Dataset Analysis to Support Verification of GPM Science Requirements
No abstract available
Precipitation↗
Engineering topics
Thurai, M.
Publications and source records attributed to Thurai, M..
High Resolution Observations of Drop Size Distribution for GPM Ground Validation
During the Mid-latitude Continental Convective Cloud Experiment (MC3E), NASA's GPM GV Disdrometer and Radar Observations of Precipitation (DROP) Facility deployed an array of disdrometers and rain gauges in northern Oklahoma to sample, with high resolution, the drop size distribution for use in development of precipitation retrieval algorithms for the GPM core satellites. The DROP Facility instruments deployed during MC3E consisted of 16 autonomous Parsivel units, 5 two-dimensional video disdrometers (2dvds), a vertically pointing K band radar, and 32 tipping bucket rain gauges. There were several rainfall events during MC3E in which rain drops exceeding 6 mm in diameter were recorded. The disdrometer array revealed large rain drops with diameters exceeding 6 mm and 8 mm during two separate stratiform and convective rainfall events, respectively. The NPOL radar, which was scanning in high resolution RHI mode (every 40 sec) over the disdrometer array during the stratiform event, indicated a 1 km thick bright band with a differential reflectivity column of 2-3 dB extending below the melting layer to the surface where the large drops were recorded by the 2dvds. These large drops are important for GPM since they can have a great impact upon satellite precipitation retrieval, especially near the ground and below heavy convective rainfall cores where satellites have had problems depicting the rainfall.
Gatlin, Patrick N.↗
Drop Shapes Versus Fall Velocities in Rain: 2 Contrasting Examples
Rainfall retrievals from polarimetric radar measurements require the knowledge of four fundamental rain microstructure parameters, namely, drop size distribution, drop shape distribution, canting angles and drop fall velocities. Some recent measurements of all four parameters in natural rain are summarized in [1]. In this paper, we perform an in-depth analysis of two events, using two co-located 2D video disdrometers (2DVD; see [2]) both with high calibration accuracy, and a C-band polarimetric radar [3], located 15 km away. The two events, which occurred 7 days apart (on the 18th and the 25th of Dec 2009), had moderate-to-intense rainfall rates, but the second event had an embedded convection line within the storm. The line had passed over the 2DVD site, thus enabling the shapes and fall velocities to be determined as the line crossed the site. The first event was also captured in a similar manner by both the 2DVDs as well as the C-band radar. Drop fall velocity measurements for, say, the 3 mm drops show noticeable differences between the two events. Whereas for the first event, the velocity distribution showed a narrow and symmetric distribution, with a mode at the expected value (7.95 m/s, as given by the formula in [4]), the second event produced a wider distribution with a significant skewness towards lower velocities (although its mode too was close to the expected value). Moreover, the slower 3 mm drops in the second event occurred when the convection line was directly over the 2DVD site (03:35-03:45 utc), and not before nor after. A similar trend was observed in terms of the horizontal dimensions of the 3 mm drops, i.e. large fluctuations during the same time period, but not outside the period. Vertical dimensions of the drops also fluctuated but not to the same extent. Interestingly, the horizontal dimensions tended towards larger values during the 10-minute period, implying an increase in drop oblateness, which in turn indicates the possibility of the horizontal mode oscillation, one of the three fundamental modes of drop oscillations [5], albeit the most difficult one to excite.
Thurai, M.↗
DSD Characteristics of a Mid-Winter Tornadic Storm Using C-Band Polarimetric Radar and Two 2D-Video Disdrometers
Drop size distributions in an evolving tornadic storm are examined using C-band polarimetric radar observations and two 2D-video disdrometers. The E-F2 storm occurred in mid-winter (21 January 2010) in northern Alabama, USA, and caused widespread damage. The evolution of the storm occurred within the C-band radar coverage and moreover, several minutes prior to touch down, the storm passed over a site where several disdrometers including two 2D video disdrometers (2DVD) had been installed. One of the 2DVDs is a low profile unit and the other is a new next generation compact unit currently undergoing performance evaluation. Analyses of the radar data indicate that the main region of precipitation should be treated as a "big-drop" regime case. Even the measured differential reflectivity values (i.e. without attenuation correction) were as high as 6-7 dB within regions of high reflectivity. Standard attenuation-correction methods using differential propagation phase have been "fine tuned" to be applicable to the "big drop" regime. The corrected reflectivity and differential reflectivity data are combined with the co-polar correlation coefficient and specific differential phase to determine the mass-weighted mean diameter, Dm, and the width of the mass spectrum, (sigma)M, as well as the intercept parameter , Nw. Significant areas of high Dm (3-4 mm) were retrieved within the main precipitation areas of the tornadic storm. The "big drop" regime assumption is substantiated by the two sets of 2DVD measurements. The Dm values calculated from 1-minute drop size distributions reached nearly 4 mm, whilst the maximum drop diameters were over 6 mm. The fall velocity measurements from the 2DVD indicate almost all hydrometeors to be fully melted at ground level. Drop shapes for this event are also being investigated from the 2DVD camera data.
Thurai, M.↗
Drop Axis Ratio Distributions in Stratiform and Convective Rain
A fully calibrated low profile 2D video disdrometer (2DVD) has been recording many different rainfall events in Northern Alabama (USA) since June 2007. An earlier publication reported drop shapes and axis ratio distributions determined for some of the events. For one of the cases examined, a noticeable shift in the 3.5 - 3.75 mm drop axis ratio distribution was noted. In this paper, we extend the earlier work by separating the 2DVD measurements into stratiform and convective rain. The separation is made possible by using the minute-by-minute drop size distribution (DSD) measured by the 2DVD. The 1-minute DSDs are fitted to a gamma distribution, and using a simple indexing technique which involves two of the fitted parameters, periods of convective and stratiform rain are separated for a given event. The output of the DSD indexing technique is qualitatively confirmed by comparing with simultaneous time series observations from a co-located UHF profiler which continuously records height profiles of reflectivity, Doppler mean and spectral width, all of which enable the identification of bright-band periods and, furthermore, periods of moderate and deep convection. Excellent consistency is found between the output of the DSD-based separation method and the profiler observations. Next, we utilize the output of DSD index-based separation method to flag the periods of severe convection for a given event. Drop axis ratios during the flagged periods are derived and compared with those during stratiform rain periods. Five cases have been considered. Axis ratio distributions do not show appreciable differences between stratiform and convective periods for four of the cases. The fifth case (the same case as reported earlier) shows a shift in the 3.5 - 3.75 mm drop axis ratios during a prolonged period of convection. The contoured shapes for these drops determined from the 2DVD camera data indicate the possibility of non-axisymmetric oscillations, compared with the contoured images for other events which fit well to our reference drop shapes. For all of above cases, observations from a C-band polarimetric radar - situated 15 km away are examined. The variations between the co-polar radar reflectivity and the differential reflectivity as well as the specific differential phase are compared with the 2DVD data based scattering calculations for the 5 events. The implications will be discussed.
Thurai, M.↗