This paper documents the production and validation of retrieved rainfall data obtained from satellite-borne microwave radiometers by the Global Satellite Mapping of Precipitation (GSMaP) Project. Using various attributes of precipitation derived from Tropical Rainfall Measuring Mission (TRMM) satellite data, the GSMaP has implemented hydrometeor profiles derived from Precipitation Radar (PR), statistical rain/no-rain classification, and scattering algorithms using polarization-corrected temperatures (PCTs) at 85.5 and 37 GHz. Combined scattering-based surface rainfalls are computed depending on rainfall intensities. PCT85 is not used for stronger rainfalls, because strong depressions of PCT85 are related to tall precipitation-top heights. Therefore, for stronger rainfalls, PCT37 is used, with PCT85 used for weaker rainfalls. With the suspiciously strong rainfalls retrieved from PCT85 deleted, the combined rainfalls correspond well to the PR rain rates over land. The GSMaP algorithm for the TRMM Microwave Imager (TMI) is validated using the TRMM PR, ground radar [Kwajalein (KWAJ) radar and COBRA], and Radar Automated Meteorological Data Acquisition System (AMeDAS) precipitation analysis (RA). Monthly surface rainfalls retrieved from six microwave radiometers (GSMaP_MWR) are compared with the gauge-based dataset. Rain rates retrieved from the TMI (GSMaP_TMI) are in better agreement with the PR estimates over land everywhere except over tropical Africa in the boreal summer. Validation results of the KWAJ radar and COBRA show a good linear relationship for instantaneous rainfall rates, while validation around Japan using the RA shows a good relationship in the warm season. Poor results, connected to weakprecipitation cases, are found in the cold season around Japan.

This paper presents an inter-comparison of rainfall parameters (median volume diameter and rain rate) using C-band polarimetric radar, a 2D-video disdrometer and a 400 MHz profiler for the Baiu front event of 8-9 June 2005 in Okinawa, Japan. These instruments are part of the Okinawa Sub-Tropical Environment Remote Sensing Center, operated by the National Institute of Information and Communications Technology (NICT). The 2D-video disdrometer is used to derive the mean axis ratio of raindrops versus drop diameter, as well as the drop size distribution for the Baiu event. The data are then used to simulate various relations between polarimetric scattering parameters such as: specific attenuation ðA h Þ, and specific differential attenuation ðA dp Þ, versus specific differential phase ðK dp Þ which are required to correct the measured reflectivity at horizontal polarization ðZ h Þ, and the differential reflectivity ðZ dr Þ for rain attenuation. The 2D-video disdrometer data are also used to arrive at retrieval formulas for median volume diameter ðD o Þ from radar Z dr , and rain rate from radar K dp .The These inter-comparisons demonstrate the accuracy of C-band polarimetric radar to retrieve important rainfall parameters, as well as the accurate correction for rain attenuation using differential propagation phase.

Characteristics of polarimetric radar variables in three rainfall types in a Baiu front event over the East China Sea observed on 1 June 2004 were studied and compared using a C-band polarimetric radar, the CRL Okinawa bistatic polarimetric radar (COBRA). The selected rainfalls are common types in the Baiu season in this area: (1) stratiform type (ST), (2) isolated convective type (ICT) and (3) embedded convective type (ECT). ST was characterized by an obvious bright band in the field of radar reflectivity (Z hh ). ICT and ECT had almost the same 30-dBZ echo-top height of about 5.5 km, and their strong echo region (Z hh > 40 dBZ) did not exceed the 0 C level (4.4 km altitude) even in their mature stages. Around the 0 C level, overall decrease in correlation coe‰cient between horizontal and vertical polarization signals (r hv ) and increase in di¤erential reflectivity (Z dr ) were observed in ST and ECT, which indicated the presence of a layer of mixed-phase precipitation. By contrast, significant decrease in r hv and increase in Z dr were not found in ICT. At lower levels, Z dr ranged from 0 to 1.5 dB and most of r hv were higher than 0.98 in ST and ECT. The values of Z dr and r hv had wider variations in ICT. The characteristics of the vertical profiles of Z dr and r hv in ECT are consequently more similar to those in ST rather than to those in ICT, although their echo-top heights of 30 dBZ and maximum Z hh near the surface were almost equal.Corresponding author and current a‰liation: Yukari Shusse, Hydrospheric Atmospheric Research Center, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8601, Japan. E-mail: shusse@rain.hyarc.nagoya-u.ac.jp 6 2009, Meteorological Society of Japan Near the surface, su‰ciently below the 0 C level, ICT had larger Z dr and lower r hv than ECT in the region with Z hh stronger than 45 dBZ. This indicates that larger raindrops were more dominant in ICT than in ECT at the same Z hh in the intense rainfall regions. This di¤erence of dominant raindrop sizes appears to reflect the difference in precipitation growth processes between ICT and ECT.

No abstract

No abstract