• Amitai, E., , Llort X. , , and Sempere-Torres D. , 2009: Comparison of TRMM radar rainfall estimates with NOAA next-generation QPE. J. Meteor. Soc. Japan, 87A, 109118, doi:10.2151/jmsj.87A.109.

    • Search Google Scholar
    • Export Citation
  • Anagnostou, M. S., , Morales C. A. , , and Dinku T. , 2001: The use of TRMM Precipitation Radar observations in determining ground radar calibration biases. J. Atmos. Oceanic Technol., 18, 616628, doi:10.1175/1520-0426(2001)018<0616:TUOTPR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bellon, A., , Lee G. , , and Zawadzki I. , 2005: Error statistics of VPR corrections in stratiform precipitation. J. Appl. Meteor., 44, 9981015, doi:10.1175/JAM2253.1.

    • Search Google Scholar
    • Export Citation
  • Bolen, S., , and Chandrasekar V. , 2000: Quantitative cross validation of space-based and ground-based radar observations. J. Appl. Meteor., 39, 20712079, doi:10.1175/1520-0450(2001)040<2071:QCVOSB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Cao, Q., , Hong Y. , , Qi Y. , , Wen Y. , , Zhang J. , , Gourley J. J. , , and Liao L. , 2013: Empirical conversion of vertical profile of reflectivity from Ku-band to S-band frequency. J. Geophys. Res. Atmos., 118, 18141825, doi:10.1002/jgrd.50138.

    • Search Google Scholar
    • Export Citation
  • Chen, S., and et al. , 2013: Evaluation of spatial errors of precipitation rates and types from TRMM spaceborne radar over the southern CONUS. J. Hydrometeor., 14, 18841896, doi:10.1175/JHM-D-13-027.1.

    • Search Google Scholar
    • Export Citation
  • Cho, Y.-H., , Lee G. W. , , Kim K.-E. , , and Zawadzki I. , 2006: Identification and removal of ground echoes and anomalous propagation using the characteristics of radar echoes. J. Atmos. Oceanic Technol., 23, 12061222, doi:10.1175/JTECH1913.1.

    • Search Google Scholar
    • Export Citation
  • Doviak, R. J., , and Zrnić D. S. , 1993: Doppler Radar and Weather Observations.2nd ed. Dover Publications, 562 pp.

  • Gabella, M., , Joss J. , , Perona G. , , and Michaelides S. , 2006: Range adjustment for ground-based radar, derived with the spaceborne TRMM Precipitation Radar. IEEE Trans. Geosci. Remote Sens., 44, 126133, doi:10.1109/TGRS.2005.858436.

    • Search Google Scholar
    • Export Citation
  • GAMIC, 2008: ENIGMA 3+ theory of operation. GAMIC mbH, 84 pp.

  • Heymsfield, G. M., , Geerts B. , , and Tian L. , 2000: TRMM Precipitation Radar reflectivity profiles as compared with high-resolution airborne and ground-based radar measurements. J. Appl. Meteor., 39, 20802102, doi:10.1175/1520-0450(2001)040<2080:TPRRPA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Iguchi, T., , Kozu T. , , Meneghini R. , , Awaka J. , , and Okamoto K. , 2000: Rain profiling algorithm for the TRMM Precipitation Radar. J. Appl. Meteor., 39, 20382052, doi:10.1175/1520-0450(2001)040<2038:RPAFTT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Iguchi, T., , Kozu T. , , Kwiatkowski J. , , Meneghini R. , , Awaka J. , , and Okamoto K. , 2009: Uncertainties in the rain profiling algorithm for the TRMM Precipitation Radar. J. Meteor. Soc. Japan, 87A, 130, doi:10.2151/jmsj.87A.1.

    • Search Google Scholar
    • Export Citation
  • Kirstetter, P. E., and et al. , 2012: Toward a framework for systematic error modeling of spaceborne radar with NOAA/NSSL ground radar–based National Mosaic QPE. J. Hydrometeor., 13, 12851300, doi:10.1175/JHM-D-11-0139.1.

    • Search Google Scholar
    • Export Citation
  • Kirstetter, P. E., , Andrieu H. , , Delrieu G. , , and Boudevillain B. , 2013: A physically based identification of vertical profiles of reflectivity from volume scan radar data. J. Appl Meteor. Climatol., 52, 16451663, doi:10.1175/JAMC-D-12-0228.1.

    • Search Google Scholar
    • Export Citation
  • Langston, C., , Zhang J. , , and Howard K. , 2007: Four-dimensional dynamic radar mosaic. J. Atmos. Oceanic Technol., 24, 776790, doi:10.1175/JTECH2001.1.

    • Search Google Scholar
    • Export Citation
  • Lee, G. W., , and Zawadzki I. , 2006: Radar calibration by gage, disdrometer, and polarimetry: Theoretical limit caused by the variability of drop size distribution and application to fast scanning operational radar data. J. Hydrol., 328, 8397, doi:10.1016/j.jhydrol.2005.11.046.

    • Search Google Scholar
    • Export Citation
  • Liao, L., , and Meneghini R. , 2009: Validation of TRMM Precipitation Radar through comparison of its multiyear measurements with ground-based radar. J. Appl. Meteor. Climatol., 48, 804817, doi:10.1175/2008JAMC1974.1.

    • Search Google Scholar
    • Export Citation
  • Liao, L., , Meneghini R. , , and Iguchi T. , 2001: Comparisons of rain rate and reflectivity factor derived from the TRMM Precipitation Radar and the WSR-88D over the Melbourne, Florida, site. J. Atmos. Oceanic Technol., 18, 19591974, doi:10.1175/1520-0426(2001)018<1959:CORRAR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Meneghini, R., , Jones J. A. , , Iguchi T. , , Okamoto K. , , and Kwiatkowski J. , 2004: A hybrid surface reference technique and its application to the TRMM Precipitation Radar. J. Atmos. Oceanic Technol., 21, 16451658, doi:10.1175/JTECH1664.1.

    • Search Google Scholar
    • Export Citation
  • Park, S.-G., , and Lee G. , 2010: Calibration of radar reflectivity measurements form KMA operational radar network. Asia–Pac. J. Atmos. Sci., 46, 243259, doi:10.1007/s13143-010-1010-3.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and Houze R. A. Jr., 2000: Comparison of radar data from the TRMM satellite and Kwajalein ocean validation site. J. Appl. Meteor., 39, 21512164, doi:10.1175/1520-0450(2001)040<2151:CORDFT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schwaller, M. R., , and Morris K. R. , 2011: A ground validation network for the global precipitation measurement mission. J. Atmos. Oceanic Technol., 28, 301319, doi:10.1175/2010JTECHA1403.1.

    • Search Google Scholar
    • Export Citation
  • Siggia, A., , and Passarelli R. , 2004: Gaussian model adaptive processing (GMAP) for improved ground clutter cancellation and moment calculation. Proceedings of ERAD (2004), Copernicus GmBH, 67–73. [Available online at http://copernicus.org/erad/2004/online/ERAD04_P_67.pdf.]

  • TRMM Precipitation Radar Team, 2011: Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar algorithm: Instruction manual for version 7. JAXA/NASA, 170 pp. [Available online at www.eorc.jaxa.jp/TRMM/documents/PR_algorithm_product_information/pr_manual/PR_Instruction_Manual_V7_L1.pdf.]

  • Wang, J., , and Wolff D. , 2009: Comparisons of reflectivities from the TRMM Precipitation Radar and ground-based radars. J. Atmos. Oceanic Technol., 26, 857875, doi:10.1175/2008JTECHA1175.1.

    • Search Google Scholar
    • Export Citation
  • Wen, Y., , Hong Y. , , Zhang G. , , Schuur T. J. , , Gourley J. J. , , Flamig Z. , , Morris K. R. , , and Cao O. , 2011: Cross validation of spaceborne radar and ground polarimetric radar aided by polarimetric echo classification of hydrometeor types. J. Appl Meteor. Climatol., 50, 13891402, doi:10.1175/2011JAMC2622.1.

    • Search Google Scholar
    • Export Citation
  • Yuter, S. E., , and Houze R. A. 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, doi:10.1175/1520-0493(1995)123<1941:TDKAME>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., , Howard K. , , and Gourley J. J. , 2005: Three-dimensional multiple radar reflectivity mosaic. J. Atmos. Oceanic Technol., 22, 3042, doi:10.1175/JTECH-1689.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Analysis domain of the S-band radars operated by the Korea Meteorological Administration (KMA; triangle) and by the Ministry of Land, Infrastructure and Transport (MOLIT; square). The terrain height map is generated with data released from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model, version 2 (GDEM V2), supported by NASA.

  • View in gallery

    Reflectivity fields at 2.5-km height for four widespread rain events on 15, 18, and 19 Jun and 14 Jul 2012 from (left) PR observations and (right) GR composite with GR coverage of 150 km.

  • View in gallery

    As in Fig. 2, but for three convective events on 30 Jun and 10 and 13 Jul 2012 and one typhoon event on 18 Jul 2012.

  • View in gallery

    Vertical cross sections of reflectivity observed from (left) PR observations and (right) GR composite for the widespread rain cases, following the center of the PR swath shown as dotted lines in Fig. 2 that corresponds to the dashed line at 2.5-km height in these graphs.

  • View in gallery

    As in Fig. 4, but for three convective events and one typhoon.

  • View in gallery

    CFADs for reflectivity measurements from (left) PR observations and (right) GR composite for the widespread rain events. Black dashed lines in the vertical indicate 0 and 18 dBZ.

  • View in gallery

    As in Fig. 6, but for three convective events and one typhoon.

  • View in gallery

    Mean reflectivity profiles for (left) widespread and (right) convective cases with the radar coverage up to 60 km. The mean at each level is computed with reflectivity larger than 18 dBZ [(PR and GR) ≥ 18 dBZ; PR in red and GR in black circles with dashed lines] and larger than 0 dBZ [(PR and GR) ≥ 0 dBZ, where PR and GR values are set for 0 dBZ if they are smaller than 18 dBZ; PR in violet and GR in blue]. The size of circles is proportional to the number of valid points. The dotted lines indicate the percentiles of 10% and 90% around the mean larger than 18 dBZ.

  • View in gallery

    As in Fig. 8, but for radar coverage up to 150 km.

  • View in gallery

    Scatterplots (black) of reflectivity observed at 2.5-km height by PR and GR for (left) widespread and (right) convective cases. The matched points with reflectivity larger than 18 dBZ are in blue and are used to compute BIAS, MAD, RMSD, and CORR.

  • View in gallery

    Point-to-point statistics in the vertical for the widespread cases. The size of circles is proportional to the number of points. (a) CORR, (b) MAD, and (c) BIAS (red) together with RMSD (black) computed with reflectivity pixels exceeding 18 dBZ.

  • View in gallery

    As in Fig. 11, but for three convective cases and one typhoon case.

  • View in gallery

    RMSD values between PR and GR reflectivities (larger than 18 dBZ) in the vertical for a (a) widespread and (b) convective case. The black line is computed with the PR–GR matched at the shortest time. The green line is done with the second-closest time matching. The circle size is proportional to the number of points used in the computation.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 65 65 9
PDF Downloads 28 28 5

Cross Validation of TRMM PR Reflectivity Profiles Using 3D Reflectivity Composite from the Ground-Based Radar Network over the Korean Peninsula

View More View Less
  • 1 Research and Training Team for Future Creative Astrophysicists and Cosmologists, Center for Atmospheric Remote Sensing, Department of Astronomy and Atmospheric Sciences, Kyungpook National University, Daegu, South Korea
© Get Permissions
Full access

Abstract

The Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) measures reflectivity downward from space and provides observations of the vertical distributions of precipitation over land as well as the ocean. It overpasses the southern part of the Korean Peninsula where (i) a dense network of operational S-band scanning radars is available and (ii) various types of precipitation occur. By utilizing a 3D reflectivity composite from the ground S-band radar (GR) observations, this paper shows a comparison of reflectivity profiles observed with both PR and GR focusing on their vertical structure. For four cases of widespread rain, visual and statistical analyses show that PR attenuation-corrected reflectivity agrees closely with reflectivity observed from the GR composite below the melting layer. Above and within the melting layer, PR is affected critically by its sensitivity while GR beam broadening at far ranges causes systematic differences in the PR–GR comparisons. For four cases of convective rain, PR underestimates the mean reflectivities by 1–3 dB compared with those from GR at low levels where precipitation attenuation is significant toward the ground. In these cases, the low sensitivity of PR results in a small number of matched points for weak echoes. Also, the PR–GR discrepancy for the convective case is more affected by time mismatching.

Corresponding author address: GyuWon Lee, Department of Astronomy and Atmospheric Sciences, Kyungpook National University, 80 Daehakro, Bukgu, Daegu 702-701, South Korea.E-mail: gyuwon@knu.ac.kr

Abstract

The Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) measures reflectivity downward from space and provides observations of the vertical distributions of precipitation over land as well as the ocean. It overpasses the southern part of the Korean Peninsula where (i) a dense network of operational S-band scanning radars is available and (ii) various types of precipitation occur. By utilizing a 3D reflectivity composite from the ground S-band radar (GR) observations, this paper shows a comparison of reflectivity profiles observed with both PR and GR focusing on their vertical structure. For four cases of widespread rain, visual and statistical analyses show that PR attenuation-corrected reflectivity agrees closely with reflectivity observed from the GR composite below the melting layer. Above and within the melting layer, PR is affected critically by its sensitivity while GR beam broadening at far ranges causes systematic differences in the PR–GR comparisons. For four cases of convective rain, PR underestimates the mean reflectivities by 1–3 dB compared with those from GR at low levels where precipitation attenuation is significant toward the ground. In these cases, the low sensitivity of PR results in a small number of matched points for weak echoes. Also, the PR–GR discrepancy for the convective case is more affected by time mismatching.

Corresponding author address: GyuWon Lee, Department of Astronomy and Atmospheric Sciences, Kyungpook National University, 80 Daehakro, Bukgu, Daegu 702-701, South Korea.E-mail: gyuwon@knu.ac.kr

1. Introduction

The global knowledge on the vertical distribution of precipitation is essential for hydrometeorological monitoring and forecasting as well as evaluating climatological model simulations. Although the true distribution is not exactly known, both spaceborne and ground-based radars provide reflectivity measurements of precipitation. The Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) orbiting around the tropics and subtropics has provided reflectivity observations since 1997 (e.g., Iguchi et al. 2000). To assess the quality of PR’s estimates, numerous researchers (e.g., Schumacher and Houze 2000; Liao et al. 2001; Schwaller and Morris 2011) have performed cross validations of TRMM PR reflectivity with independent volumetric reflectivity measurements from ground-based scanning radars (GRs). The discrepancies between PR and GR measurements were explained in part by (i) ground radar calibration issues in reflectivity measurements (e.g., Anagnostou et al. 2001; Wang and Wolff 2009), (ii) PR’s limited sensitivity (e.g., Heymsfield et al. 2000), (iii) scattering cross section at different radar frequencies associated with hydrometeor distribution and different viewing geometries (e.g., Bolen and Chandrasekar 2000; Liao and Meneghini 2009; Wen et al. 2011), and (iv) PR’s attenuation correction algorithms that still underestimate convective rain (e.g., Liao and Meneghini 2009; Schwaller and Morris 2011). Over the United States, PR rain estimates have also been compared with those from the GR composite generated with national networks of ground-based scanning radars over enlarged coverage. For example, Amitai et al. (2009) showed that the contribution of intense rain observed from PR to the total amount of rain is less than that from the GR composite. Kirstetter et al. (2012) performed spatial structure analysis of surface rainfall, explaining the PR–GR discrepancy is possibly due to PR’s lack of sampling the variability of small, disorganized convective rainfall structures together with inaccurate conversion from reflectivity to rainfall intensity, and sources (ii)–(iv) mentioned above.

The PR overpasses the southern part of the Korean Peninsula (32°–36°N latitude). This area features complex mountainous terrain surrounded by the sea and frequent occurrence of various types of precipitation, including typhoons. Also, a dense network of operational ground-based S-band radars is available in the area, which can be used for ground validation of PR observations on board both the TRMM system and the Global Precipitation Measurement (GPM) satellite that was recently launched.

The objective of this paper is to compare TRMM PR reflectivity with GR reflectivity composites over the Korean Peninsula, focusing on the vertical. The vertical profiles of reflectivity (VPRs) are the result of the microphysical processes in precipitation and characterize the storm structures that are important to determine rainfall estimates near the surface. However, both PR and GR observations have their own limitations to estimate VPR, for example, errors in attenuation correction for PR (e.g., Iguchi et al. 2000) and beam broadening effect for GR (e.g., Bellon et al. 2005). By combining information from multiple ground-based radars, it is expected not only to enlarge the analysis domain but to mitigate some inherent weaknesses of volume scan data, for example, (i) the signal loss due to beam blockage, (ii) the lack of data in the cone of silence or in the low level below the radar horizon due to the curvature of the Earth, and (iii) beam broadening effects at far range. However, because of resolution differences between PR and GR measurements, particularly in the vertical, the VPR comparisons are not straightforward. Therefore, this work aims to illustrate some of the known problems in the comparison between PR and GR measurements. This allows us to understand their limitations and to better focus on the cause of differences in the comparison depending on precipitation types and measurement heights. We first address the differences in data and matching scheme (section 2). The comparison results are presented and analyzed in section 3 for selected cases of different precipitation types. The summary and conclusion are in section 4.

2. Data

a. Reflectivity factor from the TRMM PR 2A25 product

The TRMM PR measures the reflectivity factor at Ku band (13.8 GHz and approximately 2.2-cm wavelength) from 402 km above the earth’s surface. Its downward measurements at such a short wavelength are seriously affected by atmospheric attenuation due to precipitation, cloud liquid water, water vapor, and molecular oxygen. The 2A25 is one of the TRMM PR orbital products providing 3D profiles of attenuation-corrected effective reflectivity factor (dBZ) based on other products of TRMM PR: received power calibration (1B21), measured radar reflectivity (1C21), microphysics such as rain type and bright band (2A23), and rain attenuation correction (2A21). In this work, we have used the latest version of 2A25 (version 7). The attenuation corrections are based on the estimation of two-way path-integrated attenuation with several assumptions using both (i) the latest hybrid approach of the surface reference technique (SRT) and (ii) the Hitschfeld–Bordan (HB) solution (Meneghini et al. 2004; Iguchi et al. 2009); that is, the SRT estimates attenuation by subtracting the apparent surface cross section measured in rain from the surface cross section measured outside rain as a reference. The HB method initially assumes a relationship between specific attenuation and the effective radar reflectivity factor. By matching the SRT attenuation estimates, the HB solution is adjusted with phase state of hydrometeors, temperature, and a model for the drop size distribution.

A PR scan is composed of 49 sample beams within the cross-track swath of 245 km. The horizontal resolution (footprints of the beam) is approximately 5 km at the scan edges (±17° from the center of the path) in the lower atmosphere, and the vertical resolution is 250 m. The PR’s minimum detectable signal is limited approximately above 17–18 dBZ (TRMM Precipitation Radar Team 2011). This sensitivity limitation implies that the measured returns come from rain, melting snow, and ground clutter and hardly from snow (only from intense snow storms).

b. Reflectivity factor from the composite of GR network

Conventional ground-based scanning S-band radars observe the reflectivity factor at certain elevation angles along the earth’s surface. At this wavelength (about 10 cm), the effect of path attenuation by precipitation is generally negligible. The southern part of the Korean Peninsula is covered by six S-band radars (Fig. 1) located at Gosan (GSN; 103 m MSL), Jindo (JNI; 499 m MSL), Osungsan (KSN; 231 m MSL), Guducksan (PSN; 547 m MSL), Sungsan (SSP; 72 m MSL), and Biseulsan (BSL; 1085 m MSL). The first five are Doppler radars that transmit a horizontally polarized beam with a pulse length of 1.0 μs using a Klystron transmitter with a peak power of 750 kW. The 3-dB beamwidth is 1.0°, and the resolution of the range bin is 250 m. The volumetric scan is composed of 12–15 elevation angles (0.0°–24.0°) with horizontal coverage up to 240 km and available every 10 min. The BSL dual-polarization radar uses 45° slant-polarized transmission and simultaneous dual-channel reception (both horizontally and vertically). It scans only lower elevation angles (−0.5°, 0.0°, 0.8°, 1.2°, and 1.6°), providing rapidly updated reflectivity observations every 2.5 min up to 150 km in range with a bin resolution of 125 m.

Fig. 1.
Fig. 1.

Analysis domain of the S-band radars operated by the Korea Meteorological Administration (KMA; triangle) and by the Ministry of Land, Infrastructure and Transport (MOLIT; square). The terrain height map is generated with data released from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model, version 2 (GDEM V2), supported by NASA.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

The generation of the composites requires several steps to assure data quality. First, we have used the input reflectivity data that are filtered for nonmeteorological targets (e.g., ground clutter) by the radar provider’s commercial signal processor. Although the processors are different among the radars, the method is mostly based on the infinite impulse response (IIR; GAMIC 2008) or Gaussian model adaptive processing (GMAP; Siggia and Passarelli 2004). However, this may still leave some residual nonprecipitating echoes such as chaff, anomalous propagation, interferences, or strong second-trip echoes. Such echoes can be further corrected, in particular at low levels with fuzzy logic algorithms (e.g., Cho et al. 2006) or radar reliable index product with rain gauge measurements (e.g., Kirstetter et al. 2012), which is beyond the scope of this work. Instead, we chose the less contaminated cases.

Second, it is important to estimate a system gain bias among different radars because reflectivity is affected by each radar characteristic such as transmission, antenna gain, receiver gain, loss factors in the waveguide, and rotary joints (Doviak and Zrnić 1993). In general, a well-calibrated system allows a system bias within −1 and +1 dB among the operational radars (e.g., Park and Lee 2010). For the cases used in this work, we have calibrated the systematic bias as shown in Table 1 by (i) obtaining a reference calibration from the BSL radar based on self-consistency of dual-polarization measurements (Lee and Zawadzki 2006) over seven intense rainy days (5, 6, 11, 13, 14, 15, and 17 July 2012) and (ii) comparing BSL data with those from the neighboring radars at the closest time and averaging within rainy days in June–August 2012. If the radar is too far from the BSL radar, the comparison is done with the nearby radar that is already calibrated with the BSL radar, as similarly done by Park and Lee (2010).

Table 1.

Summary of the calibrated biases of GR reflectivity.

Table 1.

To make the composite, we have converted the individual radar data (originally in spherical coordinates) onto a 3D Cartesian grid according to Langston et al. (2007) by taking into account (i) the beam geometry (the radar sampling volume increases with the square of the range) and propagation assuming a standard atmosphere, (ii) along-ground distance computed based on the great circle principle, and (iii) conversion into geographical coordinates and projection onto a Lambert conformal grid. The common Cartesian grid spacing is set to 5 km in the horizontal and 0.25 km in the vertical to be similar to the PR footprint size. We have used observations up to a maximum range of 150 km from the individual radars. The composites have been generated by simply averaging the reflectivity values of those radar bins whose center is located within each volumetric grid point of 5 km × 5 km × 0.25 km. Last, reflectivity values in the overlapping areas by multiple radars are also averaged with equal weights for each GR. Consequently, there are points where no data are available: (i) between far-spaced elevation angles (particularly at high elevations, the density of GR data in the vertical is lower than PR), (ii) below the lowest elevation angles, and (iii) at the cone of silence. In this way, we guarantee the quality of the data and avoid further interpolation or filtering (e.g., Zhang et al. 2005).

c. Matching PR and GR observations

The matching between the PR and GR data is not straightforward because of their differences in temporal and spatial resolutions and viewing geometry (compared in Table 2). While GR scans every 5–6 min (composite every 10 min) at fixed elevation angles, PR scans along and cross track of the spacecraft orbit (about 7 km s−1). It typically takes 1–2 min to cross from the west to the east boundaries of the analysis domain (Fig. 1). Although PR passes over the Korean Peninsula three or four times a day, data availability for the comparison also depends on both the availability of the GR composite and the presence of rainy events over the area. We have chosen eight instantaneous cases for the PR–GR comparisons. Table 3 shows the events selected during the East Asian summer monsoon period in June–July 2012. During such periods, significant rainfall often occurs along a quasi-stationary front moving south to north (and vice versa) over the Korean Peninsula. The selected GR composite (available every 10 min) is chosen to be the closest to the PR time. In this way, the time difference between the PR and GR observations is shorter than 5 min.

Table 2.

A comparison of different characteristics between TRMM PR and GR composite observations.

Table 2.
Table 3.

Selected widespread and convective cases for reflectivity comparisons between TRMM PR and GR composite. Local time is indicated as Korean standard time (KST; UTC + 9), and the PR overpasses within 2 min over the domain shown in Fig. 1. GR composites are available every 10 min and are selected at the closest time to either PR start or end time.

Table 3.

We have subjectively classified the eight cases into two categories; that is, widespread and convective rain according to the dominant type of precipitation based on visual inspection of the 3D reflectivity fields. The former actually includes a mesoscale convective system (MCS) containing both convective and stratiform precipitating regions. The latter includes a typhoon case. In this way, we can compare reflectivity profiles estimated from both systems event by event [alternatively, some studies (e.g., Kirstetter et al. 2012; Chen et al. 2013) have used the PR rain type product in the context of QPE assessment].

To match both measurements in space, some researchers have proposed to select the volumetric data measured at the geometric intersection of the PR rays and the individual GR elevation sweeps (e.g., Heymsfield et al. 2000; Bolen and Chandrasekar 2000; Schwaller and Morris 2011). Others have interpolated the PR and GR data onto a fixed three-dimensional common grid (e.g., Liao et al. 2001; Amitai et al. 2009). We chose the latter for the statistical comparison, using the GR grid as the reference and remapping the PR data onto it by linear interpolation.

3. PR–GR comparison

a. Visual comparisons

Four widespread events are presented in Fig. 2 showing both the PR horizontal reflectivity fields at 2.5 km in height (Fig. 2, left) and the corresponding GR reflectivity composites over a common grid (Fig. 2, right). The ring shapes in both PR and GR reflectivity maps appear because the lack of coverage in the GR composite is excluded from the common domain. Similarly, three convective events and a typhoon case are presented in Fig. 3, showing rainfall over a relatively small area compared to the widespread cases. In general, the reflectivity patterns above 18 dBZ show clear correspondence between PR and GR measurements for both widespread and convective cases at this height. However, small-scale variations in both intensity and structures of GR measurements appear differently from those of PR measurements. Also, it needs to be noted that GR observations do show the presence of weak precipitation below the PR sensitivity (18 dBZ); for example, the areas below the PR minimum detectable threshold (provided as 0 dBZ in the 2A25 product) are plotted in medium blue (between −0.1 and 0.1 dBZ in Figs. 25). From the perspective of GR, the values below PR minimum detectable threshold correspond to echo-free regions (gray), clear-air echoes (light blue), and weak echoes (dark blue).

Fig. 2.
Fig. 2.

Reflectivity fields at 2.5-km height for four widespread rain events on 15, 18, and 19 Jun and 14 Jul 2012 from (left) PR observations and (right) GR composite with GR coverage of 150 km.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Fig. 3.
Fig. 3.

As in Fig. 2, but for three convective events on 30 Jun and 10 and 13 Jul 2012 and one typhoon event on 18 Jul 2012.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Fig. 4.
Fig. 4.

Vertical cross sections of reflectivity observed from (left) PR observations and (right) GR composite for the widespread rain cases, following the center of the PR swath shown as dotted lines in Fig. 2 that corresponds to the dashed line at 2.5-km height in these graphs.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Fig. 5.
Fig. 5.

As in Fig. 4, but for three convective events and one typhoon.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

To see the reflectivity differences in the vertical, cross-sectional views for PR and GR with height are presented in Figs. 4 and 5 for both widespread and convective cases, respectively. The x axes follow the center of the PR track shown in Figs. 2 and 3 at nadir (where the PR data are less contaminated by clutter and have the best vertical resolution). The coverage of the GR composite becomes small at this low level because of the upward propagation of radar beam with respect to the earth’s surface in standard atmospheric conditions, as shown in Figs. 4 and 5. We can also see remaining ground clutter contamination in the GR composite near the surface. This clutter appears at low levels, possibly associated with moving ships (e.g., see 50–60 dBZ between 129°–130°E on 30 June and 10 and 14 July).

In the widespread rain events (Fig. 4), the GR composites show wider bright bands (the peaks appear between 4 and 5 km in height) than PR observations, which suggest the effect of GR beam broadening, in particular at far ranges, as mentioned in section 2b. PR reflectivity values are also smaller than those from GR at localized areas (e.g., between 2 and 4 km in height at 124°–127°E on 15 and 18 June and at 126°–126.5°E on 14 July). These discrepancies appear possibly because of the weakness of the PR attenuation-correction algorithm that may underestimate reflectivity toward low levels. The impact of nonuniform beam filling in the SRT might produce negative biases with large gradients across the beam as those in convective storms (Iguchi et al. 2009). The temporal differences between PR and GR measurements can also be significant and can result in clear differences in location and evolution of small-scale storms. This effect is more evident in the convective cases shown in Fig. 5. At high levels, the overlapping areas between two measurements become relatively small because of the GR’s beam pattern with height and the PR’s limitation to measure low reflectivities in the solid precipitation areas.

b. Comparisons of mean vertical profiles

Contoured frequency by altitude diagrams (CFADs; e.g., Yuter and Houze 1995) are plotted for widespread (Fig. 6) and convective (Fig. 7) rain cases. The graphs show the distribution of volumetric reflectivity observations as a function of height; that is, at a given height, the number of points in a reflectivity interval (5 dBZ) is normalized by the total number of points. Here, the analysis has been done within the PR–GR common domain, taking the GR coverage up to 60 km in range to minimize the effect of GR beam broadening.

Fig. 6.
Fig. 6.

CFADs for reflectivity measurements from (left) PR observations and (right) GR composite for the widespread rain events. Black dashed lines in the vertical indicate 0 and 18 dBZ.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Fig. 7.
Fig. 7.

As in Fig. 6, but for three convective events and one typhoon.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Regardless of precipitation types, PR CFADs (Figs. 6 and 7, left) clearly show a high percentage of 0 dBZ and few detections below 18 dBZ, whereas GR can observe weak echoes missed by PR. The CFAD values at the lowest GR reflectivity indicate the precipitation-free points (i.e., clear air) within the radar coverage.

For the cases of widespread rain (Fig. 6), the distributions of PR and GR reflectivities larger than 18 dBZ show similar vertical patterns below 6 km in height; for example, the frequency of high reflectivity increases up to the brightband peak (for a fixed frequency level, the increase in reflectivity is on the order of 5–10 dB, depending on the case). For solid precipitation (above the bright band), besides the effect of PR sensitivity limitations to measure reflectivities below 18 dBZ depicted from GR measurements, PR seems to underestimate the frequencies of reflectivities above 18 dBZ (particularly for the cases of 15 June and 14 July).

For the selected convective cases, the area of precipitation is smaller than that for the widespread cases. This is reflected in the CFADs (Fig. 7) with lower frequencies of the reflectivities above the sensitivity thresholds. Among the convective cases, the CFAD structures are more variable than those obtained for the widespread cases; for example, for the cases of 30 June and 10 July, the precipitating system (above 18 dBZ) reaches up to 9 km, but no signature of the bright band is shown. On the case of 13 July, PR observes less frequently both strong echoes (above 30 dBZ between 2 and 6 km in height) and moderate echoes (18–30 dBZ below 3 km in height). On the case of 18 July, the typhoon shows reflectivity distributions reaching up to 10 km for both GR and PR with no obvious signature of a bright band and with relatively uniform structure in the vertical. The maximum reflectivity is observed between 2 and 3 km in height and is larger for GR than PR. In general, PR observations show lower frequencies of strong convective echoes compared with those from GR. Such discrepancy is possibly due to the impact of nonuniform beam filling on PR estimates of convective echoes, which affects not only errors in the estimate of path attenuation but also the retrieval of microphysical parameters regarding properties (phase state, shape, nonuniform distributions, etc.) of precipitation particles (Iguchi et al. 2009).

Based on the CFADs of Figs. 6 and 7, the mean reflectivity profiles of both PR and GR measurements have also been computed at each level with more than 20 valid points. The mean can be computed by averaging reflectivity arithmetically either on the log scale (dBZ) or in linear units, though the latter weighs more toward the strong echoes. For this work, we have opted for the former (note that the differences in the PR–GR comparisons due to such unit conversion were not found significant in these eight cases). Figure 8 shows first the mean profiles by averaging only reflectivity values above the PR sensitivity threshold [(PR and GR) ≥ 18 dBZ]. For the widespread cases (Fig. 8, left), the mean profiles between PR (red) and GR (black) agree up to 4 km in height. Around and above the bright bands, PR mean values are smaller than those from GR. The heights of the brightband peaks do not always agree [e.g., slightly lower (~250 m) on the PR than the GR for both 15 and 18 June 2012]. For the convective cases (Fig. 8, right), GR mean values up to 4 km are 1–3 dB larger than those of PR and the PR–GR differences in height are variable among cases with a small number of matching points.

Fig. 8.
Fig. 8.

Mean reflectivity profiles for (left) widespread and (right) convective cases with the radar coverage up to 60 km. The mean at each level is computed with reflectivity larger than 18 dBZ [(PR and GR) ≥ 18 dBZ; PR in red and GR in black circles with dashed lines] and larger than 0 dBZ [(PR and GR) ≥ 0 dBZ, where PR and GR values are set for 0 dBZ if they are smaller than 18 dBZ; PR in violet and GR in blue]. The size of circles is proportional to the number of valid points. The dotted lines indicate the percentiles of 10% and 90% around the mean larger than 18 dBZ.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Although GR measures precipitation and some clear echoes with reflectivities below 18 dBZ, we have set those to 0 dBZ values to compute the mean profile above the GR sensitivity threshold [(PR and GR) ≥ 0 dBZ]. This allows us to include the values with weak reflectivity echoes (as detected by GR) in the analysis of the mean VPR of both instruments. For the widespread cases (Fig. 8, left; PR mean in violet and GR mean in blue), the shapes of the mean reflectivity profile resemble those computed with reflectivity above 18 dBZ (e.g., the height of the brightband peak). On the other hand, for the convective cases (Fig. 8, right), the reflectivity discrepancies between PR and GR are quite different from those of the widespread and increase downward between 1.5 and 4 km in height. This growing tendency that is less captured by VPR computed with reflectivity above 18 dBZ possibly indicates in part that the attenuation correction is not sufficient for the convective cases.

Similarly, Fig. 9 presents the results obtained with GR observations up to 150 km in range. Increasing the size of the domain, the mean profile extends to a higher level, particularly for the convective cases. However, the discrepancies between PR and GR observations seem to be larger; for example, above 4–5 km in height, the PR–GR reflectivity differences are systematically larger at a large domain than those obtained at a small domain (cf. Figs. 8 and 9). Also, for the widespread cases, the thickness of bright bands becomes broader, and the brightband peaks are not as clear as those obtained for the smaller domain. Both PR and GR mean reflectivity values are reduced below the bright bands because of oversmoothing at the larger domain. This illustrates the effect of GR’s beam broadening mentioned in the previous sections. For the convective cases, the PR–GR differences (GR 1–3 dB larger than PR below 4–5 km in height) seem to be less dependent on the domain size. To reduce such effects in the interpretation of the PR–GR comparisons [especially around the melting layer in stratiform cases; see the complete analysis of Bellon et al. (2005)], the rest of the analyses are done over GR coverage up to 60 km. Although it needs to be pointed out that the 3-dB beamwidth of GR at 60 km is on the order of 1 km (i.e., the real vertical resolution of GR observations is still coarser than that of PR), we have chosen this maximum range as a compromise between vertical resolution of GR observations and the number of points available for the PR–GR comparison.

Fig. 9.
Fig. 9.

As in Fig. 8, but for radar coverage up to 150 km.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

c. Point-to-point comparisons

To better understand the PR–GR differences shown in the mean reflectivity profiles, we have compared concurrent PR and GR reflectivity observations point by point. Figure 10 shows an example of the scatterplots at a given height, 2.5 km, for the widespread (Fig. 10, left) and convective (Fig. 10, right) cases. There are a large number of points that are below the minimum detectable values of PR that are above the minimum detectable value of GR measurements (see black points where PR indicates 0 dBZ). The blue points indicate the PR–GR pairs for reflectivity larger than 18 dBZ and are used to compute some PR–GR statistics [Pearson correlation coefficient (CORR), mean absolute difference (MAD), root-mean-square difference (RMSD), and bias (BIAS); dBZ] with the following equations:
e1
e2
e3
and
e4
where angle brackets indicate the mean reflectivity for n points. For the widespread cases (Fig. 10, left), the biases at 2.5 km are relatively small compared with those for the convective cases (Fig. 10, right). The small and nonsystematic BIAS supports successful calibration of GRs. GR reflectivities are higher than those of PR, particularly for strong echoes (above 35 dBZ) for most convective cases, which is possible because of more attenuated PR reflectivity at such a low level (e.g., Heymsfield et al. 2000). The correlation coefficient is larger for the widespread cases, indicating that both PR and GR estimates are similar and less variable in the horizontal (e.g., less affected by location mismatching or nonuniform beam filling) than those for the convective cases. Among the convective cases, the typhoon case shows the smallest correlation values between the PR and the GR estimates at 2.5 km.
Fig. 10.
Fig. 10.

Scatterplots (black) of reflectivity observed at 2.5-km height by PR and GR for (left) widespread and (right) convective cases. The matched points with reflectivity larger than 18 dBZ are in blue and are used to compute BIAS, MAD, RMSD, and CORR.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

The above point-to-point statistics are also presented with different heights for widespread (Fig. 11) and convective (Fig. 12) rain cases. The results are summarized as follows.

  • Near surface (below 1.5km): In general, PR estimations appear larger than those from GR composite, though both PR and GR estimates suffer from ground clutter contamination, and thus the number of matched points is small.
  • Between 1.5 and 4km (rain region): For the widespread cases, high correlation coefficients (between 0.75 and 0.9; shown in Fig. 11a) and small MAD (less than 4 dB; shown in Fig. 11b) confirm the close similarity between PR and GR. This is not so clear for the convective cases (Figs. 12a,b), where the number of matched pairs is also much smaller. The BIAS values for the widespread cases are relatively small (see Fig. 11c) with slight overestimation by PR. These small and positive BIAS values (below 1 dB) are possibly because of (i) scattering differences for some large raindrops (Mie scattering at Ku band, Rayleigh scattering at S band) and (ii) different viewing geometry for the raindrops (nadir view by PR; slantwise horizontal view by GR). These results are consistent with PR–GR comparisons obtained from reflectivity simulations at echoes below 40 dBZ done by several authors (Liao and Meneghini 2009; Wen et al. 2011; Cao et al. 2013). On the other hand, for the convective cases, the biases are negative (Fig. 12c), suggesting that the backscattered power at PR Ku band is weaker than that of GR S band because of path attenuation. The convective cases show larger RMSDs than those obtained from the widespread cases. This is possible because of the small-scale variability of the precipitation fields mentioned earlier and time mismatching.
  • Between 4 and 6km (within the bright bands): The MAD, BIAS, and RMSD values clearly appear large at this layer for widespread cases (Fig. 11). Such discrepancies can be from both a mismatch in the brightband height and backscattering differences between PR and GR for complex mixed-phase precipitation in the melting layer. For the convective cases, these features are less evident because of the small sample size and weak or no bright band.
  • Above 6km (solid precipitation region): Compared to the lower levels, the number of matched points is small because of both the sensitivity limitations of PR and the poor coverage of GR (as shown in CFADs). For the widespread cases, CORR values become poor and the BIAS values are negative (1–3 dB larger by GR than by PR). At those heights, the type of hydrometeors would be mostly ice particles (dry snow), and the effect of attenuation on PR is negligible (e.g., Liao and Meneghini 2009; Cao et al. 2013). Therefore, the reason for such differences could be attributed to other factors such as scattering differences between PR and GR. For example, Liao and Meneghini (2009) and Wen et al. (2011) showed that PR reflectivity values were up to 5 dB smaller than those of GR in dry snow regions based on the simulations of the backscattering cross sections. Unlike in their comparisons (where GR observations were found to be smaller than PR observations), the negative BIAS shown in Fig. 11 is reasonably consistent with their simulation results.
Fig. 11.
Fig. 11.

Point-to-point statistics in the vertical for the widespread cases. The size of circles is proportional to the number of points. (a) CORR, (b) MAD, and (c) BIAS (red) together with RMSD (black) computed with reflectivity pixels exceeding 18 dBZ.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Fig. 12.
Fig. 12.

As in Fig. 11, but for three convective cases and one typhoon case.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

Finally, we have analyzed the effect of time mismatching in our results. For the calculation of the RMSD, we have used the second-closest GR composite in time. Figure 13 shows the results compared with those presented in Figs. 11c and 12c (i.e., using the closest GR composite in time) for a widespread case on 14 July 2012 (Fig. 13a) and a convective case on 13 July 2012 (Fig. 13b). For both cases, the RMSDs at 2-min differences are obviously smaller than those at 8-min differences. The convective case is more affected by time mismatching, showing a larger increase in RMSDs than the widespread case.

Fig. 13.
Fig. 13.

RMSD values between PR and GR reflectivities (larger than 18 dBZ) in the vertical for a (a) widespread and (b) convective case. The black line is computed with the PR–GR matched at the shortest time. The green line is done with the second-closest time matching. The circle size is proportional to the number of points used in the computation.

Citation: Journal of Hydrometeorology 16, 2; 10.1175/JHM-D-14-0092.1

4. Summary and conclusions

Both spaceborne Precipitation Radar (PR) and ground-based scanning radars (GR) measure reflectivity that provides information about the volumetric distribution of precipitation. This study has set up a framework for the comparison of PR measurements and GR composites focusing on the vertical structure of reflectivity for eight summertime rain cases over the southern Korean Peninsula. GR composites have been produced from the observations of six operational S-band ground radars. The observations from each GR have been processed to reduce ground clutter and have been calibrated among other radars using a dual-polarization radar as a reference.

Because of the differences between PR and GR data in spatial and temporal resolutions and viewing geometry, the data matching is done (i) at closest in time (2–5 min differences), (ii) by interpolating PR data linearly over the 3D Cartesian grid of the GR composite that is set with PR spatial resolutions (5 km × 5 km × 0.25 km) for statistical analyses, and (iii) by taking into account GR beam broadening (e.g., the GR composite generated with radar coverage up to 60 km without data interpolation in the vertical).

Based on the dominant precipitation patterns over the field, four widespread and convective (including one typhoon) rain cases have been analyzed. GR composites have a better ability to capture the structure of small-scale convective cells with stronger intensities at low levels than PR observations. On the other hand, PR seems to better reproduce the location and thickness of bright bands because of the better PR vertical resolution, while GR bright bands are affected by beam broadening with range. One possibility to compensate for the effect of GR beam broadening in PR–GR comparison could be applying some range-dependent adjustment factor based on an idealized VPR [as done by Gabella et al. (2006) and Kirstetter et al. (2013)]. The statistical comparisons in the vertical show clear differences between PR and GR observations:

  • In the widespread rain cases, the vertical distributions of the two measurements are similar below the melting layer. However, those above and within the melting layer are quite different because the two instruments have their limitations: PR is affected by its sensitivity limit (around 18 dBZ), and GR beam broadening at far ranges causes systematic differences in the PR–GR comparisons. The scattering differences between PR and GR can also be a source of relatively large discrepancy in the upper levels (with biases up to 3 dB), as concluded by Wen et al. (2011) based on backscattering simulations.
  • In the convective rain cases, the mean vertical reflectivities measured from PR at low levels (between 1.5 and 4 km) are underestimated compared with GR observations (1–3 dB). Although the presented cases did not contain many strong echoes (mostly below 35 dBZ even for the typhoon case), such PR–GR discrepancies seem to be present. The PR attenuation-correction algorithm may still underperform for those small-scale moderate rainfall cells (nonuniform beam filling), indicated by growing discrepancy in the mean reflectivity downward in the layer. At the same time, we have shown that the effect of time mismatching is not negligible for the convective cases and can explain a significant part of the RMSD for time offsets of a few minutes between PR and GR observations.

The small differences between PR and GR CFADs in widespread rain (up to the melting layer) indicate that GR calibration has been successfully maintained under reasonable limits. The feasibility of using PR as an external reference for the calibration of individual GRs can be further investigated [as similarly suggested by Anagnostou et al. (2001) and Wang and Wolff (2009)] and compared with the dual-polarization self-consistency approach. However, some discrepancies in PR and GR measurements (due to factors such as the type of precipitation, beam broadening, scattering differences in snow, and attenuation) analyzed in this work should be taken into account when using PR as calibration reference for GRs.

The presented framework will be used for the analysis of the observations of the Global Precipitation Measurement (GPM) PR. Because both PR and GR observations suffer from their own limitations, it could also be of interest to add independent reflectivity measurements from ground-based X-band scanning and/or vertically pointing radars to cover data-void locations in the validation of PR products.

Acknowledgments

This study was supported by Research for the Meteorological and Earthquake Observation Technology and Its Application (I) (NIMR-2012-B-3) of the National Institute of Meteorological Research and by the Development and Application of Cross Governmental Dual-Pol Radar Harmonization (WRC-2013-A-1) project of the Weather Radar Center, Korea Meteorological Administration, in 2015. The authors are grateful to Cheong-Ryong Lee (M.Sc. student at Kyungpook National University) and Soo-Hyun Kwon (Ph.D. student at Kyungpook National University) for their help in the calibration of GR radar network. The TRMM data used in this study were acquired as part of the NASA’s Earth–Sun System Division and archived and distributed by the Goddard Earth Sciences (GES) Data and Information Services Center (DISC).

REFERENCES

  • Amitai, E., , Llort X. , , and Sempere-Torres D. , 2009: Comparison of TRMM radar rainfall estimates with NOAA next-generation QPE. J. Meteor. Soc. Japan, 87A, 109118, doi:10.2151/jmsj.87A.109.

    • Search Google Scholar
    • Export Citation
  • Anagnostou, M. S., , Morales C. A. , , and Dinku T. , 2001: The use of TRMM Precipitation Radar observations in determining ground radar calibration biases. J. Atmos. Oceanic Technol., 18, 616628, doi:10.1175/1520-0426(2001)018<0616:TUOTPR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bellon, A., , Lee G. , , and Zawadzki I. , 2005: Error statistics of VPR corrections in stratiform precipitation. J. Appl. Meteor., 44, 9981015, doi:10.1175/JAM2253.1.

    • Search Google Scholar
    • Export Citation
  • Bolen, S., , and Chandrasekar V. , 2000: Quantitative cross validation of space-based and ground-based radar observations. J. Appl. Meteor., 39, 20712079, doi:10.1175/1520-0450(2001)040<2071:QCVOSB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Cao, Q., , Hong Y. , , Qi Y. , , Wen Y. , , Zhang J. , , Gourley J. J. , , and Liao L. , 2013: Empirical conversion of vertical profile of reflectivity from Ku-band to S-band frequency. J. Geophys. Res. Atmos., 118, 18141825, doi:10.1002/jgrd.50138.

    • Search Google Scholar
    • Export Citation
  • Chen, S., and et al. , 2013: Evaluation of spatial errors of precipitation rates and types from TRMM spaceborne radar over the southern CONUS. J. Hydrometeor., 14, 18841896, doi:10.1175/JHM-D-13-027.1.

    • Search Google Scholar
    • Export Citation
  • Cho, Y.-H., , Lee G. W. , , Kim K.-E. , , and Zawadzki I. , 2006: Identification and removal of ground echoes and anomalous propagation using the characteristics of radar echoes. J. Atmos. Oceanic Technol., 23, 12061222, doi:10.1175/JTECH1913.1.

    • Search Google Scholar
    • Export Citation
  • Doviak, R. J., , and Zrnić D. S. , 1993: Doppler Radar and Weather Observations.2nd ed. Dover Publications, 562 pp.

  • Gabella, M., , Joss J. , , Perona G. , , and Michaelides S. , 2006: Range adjustment for ground-based radar, derived with the spaceborne TRMM Precipitation Radar. IEEE Trans. Geosci. Remote Sens., 44, 126133, doi:10.1109/TGRS.2005.858436.

    • Search Google Scholar
    • Export Citation
  • GAMIC, 2008: ENIGMA 3+ theory of operation. GAMIC mbH, 84 pp.

  • Heymsfield, G. M., , Geerts B. , , and Tian L. , 2000: TRMM Precipitation Radar reflectivity profiles as compared with high-resolution airborne and ground-based radar measurements. J. Appl. Meteor., 39, 20802102, doi:10.1175/1520-0450(2001)040<2080:TPRRPA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Iguchi, T., , Kozu T. , , Meneghini R. , , Awaka J. , , and Okamoto K. , 2000: Rain profiling algorithm for the TRMM Precipitation Radar. J. Appl. Meteor., 39, 20382052, doi:10.1175/1520-0450(2001)040<2038:RPAFTT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Iguchi, T., , Kozu T. , , Kwiatkowski J. , , Meneghini R. , , Awaka J. , , and Okamoto K. , 2009: Uncertainties in the rain profiling algorithm for the TRMM Precipitation Radar. J. Meteor. Soc. Japan, 87A, 130, doi:10.2151/jmsj.87A.1.

    • Search Google Scholar
    • Export Citation
  • Kirstetter, P. E., and et al. , 2012: Toward a framework for systematic error modeling of spaceborne radar with NOAA/NSSL ground radar–based National Mosaic QPE. J. Hydrometeor., 13, 12851300, doi:10.1175/JHM-D-11-0139.1.

    • Search Google Scholar
    • Export Citation
  • Kirstetter, P. E., , Andrieu H. , , Delrieu G. , , and Boudevillain B. , 2013: A physically based identification of vertical profiles of reflectivity from volume scan radar data. J. Appl Meteor. Climatol., 52, 16451663, doi:10.1175/JAMC-D-12-0228.1.

    • Search Google Scholar
    • Export Citation
  • Langston, C., , Zhang J. , , and Howard K. , 2007: Four-dimensional dynamic radar mosaic. J. Atmos. Oceanic Technol., 24, 776790, doi:10.1175/JTECH2001.1.

    • Search Google Scholar
    • Export Citation
  • Lee, G. W., , and Zawadzki I. , 2006: Radar calibration by gage, disdrometer, and polarimetry: Theoretical limit caused by the variability of drop size distribution and application to fast scanning operational radar data. J. Hydrol., 328, 8397, doi:10.1016/j.jhydrol.2005.11.046.

    • Search Google Scholar
    • Export Citation
  • Liao, L., , and Meneghini R. , 2009: Validation of TRMM Precipitation Radar through comparison of its multiyear measurements with ground-based radar. J. Appl. Meteor. Climatol., 48, 804817, doi:10.1175/2008JAMC1974.1.

    • Search Google Scholar
    • Export Citation
  • Liao, L., , Meneghini R. , , and Iguchi T. , 2001: Comparisons of rain rate and reflectivity factor derived from the TRMM Precipitation Radar and the WSR-88D over the Melbourne, Florida, site. J. Atmos. Oceanic Technol., 18, 19591974, doi:10.1175/1520-0426(2001)018<1959:CORRAR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Meneghini, R., , Jones J. A. , , Iguchi T. , , Okamoto K. , , and Kwiatkowski J. , 2004: A hybrid surface reference technique and its application to the TRMM Precipitation Radar. J. Atmos. Oceanic Technol., 21, 16451658, doi:10.1175/JTECH1664.1.

    • Search Google Scholar
    • Export Citation
  • Park, S.-G., , and Lee G. , 2010: Calibration of radar reflectivity measurements form KMA operational radar network. Asia–Pac. J. Atmos. Sci., 46, 243259, doi:10.1007/s13143-010-1010-3.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and Houze R. A. Jr., 2000: Comparison of radar data from the TRMM satellite and Kwajalein ocean validation site. J. Appl. Meteor., 39, 21512164, doi:10.1175/1520-0450(2001)040<2151:CORDFT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schwaller, M. R., , and Morris K. R. , 2011: A ground validation network for the global precipitation measurement mission. J. Atmos. Oceanic Technol., 28, 301319, doi:10.1175/2010JTECHA1403.1.

    • Search Google Scholar
    • Export Citation
  • Siggia, A., , and Passarelli R. , 2004: Gaussian model adaptive processing (GMAP) for improved ground clutter cancellation and moment calculation. Proceedings of ERAD (2004), Copernicus GmBH, 67–73. [Available online at http://copernicus.org/erad/2004/online/ERAD04_P_67.pdf.]

  • TRMM Precipitation Radar Team, 2011: Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar algorithm: Instruction manual for version 7. JAXA/NASA, 170 pp. [Available online at www.eorc.jaxa.jp/TRMM/documents/PR_algorithm_product_information/pr_manual/PR_Instruction_Manual_V7_L1.pdf.]

  • Wang, J., , and Wolff D. , 2009: Comparisons of reflectivities from the TRMM Precipitation Radar and ground-based radars. J. Atmos. Oceanic Technol., 26, 857875, doi:10.1175/2008JTECHA1175.1.

    • Search Google Scholar
    • Export Citation
  • Wen, Y., , Hong Y. , , Zhang G. , , Schuur T. J. , , Gourley J. J. , , Flamig Z. , , Morris K. R. , , and Cao O. , 2011: Cross validation of spaceborne radar and ground polarimetric radar aided by polarimetric echo classification of hydrometeor types. J. Appl Meteor. Climatol., 50, 13891402, doi:10.1175/2011JAMC2622.1.

    • Search Google Scholar
    • Export Citation
  • Yuter, S. E., , and Houze R. A. 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, doi:10.1175/1520-0493(1995)123<1941:TDKAME>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, J., , Howard K. , , and Gourley J. J. , 2005: Three-dimensional multiple radar reflectivity mosaic. J. Atmos. Oceanic Technol., 22, 3042, doi:10.1175/JTECH-1689.1.

    • Search Google Scholar
    • Export Citation
Save