (2002) argued that the diurnal cycle in cloud cover around tropical cyclones was related to changes in radiational cooling. In particular, as the sun sets, net radiational cooling results in subsidence, and this subsidence is stronger in clear areas than in cloudy areas (e.g., Gray and Jacobson 1977). Thus, stronger subsidence would be expected to be present beyond the outer edge of the cirrus canopy relative to regions closer to the storm core under the canopy, and this radial gradient in subsidence may cause the cirrus canopy to shrink over the course of a night. The hypothesis put forth by Kossin (2002) requires convection to create and maintain the cirrus canopy around a tropical cyclone, but the cyclical nature of the canopy is driven by the nightly erosion of the canopy by subsidence, not a morning increase in inner-core convection. Hobgood (1986) also suggested that the diurnal cycle of tropical cyclone cold cloud tops is tied to the diurnal cycle of radiation near cloud tops and the differences in net radiation between clear and cloudy regions.

In addition to a single-peaked diurnal cycle of cold cloud tops around tropical cyclones, a semidiurnal cycle has also been observed (e.g., Browner et al. 1977; Lajoie and Butterworth 1984; Steranka et al. 1984; Kossin 2002). Lajoie and Butterworth (1984) hypothesized that this semidiurnal cycle was the result of the influence of convective downdrafts on the boundary layer. In particular, downdrafts from the early morning peak of convection may act to cool and dry the boundary layer resulting in a subsequent minimum in convection and cold cloud tops. After the boundary layer recovers, a new round of convection can develop resulting in a second maximum in cold cloud tops. Li and Wang (2012) also hypothesized that the quasi-periodic nature of outer rainbands in their numerically simulated tropical cyclone was due to similar thermodynamic effects of convective downdrafts on the boundary layer and subsequent recovery. In contrast, Kossin (2002) hypothesized that the semidiurnal cycle in cloud cover around tropical cyclones is due to oscillations in local lapse rates linked to the solar semidiurnal atmospheric tide S2.

While some studies have examined the diurnal cycle of cold cloud tops around tropical cyclones, others have examined the diurnal cycle of rainfall around these storms (Lajoie and Butterworth 1984; Jiang et al. 2011; Wu et al. 2015; Bowman and Fowler 2015). Using data from an island in the southwest Pacific, Lajoie and Butterworth (1984) found a diurnal signal in tropical cyclone rainfall with a maximum (minimum) in rainfall near 0500 (1900) local standard time (LST). Wu et al. (2015) used the gridded 3B42 TRMM rainfall product to examine the rainfall around tropical cyclones in all major regions around the world where these storms form. While they found some variability in the phasing of the diurnal cycle with different storm intensities, storm locations, and/or radii, a peak in tropical cyclone rainfall was nearly always found in the morning hours (typically 0300–0600 LST), consistent with general tropical oceanic rainfall (e.g., Nesbitt and Zipser 2003; Bowman et al. 2005). Jiang et al. (2011) and Bowman and Fowler (2015) also found a morning maximum (0430–0730 LST) in tropical cyclone rainfall using the 3B42 product.

The presence of a diurnal cycle in tropical cyclone rainfall, in addition to that observed using IR brightness temperatures, suggests that this cycle may be related to processes occurring through a deep layer of the troposphere, not only at cloud top. In addition, DTV14 provide evidence that the tropical cyclone diurnal signal is present through a deep layer of the storm through the use of 37- and 85–91-GHz passive microwave imagery. Motivated by their case studies and the apparent utility of microwave frequencies, which allows for a more direct investigation of the rainfall organization underneath the cirrus canopy, the goal of this study is to use passive and active microwave measurements from the TMI and PR, respectively, to better understand the tropical cyclone diurnal cycle throughout a large depth of the storm’s clouds.

2. Data and methodology

The TRMM satellite was launched in November 1997 with a 35° orbit inclination angle and an initial altitude of ~350 km (Kummerow et al. 1998). An orbit boost to ~400 km in 2001 extended the mission duration into early 2015. The inclination of TRMM’s orbit allowed it to sample the full diurnal cycle every ~6 weeks (Simpson et al. 1988). The PR on board TRMM operated at 13.8 GHz with a swath width of 247 km (215 km prior to orbit boost), a vertical resolution of 0.25 km, and a nadir horizontal resolution of ~5.0 km [~4.3 km prior to orbit boost; Kummerow et al. (1998); Kozu et al. (2001)].

The TMI is a nine-channel, conically scanning passive microwave radiometer with a swath width of 878 km [~760 km prior to orbit boost; Kummerow et al. (1998)]. The TRMM 2A12 algorithm provides surface rainfall rates derived from TMI measurements with a horizontal resolution similar to that of the TMI 85.5-GHz channels (~7 km × 5 km). The version of this algorithm used here (V7) is a Bayesian scheme that utilizes cloud-resolving model simulations to create a priori databases that are constrained by PR reflectivity and TMI brightness temperature observations (Kummerow et al. 2011). These databases include simulated brightness temperatures at TMI frequencies, and rainfall rates are retrieved by matching observed TMI brightness temperatures with those in the databases. The V7 2A12 algorithm also outputs a probability of precipitation for each pixel over water. Because a given set of observed TMI brightness temperatures is probably similar to at least one database entry with rain, many pixels may be assigned a nonzero rain rate but be associated with a very small probability of precipitation. Thus, the probability of precipitation is an important parameter in assessing whether a particular pixel is actually a raining pixel.

The TRMM satellite also carried a five-channel imaging spectroradiometer called the Visible and Infrared Scanner (VIRS; Kummerow et al. 1998). This instrument measured radiances in the visible to IR (0.6–12.0 μm) parts of the electromagnetic spectrum over an 833-km swath width (720 km before orbit boost). The horizontal resolution of VIRS was ~2.4 km (~2.2 km prior to boost).

The University of Utah TRMM database uses data from the various sensors carried on the TRMM satellite to identify precipitation features and includes various statistics for these features (Nesbitt et al. 2000; Liu et al. 2008). The first step in the creation of this database was to collocate data from the various TRMM instruments and store the resulting pixel level data in the level-1 data files. Here we use the PR reflectivity profiles, TMI 2A12 surface rain rates and precipitation probabilities, and VIRS channel 4 (10.8 μm) brightness temperatures from the level-1, V7 data files of the precipitation feature database (data are available online at http://trmm.chpc.utah.edu/data.html).

Following Jiang et al. (2011), the tropical cyclone center locations from the National Hurricane Center best track data (Landsea and Franklin 2013) were interpolated to the time of each TRMM overpass. These interpolated center locations were then used to extract all PR and TMI pixels within 1000 km of the center of each tropical cyclone in the Atlantic basin between 1998 and 2011 (208 storms). Data were composited according to radius from the storm center (100-km bins from 0 to 1000 km) and LST (3-h bins centered on 0000, 0300, 0600 LST, etc.) using data only when storms were >300 km from land (DTV14) and were at least at tropical storm strength [maximum sustained wind ≥34 kt (1 kt = ~0.5144 m s−1). We also tried compositing the data relative to the radius of maximum wind using values of this radius from the extended best track dataset produced by Colorado State University (Demuth et al. 2006) and/or the time of local sunset, but the results were similar to those using uniform 100-km radial intervals and relative to LST.

Specifically, PR data were composited by calculating the percentage of pixels in each radial and time bin with reflectivity ≥20 dBZ at various heights (2–14 km in increments of 2 km). Prior to compositing the TMI rainfall data, rain rates of all pixels with probability of precipitation <50% were set to 0 mm h−1 as recommended in NASA (2016). Then TMI rainfall composites were created by calculating the percentage of pixels with rain (i.e., rain rate > 0 mm h−1) and separately calculating the mean rain rate as a function of radius and time. We also examined the composite coverage by 37.0- and 85.5-GHz polarization-corrected temperatures (Spencer et al. 1989; Cecil et al. 2002) below various thresholds (260–275 K in increments of 5 K for 37.0 GHz and 200–270 K in increments of 10 K for 85.5 GHz), but these composites were found to be noisy and not very useful for examining the tropical cyclone diurnal cycle. Both PR and TMI rainfall composites were created using data from all overpasses of storms that met the intensity and distance from land thresholds above. Subsets were tested using hurricanes only and, also, using only data for storms experiencing 200–850-hPa wind shear ≤15 kt (DTV14) using wind shear information from the Statistical Hurricane Intensity Prediction Scheme developmental dataset (available online at http://rammb.cira.colostate.edu/research/tropical_cyclones/ships/developmental_data.asp).

Each TRMM overpass does not necessarily sample the full area of a radial bin. To account for this, the effective sample size Ne is used. The effective sample size provides a better estimate of the number of independent samples used in each composite compared to the number of individual satellite pixels. Specifically, Ne provides an estimate of the number of satellite overpasses that sampled a combined area equal to the full area of a particular radial band calculated according to
e1
where Ar is the area of a particular radial band r, N is the total number of satellite overpasses that sampled r, and is the area observed on the kth satellite overpass.

We focus on composite analyses in this paper because the irregular sampling from TRMM and additional passive microwave sensors [e.g., Special Sensor Microwave Imager (SSM/I), Special Sensor Microwave Imager/Sounder (SSMIS)] is not suitable for diurnal cycle analysis of individual storms. Results from several case studies that combined data from multiple radiometers (not shown) confirm this conclusion.

3. Results

a. VIRS composites

For a more direct comparison to previous work that examined the tropical cyclone diurnal cycle using satellite IR brightness temperatures, we created composites using VIRS IR brightness temperature data. Specifically, the purpose of these IR composites is to illustrate that the methodology (and sample) employed here leads to different results than those of DTV14. Figure 1 shows the mean IR brightness temperature as a function of LST and radius using Atlantic storms of different intensities, and Table 1 shows Ne used for these composites. Table 1 shows that for a given radius, there does not appear to be any consistent trends in Ne with respect to time. The effective sample size tends to decrease at large radii because we did not include overpasses that merely graze the edge of a tropical cyclone on one side.

Fig. 1.
Fig. 1.

Composite VIRS IR brightness temperature (BT) for (a) inner radii (<500 km) and (b) outer radii (500–1000 km) as a function of LST for Atlantic tropical cyclones of at least tropical storm strength [i.e., maximum sustained wind (VMAX) ≥34 kt] regardless of 200–850-hPa vertical wind shear magnitude. (c),(d) As in (a),(b), but using Atlantic tropical cyclones of at least Saffir–Simpson category 2 intensity (i.e., VMAX ≥ 83 kt) and wind shear ≤15 kt. Note that the scale on the ordinate is inverted.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

Table 1.

Effective sample size for the VIRS composites as a function of radius and LST (sampling window is ±1.5 h) using all Atlantic storms of at least (top) tropical storm strength (i.e., VMAX ≥ 34 kt) regardless of 200–850-hPa vertical wind shear magnitude and (bottom) Saffir–Simpson category 2 strength (VMAX ≥ 83 kt) under the influence of wind shear ≤15 kt.

Table 1.

Figures 1a and 1b show VIRS composites using all Atlantic tropical cyclones of at least tropical storm intensity regardless of 200–850-hPa vertical wind shear magnitude (i.e., tropical storm composites), and Figs. 1c and 1d show similar composites except using storms of at least Saffir–Simpson category 2 intensity under the influence of wind shear ≤15 kt (i.e., category 2 composites). The low intensity threshold and lack of a shear threshold for Figs. 1a and 1b maximizes our sample size, while the thresholds in Figs. 1c and 1d allow the closest comparison to the composites of DTV14. Note that for Fig. 1 and subsequent figures, 100-km radial bands that showed similar patterns were combined to make the plots easier to read. The VIRS tropical storm composites suggest a weak diurnal cycle for the 0–200-km radii with a maximum brightness temperature between 1330 and 2230 LST. (DTV14 found a maximum brightness temperature for the 200-km radius slightly later at 2000–0200 LST.) The 0–200-km radii have two relative minima in brightness temperature at 0000 and 0600 LST. The timing of the second minima is consistent with DTV14 who found broad minima from early morning through afternoon at a radius of 200 km. The 200–300-km radius shows no evidence of a diurnal cycle in Fig. 1a, whereas the 300–500-km radii indicate a weak diurnal cycle with a minimum at 2100 LST (the minimum for the 400-km radius was found by DTV14 to occur earlier between 1400 and 2000 LST). At larger radii (Fig. 1b), the VIRS tropical storm composites provide some indication of a diurnal cycle with the coldest brightness temperatures between 1630 and 2230 LST. Radii between 500 and 800 km also suggest a secondary minimum at 0300 LST. The 600-km radius composites of DTV14 (largest radius examined in that study) indicate the occurrence of a minimum IR brightness temperature at 1600–2000 LST which is consistent with the primary minimum found here for the 500–700-km radii (Fig. 1b). The VIRS category 2 composites (Figs. 1c,d) are generally much noisier than the tropical storm composites with sample sizes too small for robust conclusions (Table 1). Where a clear diurnal cycle is indicated by the category 2 composites (i.e., 300–500-km radii), the time of maximum and minimum IR brightness temperature does not change relative to the corresponding tropical storm composites.

For a better understanding of our composite time series, a spectral analysis was carried out using the procedure described in Panofsky and Brier (1958). Specifically, autocorrelation coefficients were calculated for each time series for lags 0–m, where m is half the number of values in the time series (in this case, m = 4). Then a harmonic analysis was applied to the coefficients, and the resulting coefficients were smoothed by a 3-point weighted moving average (weights of 0.25, 0.50, and 0.25) to account for nonphysical small negative coefficients. Note that results of this spectral analysis with and without the 3-point smoothing are qualitatively similar. Similar to Bowman et al. (2005) and Wu et al. (2015), the statistical significance of the diurnal and semidiurnal harmonics were tested using the F statistic with degrees of freedom of 2 and 5. Specifically, this test determines whether the variance explained by one harmonic is statistically significantly greater than the combined variance explained by all the other harmonics. Table 2 shows the diurnal harmonic amplitude and phase for the VIRS composites. The semidiurnal harmonic amplitude and phase are not shown because this harmonic was not statistically significant for many of the composites used in this study, including VIRS. Results of the statistical analysis indicate the diurnal harmonic explains significantly more of the variance than the other harmonics combined valid at or above the 95% level (for conciseness, in the rest of the paper we will simply say that the diurnal harmonic is statistically significant) for VIRS tropical storm composites at the 0–200-, 400–500-, and 700–1000-km radii (Table 2). The diurnal harmonic of the VIRS category 2 composites is not statistically significant, except for the 300–500-km radii.

Table 2.

Amplitude (K) and phase (LST) of the diurnal harmonic for the VIRS composites as a function of radius using all Atlantic storms of at least (left) tropical storm strength (VMAX ≥ 34 kt) regardless of 200–850-hPa vertical wind shear magnitude and (right) Saffir–Simpson category 2 strength (VMAX ≥ 83 kt) under the influence of wind shear ≤15 kt. Italic (bold) values indicate that the amplitude is statistically significant at the 95% (99%) level.

Table 2.

In general, the timing of coldest IR brightness temperatures found here does not match that of DTV14 because the outward propagation of the diurnal pulse with time observed in DTV14 is not observed in Fig. 1. DTV14 found that the outward-propagating diurnal pulse is most apparent in time series of 6-h IR brightness temperature differences. It is possible that 6-h differences in VIRS IR brightness temperatures may show an outward-propagating signal, but the irregular temporal sampling by TRMM precludes this type of analysis. In addition, DTV14 only examined tropical cyclones near their maximum intensity. In contrast, the composites shown in Fig. 1 include storms at all stages of their life cycle as long as they met the indicated intensity threshold. It is possible that the outward propagation of the diurnal cloud signal around tropical cyclones is primarily a characteristic of tropical cyclones near their maximum intensity and/or embedded within favorable environmental conditions for intensification (e.g., weak vertical wind shear, high moisture, etc.). Our sample size is not sufficient for investigating this.

b. PR composites

Figures 2a and 2b show the composite percentage coverage by PR pixels with reflectivity ≥20 dBZ at a height of 2 km as a function of LST using storms of at least tropical storm strength, and Table 3 (top) shows the corresponding Ne. As before, 100-km radial bins that displayed similar patterns were combined. In absolute terms, the temporal variability associated with the radial bands between 300 and 500 km shown in Fig. 2a is small because most of that region is rain free at any time of day. To amplify and better show any potential diurnal signal, we calculated the percentage anomaly (Figs. 2c,d) from the daily mean values for each radius. The remainder of the paper will primarily focus on these percentage anomalies so that the variability can be more readily observed at all radii, even those with very small overall coverage of precipitation.

Fig. 2.
Fig. 2.

Composite percentage of PR pixels with reflectivity ≥20 dBZ at the 2-km height level for (a) inner radii and (b) outer radii as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (i.e., VMAX ≥ 34 kt). (c),(d) Percentage anomalies from the daily mean coverage of the composite values in (a) and (b), respectively.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

Table 3.

Effective sample size for the PR composites using all Atlantic storms of at least (top) tropical storm strength (VMAX ≥ 34 kt) and (bottom) hurricane strength (VMAX ≥ 64 kt) as a function of radius and LST (sampling window is ±1.5 h).

Table 3.

Similar to what Kossin (2002) and DTV14 found using IR brightness temperature data, there is little indication of a clear diurnal cycle for the innermost radius at the 2-km height using PR data (Figs. 2a,c). However, larger radii do appear to be associated with a diurnal cycle. Specifically, the radii between 100 and 500 km have peak coverage between 0130 and 0730 LST {diurnal harmonic is statistically significant for the radii between 100 and 400 km [Table 4 (top left)]}, and the 300–500-km radii have a secondary peak at 2100 LST (semidiurnal harmonic is not statistically significant). This peak coverage at 0130–0730 LST is consistent with several previous studies that found a maximum in cold cloudiness and/or rainfall around tropical cyclones during the morning hours (e.g., Muramatsu 1983; Lajoie and Butterworth 1984; Wu et al. 2015; Bowman and Fowler 2015). DTV14 found peak IR cooling between 0400 and 0800 LST for the 200-km radius, which is consistent with the time of peak coverage by 20-dBZ radar echo shown in Fig. 2 for the 100–300-km radius (i.e., 0600 LST). However, the outward propagation of the diurnal signal observed in DTV14 is not observed in the PR composites shown in Fig. 2, similar to the VIRS composites.

Table 4.

(top) Amplitude (percent) and phase (LST) of the diurnal harmonic for the PR composites as a function of radius using all Atlantic storms of at least tropical storm strength (VMAX ≥ 34 kt) at the (left) 2-, (middle) 8-, and (right) 10-km height. (bottom) As in (top), but using only storms of hurricane strength (VMAX ≥ 64 kt). Italic and bold values are as in Table 2.

Table 4.

Beyond 500 km, the time series of coverage by PR reflectivity at 2 km ≥20 dBZ and corresponding percent anomalies (Figs. 2b and 2d, respectively) are noisy without any clear indication of a diurnal cycle. Consistent with having no clear signal, the diurnal harmonic is generally not statistically significant for any of the outer radii [Table 4 (top left)]. A similar result is generally observed at outer radii for other heights in the PR composites. Thus, the following discussion pertaining to the PR composites will focus on the inner radii.

Figure 3 shows the percent anomaly in coverage by PR reflectivity ≥20 dBZ at a height of 8 and 10 km as a function of time and radius for Atlantic tropical cyclones of at least tropical storm intensity. In general, a comparison between Figs. 2 and 3 indicates that the pattern observed for inner radii at the 2-km height extends through a deep layer up to 10 km. Specifically, the radial bands between 100 and 300 km are associated with a single-peaked diurnal cycle {diurnal harmonic is statistically significant, except for the 100–200-km radius at 10 km [Tables 4 (top middle and top right)]}, and the radial bands between 300 and 500 km have a double-peaked cycle. The maximum for the 100–300-km radii occurs at 0600 (0900) LST at a height of 8 (10) km. Similar to the 2-km height, the peak coverage for radii 300–500 km occurs at 0300 and 2100 LST at a height of both 8 and 10 km. In contrast to what is observed at lower levels, the 0–100-km radius appears to be associated with a single-peaked diurnal cycle at both 8 and 10 km with the maximum/minimum occurring slightly later at 10 km (0300/1800 LST) than at 8 km (0000/1500 LST). This diurnal harmonic for the 0–100-km radius is statistically significant only at the 10-km height.

Fig. 3.
Fig. 3.

Composite percentage of PR pixels with reflectivity ≥20 dBZ at the (a) 8- and (b) 10-km height level for inner radii as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (VMAX ≥ 34 kt) displayed as percentage anomalies from the daily mean coverage.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

Steranka et al. (1984) found a maximum (minimum) in cold cloudiness coverage over the inner core at 0100 (1300) LST for hurricanes and at 0800 (2000) LST for tropical storms. The timing of maximum 20-dBZ reflectivity coverage shown in Fig. 3 for the 0–100-km radius appears to be more consistent with the hurricane composite of Steranka et al. (1984), while the minima in Fig. 3 occur between the times of minimum cloudiness coverage for hurricanes and tropical storms found in Steranka et al. (1984). Wu et al. (2015) found a maximum/minimum in inner-core rainfall for weak storms (tropical storm and category 1 hurricanes) in the Atlantic basin at 0600/1800 LST (their composites for stronger hurricanes revealed no statistically significant diurnal cycle in the Atlantic basin). Thus, the rainfall maximum from Wu et al. (2015) occurs later than the maxima shown in Fig. 3 here, but the minima occur at similar times.

The PR composites for tropical cyclones with maximum sustained wind speed ≥34 kt suggest an increase in coverage by hydrometeor mass between heights of 8–10 km in the inner core during the evening hours that reaches a peak in the middle of the night (2230–0430 LST). The coverage by rainfall/convection at radii between 100 and 500 km peaks a little later in the morning (0130–1030 LST) without any indication of outward propagation with time among these radii. As the day progresses, the PR composites indicate that the coverage by rainfall/convection decreases until the afternoon hours (1800 LST) for radii between 100 and 300 km. In contrast, radii between 300 and 500 km reach a minimum in coverage somewhat earlier (1200 LST) before rapidly increasing to another peak in coverage at 2100 LST. Kossin (2002) hypothesized that the semidiurnal oscillation observed in IR brightness temperatures with peaks in colder cloud tops near 0400 and 1600 LST was consistent with the semidiurnal atmospheric tide S2. However, the timing (2100 LST) of the later peak in coverage observed in the PR composites for radii between 300 and 500 km may not be consistent with the S2 atmospheric tide. Alternatively, the semidiurnal cycle observed for these radii may be consistent with the hypothesis of Lajoie and Butterworth (1984) and Li and Wang (2012) involving the impacts of convective downdrafts on the boundary layer described previously. It is also possible that the 300–500-km radii are actually associated with a single, broad peak in coverage interrupted by noise, which may be consistent with the finding of a lack of a statistically significant semidiurnal cycle at these radii. Note that all radii and times may be impacted to some extent by sampling fluctuations or noise.

Composites were also created using only data from storms that were at hurricane strength (Fig. 4). The corresponding sample sizes for these composites are provided in Table 3 (bottom). Composites limited to stronger hurricanes tend to be even noisier, likely due to relatively small sample sizes, and are not shown here. A comparison between Figs. 4a and 2c, Figs. 4b and 3a, and Figs. 4c and 3b indicate somewhat similar patterns. Specifically, the times of maximum and minimum coverage generally do not change between samples that are limited to hurricanes only and those that also include tropical storms. The larger amplitudes shown in Fig. 4 relative to the corresponding panels in Figs. 2 and 3 are likely due to the smaller, hurricane-only sample size used for Fig. 4 (Table 3). Thus, limiting the sample to hurricanes does not temporally shift the diurnal cycle in 20-dBZ reflectivity coverage relative to a sample that also includes tropical storms. The amplitudes do appear to get larger for the smaller sample. This is in contrast to what was found by Browner et al. (1977), Steranka et al. (1984), and Hobgood (1986) who all showed a decrease in amplitude of the tropical cyclone diurnal cycle with increasing intensity. Note that limiting composites to times when storms were experiencing 200–850-hPa wind shear ≤15 kt following DTV14 (not shown) also had little impact on the results.

Fig. 4.
Fig. 4.

Composite percentage of PR pixels with reflectivity ≥20 dBZ at the (a) 2-, (b) 8-, and (c) 10-km height level for inner radii as a function of LST for Atlantic tropical cyclones of at least hurricane strength (VMAX ≥ 64 kt) displayed as percentage anomalies from the daily mean coverage.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

c. TMI composites

The TMI swath is more than 3 times wider than the PR swath, so the TMI-based sample sizes for these composites (Table 5) are much larger than those from PR (Table 3). Figure 5 shows the composite percentage coverage by TMI raining pixels shown as percent anomalies relative to the daily mean coverage for all storms of at least tropical storm strength. The patterns shown by the TMI composites for the inner radii in Fig. 5a are quite similar to the patterns shown by the low-level PR composites in Fig. 2. In particular, the TMI rain coverage composite for radii near the storm core (0–200-km radii) shows no indication of a diurnal cycle. Farther away from the storm center, the 200–300-km radius appears to be associated with a single-peaked diurnal cycle in rain coverage with a peak at 0600 LST, and the radii between 300–500 km exhibit a double-peaked cycle with peaks at 0300 and 2100 LST. The statistical analysis [Table 6 (left)] indicates that the diurnal harmonic associated with the TMI rain coverage composites is statistically significant for the inner radii beyond 200 km, but the semidiurnal cycle is not significant (not shown) for the 300–500-km radii. In contrast to the PR composites, the TMI rain coverage composites for the outer radii (Fig. 5b) exhibit a clear diurnal cycle (diurnal harmonic is statistically significant for all outer radii at the 99% level). All outer radii show a morning peak in rain coverage (0300 LST) with a minimum in the middle of the day (0730–1630 LST). Figure 2d from the PR is actually consistent with this, but with much greater noise from that smaller sample (Table 3). Thus, the differences observed between the PR and TMI composites for the outer radii may be due to differing sample sizes. The phasing of the TMI rain coverage composites are similar to several previous studies (e.g., Lajoie and Butterworth 1984; Wu et al. 2015; Bowman and Fowler 2015), which found a morning maximum in rain rate around tropical cyclones. However, results from Wu et al. (2015) and DTV14 indicate that the diurnal cycle propagates away from the storm center with time. No such outward propagation is observed in the TMI composites shown in Fig. 5.

Table 5.

Effective sample size for the TMI composites using all Atlantic storms of at least tropical storm strength as a function of radius and LST (sampling window is ±1.5 h).

Table 5.
Fig. 5.
Fig. 5.

Composite percentage of TMI pixels with a rain rate >0 mm h−1 for (a) inner radii (<500 km) and (b) outer radii (500–1000 km) as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (i.e., VMAX ≥ 34 kt) displayed as percentage anomalies from the daily mean coverage.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

Table 6.

Amplitude and phase (LST) of the diurnal harmonic for the composite (left) percentage of TMI pixels with a rain rate >0 mm h−1 and (right) unconditional TMI rain rate (mm h−1) as a function of radius using all Atlantic storms of at least tropical storm strength. Italic and bold values are as in Table 2.

Table 6.

The composite mean TMI rain rate (Fig. 6) for all storms of at least tropical storm strength generally suggest patterns similar to the composite rain coverage (Fig. 5), except that the composite rain rates are somewhat noisier. In particular, the 0–200-km radii show no indication of a clear diurnal cycle, and the 300–500-km radii have two maxima (0300 and 2100 LST). The outer radii show a peak rain rate 0130–0730 LST. The radii between 500 and 900 km appear to be associated with another peak in rain rate at 2100 LST, but these radii are associated with maximum power at a frequency of 1 day−1 {diurnal harmonic is statistically significant for all outer radii except the 500–600-km radius [Table 6 (right)]}. One difference between the composite TMI rain coverage and rain rate is that the mean rain rate for the 200–300-km radius is associated with two peaks (0600 and 2100 LST; Fig. 6), whereas the mean rain coverage is associated with a single peak at 0600 LST (Fig. 5). Note that, similar to the PR composites, limiting the TMI composites to hurricanes only (not shown) does not substantially change the results relative to composites that also include tropical storms.

Fig. 6.
Fig. 6.

Composite unconditional TMI rain rate (includes nonraining pixels) for (a) inner radii and (b) outer radii as a function of LST for Atlantic tropical cyclones of at least tropical storm strength (i.e., VMAX ≥ 34 kt) displayed as percentage anomalies from the daily mean rain rate.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

4. Conclusions

Previous work has shown a clear diurnal cycle in cold cloud tops around tropical cyclones using satellite IR data (e.g., Browner et al. 1977; Kossin 2002; DTV14) and/or rainfall data (e.g., Bowman and Fowler 2015). The aim of this study is to use spaceborne passive and active microwave measurements to better understand the tropical cyclone diurnal cycle throughout a deep layer of the storm’s clouds. The diurnal cycle of cloudiness/rainfall around tropical cyclones may have important implications for structure and intensity changes of these storms and the forecasting of these changes.

The composite percentage coverage by PR reflectivity ≥20 dBZ at various heights suggest the presence of a diurnal signal for radii <500 km through a deep layer (2–10-km height) of the troposphere for tropical cyclones in the Atlantic of at least tropical storm strength. The inner core (0–100-km radius) appears to be associated with an increase in coverage by hydrometeor mass between heights of 8 and 10 km during the evening hours, reaching a peak in the middle of the night (2230–0430 LST) before steadily decreasing to a minimum between 1330 and 1930 LST. The coverage by rainfall/convection at radii between 100 and 500 km peaks a little later in the morning (0130–1030 LST) at 2–10-km height without any indication of outward propagation with time among these radii. As the day progresses, the PR composites indicate that the coverage of rainfall decreases until the afternoon hours (1030–1930 LST) for radii 100–500 km. After achieving a minimum at 1200 LST, the 20-dBZ coverage for radii between 300 and 500 km rapidly increases to another peak at 2100 LST. It is not clear what may be the cause of this semidiurnal oscillation for radii 300–500 km, because the timing of this cycle is not consistent with the timing of the S2 semidiurnal atmospheric tide (Kossin 2002). The PR composites for radii beyond 500 km from the storm center are somewhat noisy, and, as a consequence, do not exhibit a clear diurnal cycle. Note that our PR composites all relate well to the areal coverage of precipitation, but not necessarily to precipitation intensity; however, on the spatial scales examined here, the precipitation coverage and intensity are highly correlated.

The composite coverage by TMI rainfall exhibits a diurnal cycle for all radii between 200 and 1000 km with a maximum coverage in the morning between 0130 and 0730 LST. The 300–500-km radii also show a second peak at 2100 LST, but it is possible that the peaks in coverage at 0300 and 2100 LST for these radii are actually a single broad peak interrupted by noise. The composite TMI rain rate generally shows similar patterns to the rain coverage, except that the rain-rate composites are noisier.

One key difference between the results shown here and previous studies (e.g., DTV14; Wu et al. 2015) is that the composites shown herein do not indicate a propagation of the diurnal signal away from the storm center with time. The sampling from low Earth orbit satellites precludes exploring this in ways directly analogous to DTV14. It is also possible, though not conclusive from our results, that the outward-propagating signal is manifest near cloud top but not through deep-layer precipitation. Although Wu et al. (2015) reported an outward propagating signal in precipitation estimates, those estimates were largely derived from IR brightness temperatures.

Figure 7 compares results from the diurnal cycle analysis using the PR and TMI composites from this study to those from previous studies. Specifically, Fig. 7 shows the composite coverage by PR reflectivity ≥20 dBZ at a height of 2 km and the TMI rain coverage for the 200–300-km radius using hurricanes only. Also shown in Fig. 7 are portions of time series from DTV14 and Kossin (2002) using mean IR brightness temperatures for the 300-km radius around tropical cyclones, a time series of mean tropical cyclone rain rates from Bowman and Fowler (2015), and a time series from Bowman et al. (2005) and Nesbitt and Zipser (2003) showing the diurnal cycle of general oceanic rainfall. The PR and TMI composites, as well as all the rainfall time series, including those not limited to tropical cyclone rainfall, show a peak in the morning (~0600 LST) and a minimum during the afternoon (1500–1800 LST). It might be expected that the peak in rain rate/coverage be associated with the coldest cloud tops. However, the coldest mean IR brightness temperatures shown in Fig. 7 occur at ~1500 LST, about 9 h after the peak in rainfall and nearly coinciding with the minima in rain and echo coverage. Thus, the mean IR brightness temperatures seem to reflect an expansion in cold anvil cloud several hours after the morning peak in rainfall around tropical cyclones.

Fig. 7.
Fig. 7.

Comparison of results from this study and the literature. 20-dBZ PR echo coverage at 2-km altitude (black solid) and TMI rain coverage (black dashed) at 200–300-km radius for hurricanes only (percent anomaly). Mean IR brightness temperature at 300-km radius in 31 major hurricanes (red solid; DTV14) and Hurricane Beatriz (1999; red dashed; Kossin 2002). Mean TRMM precipitation rates from inner 500 km of tropical cyclones (blue solid; Bowman and Fowler 2015), tropical west Pacific (blue dashed; Bowman et al. 2005), and oceanic mesoscale convective systems (blue dash–dot; Nesbitt and Zipser 2003). Rain rates from the literature are scaled to fit on the same axis, and the colors of the lines and y axes correspond with one another.

Citation: Monthly Weather Review 144, 8; 10.1175/MWR-D-15-0358.1

Previous studies (e.g., Browner et al. 1977; Hobgood 1986) found that more intense tropical cyclones are associated with a weaker diurnal signal. However, the PR and TMI composites used here show little sensitivity to storm intensity. Strong vertical wind shear may create asymmetries in the distribution of rainfall and convection around tropical cyclones (e.g., Cecil 2007; Hence and Houze 2011, 2012) that may disrupt a tropical cyclone’s typical diurnal variation in structure. However, limiting composites to times when storms were associated with 200–850-hPa shear ≤15 kt did not result in a more apparent diurnal cycle in any of the composites. If anything, the shear-limited composites were somewhat noisier than composites without the shear threshold, probably due to a smaller sample size. In addition, composites created relative to LST were similar to those relative to the time of local sunset, and composites using regular 100-km radial bands showed generally similar results to those using radial bands relative to the radius of maximum wind.

Browner et al. (1977) and Muramatsu (1983) hypothesized that the diurnal cycle observed using IR brightness temperatures was the result of an early morning peak in convection near the inner core of a storm followed by an expansion of cirrus outflow over the next ~12 h. However, the diurnal cycle observed in the PR and TMI composites suggests that the tropical cyclone diurnal cycle away from the storm core is not confined to high-level cirrus clouds, but extends through a deep layer of a storm’s clouds. A diurnal cycle in the coverage and/or intensity of a tropical cyclone’s convection/rainfall and associated cycle in latent heating may be related to a diurnal signal in tropical cyclone intensity. DTV14, indeed, found a significant diurnal cycle in intensity estimates of tropical cyclones. A diurnal signal in storm intensity and/or structure may have a substantial effect on what impacts a landfalling tropical cyclone, depending on what time of day the storm makes landfall. Therefore, future work should further examine the potential relation between a tropical cyclone’s diurnal cycle in cloud cover, rainfall, and/or convection and its intensity. In addition, future work should examine tropical cyclones in other basins to determine if the diurnal cycle observed using PR and TMI data in the Atlantic is also observed elsewhere. Other future work could combine TMI data with data from other passive microwave sensors (e.g., SSM/I, Global Precipitation Measurement Microwave Imager, etc.) for a larger, more robust sample.

Acknowledgments

Funding for this research was generously provided through NASA Grants NNX12AK70G and NNM11AA01A. The authors thank Dr. Haiyan Jiang for her help in identifying TRMM orbits that passed over tropical cyclones and Dr. Jason Dunion for his helpful suggestions for conducting this work. The authors are also grateful to the University of Utah for providing the TRMM data, Colorado State University for providing the radius of maximum wind and wind shear data, and the helpful suggestions from Jon Zawislak and another anonymous reviewer that led to the improvement of the manuscript.

REFERENCES

  • Bowman, K. P., and M. D. Fowler, 2015: The diurnal cycle of precipitation in tropical cyclones. J. Climate, 28, 53255334, doi:10.1175/JCLI-D-14-00804.1.

    • Search Google Scholar
    • Export Citation
  • Bowman, K. P., J. C. Collier, G. R. North, Q. Wu, E. Ha, and J. Hardin, 2005: Diurnal cycle of tropical precipitation in Tropical Rainfall Measuring Mission (TRMM) satellite and ocean buoy rain gauge data. J. Geophys. Res., 110, D21104, doi:10.1029/2005JD005763.

    • Search Google Scholar
    • Export Citation
  • Browner, S. P., W. L. Woodley, and C. G. Griffith, 1977: Diurnal oscillation of the area of cloudiness associated with tropical storms. Mon. Wea. Rev., 105, 856864, doi:10.1175/1520-0493(1977)105<0856:DOOTAO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Cecil, D. J., 2007: Satellite-derived rain rates in vertically sheared tropical cyclones. Geophys. Res. Lett., 34, L02811, doi:10.1029/2006GL027942.

    • 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, doi:10.1175/1520-0493(2002)130<0769:RISALC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Demuth, J. L., M. DeMaria, and J. A. Knaff, 2006: Improvement of Advanced Microwave Sounding Unit tropical cyclone intensity and size estimation algorithms. J. Appl. Meteor. Climatol., 45, 15731581, doi:10.1175/JAM2429.1.

    • Search Google Scholar
    • Export Citation
  • Dunion, J. P., C. D. Thorncroft, and C. S. Velden, 2014: The tropical cyclone diurnal cycle of mature hurricanes. Mon. Wea. Rev., 142, 39003919, doi:10.1175/MWR-D-13-00191.1.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., and R. W. Jacobson Jr., 1977: Diurnal variation of deep cumulus convection. Mon. Wea. Rev., 105, 11711188, doi:10.1175/1520-0493(1977)105<1171:DVODCC>2.0.CO;2.

    • 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, doi:10.1175/2011JAS3578.1.

    • Search Google Scholar
    • Export Citation
  • Hence, D. A., and R. A. Houze Jr., 2012: Vertical structure of tropical cyclone rainbands as seen by the TRMM Precipitation Radar. J. Atmos. Sci., 69, 26442661, doi:10.1175/JAS-D-11-0323.1.

    • Search Google Scholar
    • Export Citation
  • Hobgood, J. S., 1986: A possible mechanism for the diurnal oscillations of tropical cyclones. J. Atmos. Sci., 43, 29012922, doi:10.1175/1520-0469(1986)043<2901:APMFTD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Jiang, H., C. Liu, and E. J. Zipser, 2011: A TRMM-based tropical cyclone cloud and precipitation feature database. J. Appl. Meteor. Climatol., 50, 12551274, doi:10.1175/2011JAMC2662.1.

    • Search Google Scholar
    • Export Citation
  • Kossin, J. P., 2002: Daily hurricane variability inferred from GOES infrared imagery. Mon. Wea. Rev., 130, 22602270, doi:10.1175/1520-0493(2002)130<2260:DHVIFG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kozu, T., and Coauthors, 2001: Development of precipitation radar onboard the Tropical Rainfall Measuring Mission (TRMM) satellite. IEEE Trans. Geosci. Remote Sens., 39, 102116, doi:10.1109/36.898669.

    • 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, doi:10.1175/1520-0426(1998)015<0809:TTRMMT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., S. Ringerud, J. Crook, D. Randel, and W. Berg, 2011: An observationally generated a priori database for microwave rainfall retrievals. J. Atmos. Oceanic Technol., 28, 113130, doi:10.1175/2010JTECHA1468.1.

    • Search Google Scholar
    • Export Citation
  • Lajoie, F. A., and I. J. Butterworth, 1984: Oscillation of high-level cirrus and heavy precipitation around Australian region tropical cyclones. Mon. Wea. Rev., 112, 535544, doi:10.1175/1520-0493(1984)112<0535:OOHLCA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Landsea, C. W., and J. L. Franklin, 2013: Atlantic hurricane database uncertainty and presentation of a new database format. Mon. Wea. Rev., 141, 35763592, doi:10.1175/MWR-D-12-00254.1.

    • Search Google Scholar
    • Export Citation
  • Li, Q., and Y. Wang, 2012: Formation and quasi-periodic behavior of outer spiral rainbands in a numerically simulated tropical cyclone. J. Atmos. Sci., 69, 9971020, doi:10.1175/2011JAS3690.1.

    • Search Google Scholar
    • Export Citation
  • Liu, C., E. J. Zipser, D. J. Cecil, S. W. Nesbitt, and S. Sherwood, 2008: A cloud and precipitation feature database from nine years of TRMM observations. J. Appl. Meteor. Climatol., 47, 27122728, doi:10.1175/2008JAMC1890.1.

    • Search Google Scholar
    • Export Citation
  • Muramatsu, T., 1983: Diurnal variations of satellite-measured TBB areal distribution and eye diameter of mature typhoons. J. Meteor. Soc. Japan, 61, 7790.

    • Search Google Scholar
    • Export Citation
  • NASA, 2016: Precipitation Processing System Tropical Rainfall Measuring Mission: File specification for TRMM products, version 7.004. NASA Tech. Doc., 328 pp. [Available online at http://pps.gsfc.nasa.gov/Documents/filespec.TRMM.V7.pdf.]

  • Nesbitt, S. W., and E. J. Zipser, 2003: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements. J. Climate, 16, 14561475, doi:10.1175/1520-0442-16.10.1456.

    • Search Google Scholar
    • Export Citation
  • Nesbitt, S. W., E. J. Zipser, and D. J. Cecil, 2000: A census of precipitating features in the tropics using TRMM: Radar, ice scattering, and lightning observations. J. Climate, 13, 40874106, doi:10.1175/1520-0442(2000)013<4087:ACOPFI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Olander, T. L., and C. S. Velden, 2007: The advanced Dvorak technique: Continued development of an objective scheme to estimate tropical cyclone intensity using geostationary infrared satellite imagery. Wea. Forecasting, 22, 287298, doi:10.1175/WAF975.1.

    • Search Google Scholar
    • Export Citation
  • Panofsky, H. A., and G. W. Brier, 1958: Some Applications of Statistics to Meteorology. The Pennsylvania State University, 224 pp.

  • Simpson, J., R. F. Adler, and G. R. North, 1988: A proposed Tropical Rainfall Measuring Mission (TRMM) satellite. Bull. Amer. Meteor. Soc., 69, 278295, doi:10.1175/1520-0477(1988)069<0278:APTRMM>2.0.CO;2.

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

    • Search Google Scholar
    • Export Citation
  • Steranka, J., E. B. Rodgers, and R. C. Gentry, 1984: The diurnal variation of Atlantic Ocean tropical cyclone cloud distribution inferred from geostationary satellite infrared measurements. Mon. Wea. Rev., 112, 23382344, doi:10.1175/1520-0493(1984)112<2338:TDVOAO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wu, Q., Z. Ruan, D. Chen, and T. Lian, 2015: Diurnal variations of tropical cyclone precipitation in the inner and outer rainbands. J. Geophys. Res. Atmos., 120, 111, doi:10.1002/2014JD022190.

    • Search Google Scholar
    • Export Citation
  • Yang, G.-Y., and J. Slingo, 2001: The diurnal cycle in the tropics. Mon. Wea. Rev., 129, 784801, doi:10.1175/1520-0493(2001)129<0784:TDCITT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
Save