1. Introduction and background
Precipitation in convective clouds can be produced through various microphysical processes. The warm rain process, consisting of growth of liquid precipitation via droplet collision and coalescence, occurs in convection in a variety of environments (e.g., Rogers 1967; Houze 1977; Petersen et al. 1999; Stith et al. 2002; Golding et al. 2005; Blyth et al. 2013). This can be the exclusive mechanism for precipitation development, for example in clouds exclusively or primarily existing at temperatures above 0°C (Johnson 1982; Beard et al. 1986; Szumowski et al. 1997, 1998; Leon et al. 2016). Precipitation production in convective clouds at lower temperatures may result in mixed- or ice-phase hydrometeors, with relevant growth processes including accretion and aggregation (Braham 1964, 1968; Johnson 1987). Multiple studies have indicated that initial warm rain production can be of fundamental importance to the later development of precipitation through ice processes (e.g., Koenig 1963; Cooper and Lawson 1984; Johnson 1987; Blyth and Latham 1993; Petersen et al. 1999; Stith et al. 2002). Liquid drops initially formed via collision–coalescence can serve as embryos for graupel production, potentially leading to secondary ice production via the Hallett–Mossop multiplication process (Hallett and Mossop 1974). In addition, they can directly “feed” hydrometeor growth by providing a source of supercooled liquid for accretion (Cooper and Lawson 1984; Johnson 1987; Blyth and Latham 1993; Jameson et al. 1996; Petersen et al. 1999; Stith et al. 2002; Huang et al. 2008; Kumjian and Ryzhkov 2008; Taylor et al. 2016). This study investigates the characteristics of rainfall production associated with convection in which both warm rain and ice processes are active.
Dual-polarization radar measurements are a key component in the observational investigation of these processes. Characteristic signatures are associated with hydrometeor properties such as phase, shape, and density (e.g., Doviak and Zrnić 1993). Many studies use parameters such as radar reflectivity factor ZH and differential reflectivity ZDR, the latter providing a reflectivity-weighted estimate of horizontal versus vertical backscatter from hydrometeors, to provide critical information for inferring the microphysical processes most likely to be occurring (e.g., Zhang et al. 2006; Cao et al. 2008; Carr et al. 2017). Early polarimetric radar studies of warm-based convection (i.e., convection for which the inflow source is at temperatures above 0°C) over southern England (Hall et al. 1984; Caylor and Illingworth 1987; Illingworth et al. 1987) focused on measurements of ZH and ZDR, noting that some actively growing convective cells contained “ZDR columns,” defined as areas of enhanced, positive ZDR values associated with moderate ZH values and extending above the environmental 0°C level within the cell cores. Values of ZDR of 3–4 dB were common in the columns identified in these studies, with the columns typically extending ~1–2 km in width and up to 1.5–2 km above the environmental 0°C level. The ZDR column signature was concluded to be a likely indicator of small concentrations of large liquid drops [estimated at up to 4–6 mm in diameter in the cases analyzed by Illingworth et al. (1987)] lofted above the environmental 0°C level within the updrafts of growing cells, prior to glaciation. In some cases, the ZDR columns were capped by weakly negative ZDR values, corresponding to enhanced linear depolarization ratio values reported in later studies. These signatures are more characteristic of solid hydrometeors and were identified as indicators of some liquid transitioning into ice, implying mixed-phase conditions, with the remaining liquid acting as a source for growth by accretion (Hall et al. 1984; Herzegh and Jameson 1992; Jameson et al. 1996; Bringi et al. 1997; Kumjian et al. 2012). Thus, these early studies identified the ZDR column signature, along with decreased ZDR values farther aloft, as an important indicator of the microphysical structure within developing convection.
Such studies have confirmed the existence of ZDR columns and their inferred microphysical characteristics (including the presence of small wet hail or graupel; e.g., Kumjian et al. 2014) in other environments. The studies commonly focus on deeper convection in environments with higher instability, with many using observations in the U.S. Southeast (e.g., Tuttle et al. 1989; Fulton and Heymsfield 1991; Yuter and Houze 1995; Jameson et al. 1996; Bringi et al. 1996; French et al. 1996; Ramachandran et al. 1996; Zeng et al. 2001) or the Great Plains and Intermountain West (e.g., Brandes et al. 1995; Bringi et al. 1996; Blyth et al. 1997; Kennedy et al. 2001; Knight 2006; Rowe et al. 2011; Kumjian and Ryzhkov 2008; Kumjian et al. 2012, 2014; Van Den Broeke 2016; van Lier-Walqui et al. 2016). Such signatures have also been observed in both squall lines and discrete convection in Germany, Australia, and Japan (Meischner et al. 1991; Keenan et al. 2000; Adachi et al. 2013). Enhanced positive ZDR values often extended to between −10° and −15°C, and up to −25°C in the deepest cases (Bringi et al. 1996; Keenan et al. 2000; Knight 2006). Studies incorporating airborne measurements found a correspondence between the ZDR column and the convective updraft, identifying low concentrations of large liquid drops (diameter D ≥ 5 mm) associated with the enhanced positive ZDR values (Brandes et al. 1995; Bringi et al. 1996). Measurements supported the possibility of collision–coalescence within the convective updraft, with supercooled drops acting as a source for rimed growth given partial glaciation (Bringi et al. 1991, 1996, 1997; French et al. 1996; Blyth et al. 1997).
Given the results of these studies, the ZDR column structure is characteristic of convection in which a strong warm rain process has contributed to precipitation evolution. While the signature is associated both with supercooled liquid drops and, eventually, mixed-phase hydrometeors [e.g., wet graupel or hail as in Kumjian et al. (2014)], it is used here as an indicator of warm rain production having influenced precipitation evolution of warm-based convection, through the initial growth of liquid drops and the availability of this liquid for eventual accretional growth. The ZDR column signature has also been used as a proxy for identifying convective updraft location and intensity (e.g., Tuttle et al. 1989; Brandes et al. 1995; Kennedy et al. 2001), again with a focus on deep, often supercellular, convection in high-instability environments. This use of the column signature has primarily occurred within the context of operational storm hazard detection, leading to the development of techniques for objectively locating ZDR columns (Kumjian and Ryzhkov 2008; Picca et al. 2010; Adachi et al. 2013; Kumjian et al. 2014; Snyder et al. 2015).
The goal of this study is to describe observations of ZDR columns in relatively shallow convection developing in environments in which few prior observations of convection producing such structures exists, using a set of radar observations large enough to be suitable for statistical analysis. The object of these analyses is then to identify whether distinct microphysical characteristics inferred from the radar measurements are associated with quantifiable enhancements in rain production. In addition, this study also takes advantage of objective ZDR column identification in the context of established knowledge regarding the column signature and its implications on identifying relevant microphysical processes that have occurred in the convection. Datasets collected over southwestern England in July–August 2013 during the Convective Precipitation Experiment (COPE; Leon et al. 2016) are used for this study. The warm-season synoptic environment in this region sometimes allows the development of convection that produces heavy rainfall (defined here as rainfall rates exceeding ~25 mm h−1), sometimes leading to devastating floods (Golding et al. 2005). The convection is based near 10°C, but it is often deep enough for glaciation and ice growth to occur (Leon et al. 2016). Measurements of such convection were made during COPE using a polarimetric Doppler X-band weather radar and supplemented with airborne measurements of hydrometeor characteristics and bulk cloud properties, with the aircraft also directly sampling multiple ZDR columns.
The convection observed during COPE provides a new set of measurements showing frequent development of ZDR columns in convective environments compared to expectations from past studies. Early studies of ZDR columns (e.g., Hall et al. 1984; Caylor and Illingworth 1987; Illingworth et al. 1987; see also Brown and Swann 1997) featured more limited measurements of convection that developed outside of the COPE domain. In addition, the COPE convection is distinct from that described by more recent ZDR column studies, which focus on deep convection developing in environments with higher instability and stronger vertical wind shear (e.g., Kumjian et al. 2012, 2014; Van Den Broeke 2016; van Lier-Walqui et al. 2016). In addition to showing the unexpectedly frequent occurrence of ZDR columns in relatively shallow convection, the radar measurements from COPE provide a sufficiently large set of observations to be suitable for statistical analyses, and they were supplemented with airborne measurements of cloud properties.
Subsequent sections of this paper will describe the COPE project and the specific datasets used in this study, along with the method developed for objective identification of convective cells with ZDR columns. Next, details regarding the structural properties of these cells will be presented along with implications regarding their likely microphysical characteristics. Last, a statistical summary of the rainfall produced by these cells is presented, as compared with the more general convective population.
2. Data and method
a. Data and case studies
The datasets used in this study were obtained during the COPE field campaign (Leon et al. 2016), during which airborne and ground-based instrumentation were deployed to observe warm-based convection forming over the southwest peninsula of England from July to August of 2013. Sea-breeze convergence lines commonly form over the peninsula and can act as a localized trigger for convection initiation. This provided the opportunity to sample many convective cells throughout their lifetime within a relatively small region. This study focuses on measurements from the National Centre for Atmospheric Science’s (NCAS) Doppler dual-polarization X-band weather radar (NXPol; https://www.ncas.ac.uk/index.php/en/about-amf/263-amf-main-category/amf-x-band-radar/1098-x-band-radar-overview; Neely et al. 2018), which was deployed at a fixed location throughout COPE, as indicated in Fig. 1 (Bennett 2017). The NXPol observations were supplemented with measurements made within its domain using the University of Wyoming King Air (UWKA) research aircraft, with directly coincident measurements made in some cases (e.g., Jackson et al. 2018). During COPE, the NXPol provided measurements using an effective beamwidth of ~1° with a range spacing of 150 m. Sampling strategies focused on constructing volumes of plan position indicator (PPI) scans at ~5-min resolution, using 10 incremental elevation angles in 1°–3° steps. Measured parameters include the radar reflectivity factor at horizontal polarization ZH, Doppler velocity, and spectrum width, as well as polarimetric parameters including differential reflectivity ZDR, differential phase shift, and cross-correlation coefficient. The NXPol measurements were processed following COPE to correct for data quality issues inherent in sampling convective environments using X-band radar (e.g., significant attenuation that was evident in ZH and ZDR). Basic quality control of the radar data included removing second-trip echoes using a variable threshold filter (Dufton 2016) and removing nonmeteorological echoes using fuzzy logic classification (Dufton and Collier 2015). A calibration offset of −4.25 dBZ for ZH was calculated following the self-consistency technique in Gourley et al. (2009). As vertical scans were not available, ZDR was calibrated using the measured ZH and ZDR values in drizzle, with a basic offset of −1.25 dB. The measurements were corrected for attenuation using a modification of the differential phase shift–based “ZPHI” method from Testud et al. (2000), as implemented in the Python ARM Radar Toolkit (Py-ART; Helmus and Collis 2016). Last, rainfall rate was computed using a combination of dual polarization estimators and a corrected reflectivity estimate, with a specific attenuation Ah rainfall estimator, R(Ah), used as the primary method of quantitative precipitation estimation (QPE) when possible, requiring a phase shift ≥ 5° along a given radial to be used. Recent studies have shown rainfall estimates computed from Ah to be more reliable than previously implemented QPE methods (e.g., Ryzhkov et al. 2014; Wang et al. 2014; Diederich et al. 2015), which was also found to be the case for the COPE field campaign (Dufton 2016). Where it was not possible to accurately compute R(Ah), a weighted combination of rain rates calculated from specific differential phase KDP (derived via least squares regression using the Rainbow 5 software package; http://www.de.selex-es.com/capabilities/meteorology/products/components/rainbow5) and the corrected reflectivity values was used. In these cases, the weighting for the KDP-based calculations increased with larger rain rates, and R(KDP) was used exclusively for rain rates at or above 10 mm h−1.
Measurements of ZH (dBZ) from the NXPol gridded to 2.25-km altitude, from scans valid at (a) 1402:34 UTC 29 Jul, (b) 1547:03 UTC 2 Aug, and (c) 1503:54 UTC 3 Aug 2013. The UWKA base location is indicated in red, and the NXPol and rawinsonde launch site is shown in black, with dashed lines indicating 30-, 60-, and 90-km range from radar.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
Three cases from COPE were used for this study, each featuring the development of warm-based convection that became deep enough for ice processes to become relevant. Leon et al. (2016) and Jackson et al. (2018) provide additional details regarding these cases, with the latter providing a summary table. Cloud bases were near 10°C in each case, limiting the depth between cloud base and the environmental freezing level to ~2 km or less. However, relatively low mean updraft speeds are indicative of adequate time for liquid growth via collision–coalescence. Again, the convection in each case was shallow compared to recent studies of ZDR columns, with clouds commonly extending to 4–6-km altitude (7–8 km for the deepest clouds) and developing in environments of lower shear and instability (typically 100–300 J kg−1, with equilibrium levels generally close to the deeper cloud-top heights observed in each case). The lifetime of individual convective cells was also shorter than typical for deep convection, with the ZDR column structures often developing and evolving on shorter time scales.
On 29 July, convection formed ahead of a synoptic-scale trough over the peninsula, with the primary surface cyclone centered well north of the United Kingdom. The convective mode included both a relatively linear convective complex over the peninsula, and smaller, relatively discrete convective cells (Fig. 1a). The maximum surface-based convective available potential energy (CAPE) was ~300 J kg−1, with clouds commonly reaching 4–6-km altitude. Cloud bases were near 11°C with the deeper cloud tops from −25° to −30°C (~6–7 km). The mean maximum updraft measured by aircraft was ~9 m s−1, with the strongest updraft being ~15 m s−1 (Jackson et al. 2018). Convective updraft strength and cloud depth were in the midrange of the three cases analyzed here, with environmental vertical wind shear of ~1 × 10−3 s−1, as calculated between cloud base and the deepest cloud tops from rawinsondes launched at the NXPol site in Davidstow (Cornwall, England).
On 2 August, a synoptic low pressure center was located northwest of Ireland. Convection developed in post-cold-frontal flow over the peninsula. The convective cells that developed were initially relatively discrete, but they coalesced into a linear complex (e.g., Fig. 1b) along the sea-breeze convergence line (with some separate cells off of the line) over the course of the day. Maximum CAPE values were ~620 J kg−1, with the strongest updrafts and deepest clouds occurring during this case. Cloud tops were commonly near 4–6 km, with bases near 12°C and the deepest tops from −30° to −35°C (~7–8 km). The mean maximum updraft speeds were again ~9 m s−1, with the largest observed speed being ~18 m s−1. The strongest vertical wind shear (~5.2 × 10−3 s−1) also occurred during this case.
On 3 August, the surface low pressure center had moved slightly north from the previous day. Again, convective cells were initially relatively discrete but became more linearly organized along the sea breeze convergence lines through the day (Fig. 1c). A shallower unstable environment was present relative to that of 2 August, with maximum CAPE values near ~250 J kg−1. Weaker convective updrafts and shallower clouds were evident relative to conditions on the other days. Cloud tops were commonly 4–5 km, with bases near 11°C and the deeper tops typically from −15° to −20°C (5–6 km). Typical peak updrafts measured by aircraft were ~7 m s−1, with the maximum updraft being ~14 m s−1 and vertical wind shear being near 1.7 × 10−3 s−1.
b. ZDR column identification
An automated algorithm was created to objectively identify subsets of the NXPol observations for statistical analysis using identification of three-dimensional volumes of characteristic ZH and ZDR values. The general method used here is similar to prior work developed in the context of severe weather nowcasting, in which ZDR columns are used as proxies for updraft locations in severe thunderstorms [Snyder et al. 2015; see also prior work by Picca and Ryzhkov (2010), Picca et al. (2010), and Trömel et al. (2012)], although tuned for the smaller and shallower convective cells observed during COPE by adjusting and optimizing the ZH and ZDR thresholds discussed below. In addition, non-Rayleigh effects complicate the polarimetric signatures associated with larger hydrometeors because of the shorter-wavelength radar used in this study [illustrated, for example, using theoretical simulations and disdrometer measurements for radar parameters including ZDR and KDP by Ryzhkov and Zrnić (2005)]. However, while this complicates the interpretation of the observed radar signatures, the ensembles of particles sampled by the radar are still associated with signatures consistent enough to be useful for this study (e.g., Dolan and Rutledge 2009). An advantage of this approach is its dependence on an objective signature providing a known indicator of an active warm rain process early in the convective life cycle. It provides a simple method to statistically analyze multiple radar parameters summarizing the convective cells’ structure.
The ZDR column identification algorithm is applied here to radar measurements interpolated to a Cartesian grid using the National Center for Atmospheric Research’s Radx software package (https://www.ral.ucar.edu/projects/titan/docs/radial_formats/radx.html). The Cartesian grid spacing chosen here represents a balance between the high-resolution convective structures and the range-dependent spatial resolution of the measurements. In this case, the radar measurements were regridded to 250-m grid spacing (both horizontal and vertical) to retain the detailed structural characteristics of the convection, and the analyses were limited to measurements made within 60 km of the radar to avoid the effects of beam broadening.
The technique identifies ZDR columns as contiguous three-dimensional volumes of ZDR ≥ 1 dB and ZH ≥ 10 dBZ extending at least 500 m above the environmental 0°C level (estimated from UWKA measurements in regions without measured liquid water or hydrometeors). The identification is performed above the 0°C level because of the ambiguities associated with generally larger ZDR values commonly present from liquid hydrometeors below that level. To eliminate weaker, noisier echoes and focus on cells that contained adequately resolved, well-defined ZDR columns, each cell must also contain at least one grid point immediately above the environmental 0°C level featuring ZDR ≥ 2.5 dB, along with a minimum value of ZH ≥ 25 dBZ within the cell core (although not necessarily collocated with the maximum ZDR). Note that values of ZDR ≥ 1 dB are commonly used to identify ZDR columns, with the requirement of a grid point containing ZDR ≥ 2.5 dB used here to ensure the identification of well-defined column structures. In addition, no single set of ZH and ZDR thresholds would capture every ZDR column because of the large variability in structure between cases, even within different storms occurring on any particular day. For example, some columns with large ZDR values but relatively low ZH (i.e., entirely below 25 dBZ) were visually observed (e.g., on 2 August) but are not captured by these criteria because of the requirement of at least one grid point featuring ZH ≥ 25 dBZ. However, decreasing the minimum ZH value below 25 dBZ results in columns with poorly defined structure being identified. Thus, the ZH and ZDR thresholds used here are inherently limited but represent an attempt to find the best balance between identifying as many columns as possible while limiting the number of noisy or poorly organized cases. Several factors complicate the radar-indicated convective cell structure. Because vertical wind shear was present, the ZDR columns were often somewhat tilted. The storm motion also influenced the apparent vertical structure of the columns, an artifact that has been noted with deep convection particularly with fast storm motion (e.g., Snyder et al. 2015). During COPE, the NXPol volume scans started at the lowest elevation, with elevation angle increasing after each rotation. The convection propagated during this time, resulting in a continuous horizontal displacement of each cell during a given volume scan. This motion is superimposed on any tilt with height induced by vertical wind shear. Each PPI scan took approximately 30 s to complete at each elevation angle, so the displacement of the cells between each elevation in a given volume was relatively small. For example, given a cell motion of ~8 m s−1, this displacement would be ~250 m over the course of a scan at one elevation angle. The algorithm’s requirement of three-dimensional continuity for threshold ZH and ZDR values accounts for both possibilities, allowing it to identify vertically sheared columns and both the column maximum height above 0°C and extent of any shear or measurement-induced horizontal tilt, as quantified by the change in location of the ZDR-weighted centroid over the column height.
The column identification technique was also designed to deal with variations in cell structure and data quality occurring during COPE. In addition to depending on the specified Cartesian grid spacing and maximum usable range, the identification parameters were chosen to minimize data quality issues sometimes apparent in the measurements. These included polarimetric artifacts associated with multiple-body scattering in convective cores, nonmeteorological echoes, unrealistically large ZDR values (i.e., 6–7 dB or greater) along sharp ZH gradients near cloud edge, and attenuation along radials passing through strong convective cells. The quality control and correction procedures mitigated the issue of spurious echoes and attenuation. However, while the presence of multiple-body scattering potentially indicates the presence of hail, the signature’s occurrence in X-band measurements does not conclusively indicate hail rather than large drops (e.g., Wilson and Reum 1988; Lindley and Lemon 2007). In addition to the data correction, the algorithm’s ZH and ZDR thresholds were also chosen to mitigate these issues. However, even with these measures in place, these artifacts likely reduce the sample size from the actual number of relevant convective cells.
Some examples of results from the column identification algorithm are shown in Figs. 2–4 for measurements from 1415 UTC 2 Aug 2013. The line of convection at that time is apparent from the ZH and ZDR from the 6.5° elevation scan (Figs. 2a,b). The most southerly cell had developed ZDR values of at least 3–4 dB in its core. Figures 2c and 2d show the ZH and ZDR measurements after interpolation to the 250-m Cartesian grid at 3.25-km altitude (corresponding to the approximate altitude of the measurements in Figs. 2a,b); little detail in the ZH and ZDR features has been lost as a result of the interpolation, although the shallower cells to the north that appeared in the 6.5° scan do not all extend to this altitude. Figure 3 shows vertical cross sections of ZH and ZDR located along the dashed lines in Fig. 2. The ZDR column structure produced by the southernmost cell is visually evident, with maximum ZDR values of ~3.5 dB occurring over 1 km above the 0°C level. The UWKA made collocated observations, sampling the ZDR column directly at ~4.2-km altitude (near −6°C) and observed large liquid drops (up to 8-mm diameter) in this region (Leon et al. 2016). Figure 4a shows the corresponding column identification algorithm output, in terms of the ZDR column height above the environmental 0°C level [following, e.g., Picca et al. (2010) and Kumjian et al. (2014)], indicating ZDR > 1 dB occurring up to 2.5 km above 0°C.
The 6.5° elevation scans of (a) ZH and (b) ZDR, valid at 1415 UTC 2 Aug 2013, along with the corresponding measurements of (c) ZH and (d) ZDR interpolated to a 250-m Cartesian grid at 3.25-km altitude. Dashed lines correspond to the cross section shown in Fig. 3, below. Distances on the x and y axes are relative to the NXPol location.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
Vertical cross sections of (a) ZH and (b) ZDR along line A–B in Fig. 2, with x-axis distances corresponding to that line. The dash–dotted line represents the environmental 0°C altitude.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
(a) The height of the ZDR column above the environmental 0°C level, using objective column identification for the southernmost convective cell in Fig. 2. (b) The corresponding derived rain rate, interpolated to a Cartesian grid at 1.25-km altitude. Distances on the x and y axes are from the NXPol location.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
The algorithm also identifies other metrics describing structural characteristics, including shear-induced column tilt, maximum ZH and ZDR values with respect to altitude, and the corresponding radar-derived rain rates. The example convection’s rain rates are shown in Fig. 4b, for measurements interpolated to 1.25-km altitude. The dual-polarimetric rain-rate estimates are used, with data from 1.25-km height used to avoid underestimation due to beam blockage in the lowest elevation PPI scans. Although the estimated rain rates at this altitude may not equate to those at the surface (because of, e.g., evaporation), the choice of altitude ensures the gridded data is interpolated from elevation angles of at least 1.5° within 60-km range. Because of the uncertainties inherent in radar-derived rain rates, the derived rain rates are presented here to show relative differences of rainfall intensity rather than absolute quantitative precipitation estimates.
Because the column structure is identified using ZH and ZDR measurements above the environmental 0°C level, a vertical discontinuity exists between the identified column and the near-surface rainfall. Therefore, the correspondence between the column signatures and the rainfall below can be somewhat ambiguous because of the spatial disconnect and the time necessary for precipitation to fall. However, given the stringent criteria used to identify ZDR columns, it is assumed that the presence of an organized column signature indicates that precipitation production (and any direct enhancements related to an active warm rain process) is already occurring at the time of the measurements. While there may be a time lag between the ZDR column’s initial appearance in the early stages of growing convection and the most enhanced rainfall (e.g., Kumjian et al. 2014) the signatures identified here are typically well defined within active convective cells, suggesting that this is a reasonable approximation. This is also borne out by the precipitation enhancement often visually evident below the identified ZDR columns even in association with discrete, individual convective cells. In addition, the temporal resolution is on the same order as the time necessary for precipitation to descend from the 0°C level to ~1 km above the surface. While convective updrafts will loft some of the precipitation, the aircraft-observed updraft speeds discussed in the previous section represent the local maximum values, which were less than 10 m s−1 on average. The typical updraft speeds (again, relatively weak compared to deep convection) do not exceed fall velocities for the larger drops, allowing them to fall through updraft cores rather than being ejected from the core updrafts, allowing some precipitation to fall more directly to the surface. With relatively low environmental wind shear below the 0°C level, the horizontal transport of precipitation directly resulting from the identified convection would be relatively small. The analyses used here, which incorporate derived rain-rate values below and expanded horizontally by 0.5 km about the identified ZDR column (to account for any horizontal motion), are suitable to identify the intensity of the precipitation directly resulting from the identified convection in many cases. A limitation inherent in these analyses is the potential for stronger shear aloft (e.g., on 2 August) to transport some precipitation lofted in the updraft farther horizontally from the updraft core. With this limitation in mind, the results of these analyses as applied to COPE cases will be described in detail and contrasted with the corresponding derived rainfall statistics for the more general convective population in the following sections.
3. Results of statistical analyses
a. Statistical properties of ZDR columns
The column identification algorithm was applied to all radar scans from the 29 July, 2 August, and 3 August cases to obtain a dataset suitable for statistical analysis, with analysis techniques drawing from those applied by Rowe et al. (2011) to deeper monsoonal convection developing over the southwestern United States. Convection containing ZDR columns was present in each case both as relatively discrete cells (typically representative of initial convective development) and embedded in the more widespread convective lines that developed over time.
The ZDR columns (i.e., the volume with ZDR ≥ 1 dB) had an average height of ~1 km above the environmental 0°C level (solid line in Fig. 5a) but extended to nearly 4 km above 0°C in the deepest cases. Airborne measurements near the environmental 0°C level typically featured temperature perturbations of ~2°C or less in updrafts, indicating that subfreezing temperatures are likely through much of the ZDR column structure. Even though the COPE environments featured relatively weak instability and shallow convection compared to many prior studies (e.g., Van Den Broeke 2016), the microphysical structure resulting from the initial warm rain process (i.e., supercooled liquid drops and eventual mixed-phase hydrometeors) was substantial enough to commonly be observed several kilometers above the environmental freezing level. While the vertical wind shear in these cases was also relatively weak compared to prior studies of deep convection, some shear was present in each case, skewing convective cells horizontally with altitude (e.g., the ZDR column evident in the cell shown in Fig. 3). This was quantified by tracking the ZDR-weighted centroid location with respect to altitude within each column. As indicated by the dashed line in Fig. 5a, the horizontal “tilt” of the columns was generally small relative to the scale of the convection. The median horizontal displacement of the ZDR centroid location was 250 m over the height of the columns, and was nearly always ≤ 500 m, on the order of the maximum apparent displacement expected due to storm motion. Thus, while some precipitation will be lofted downstream, particularly in the higher-shear environments on 2 August, the largest precipitation particles (i.e., those measured by radar) will more likely remain localized near the cell core.
The (a) height (solid curve) and horizontal tilt with altitude (dashed curve) of the ZDR column (km). Also shown is the frequency distribution of maximum altitude above 0°C of (b) threshold ZH values in 10-dBZ increments and (c) threshold ZDR values in 0.5-dB increments.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
The general ZH and ZDR characteristics of the ZDR columns are also summarized in Fig. 5 using frequency distributions of the maximum height above the 0°C level at which threshold ZH (Fig. 5b) and ZDR (Fig. 5c) values occurred. The statistics shown here summarize only the structure above the 0°C level. The largest ZH values (50 dBZ) most commonly occurred within ~500 m of 0°C (Fig. 5b), although a few cases produced these large values up to 2-km height. For other threshold ZH values (20–40 dBZ), a slight increase in the most frequent heights was evident for decreasing ZH, as expected. The maximum altitude of these ZH values was most commonly centered near 1 km above 0°C, with a few cases extending up to 2.5 km. The corresponding frequency distributions for threshold ZDR values are shown in Fig. 5c. The most frequent height above 0°C for a given threshold ZDR increased more dramatically as the threshold value decreased. The distribution of heights for ZDR = 1 dB mirrors the general distribution of column height, as expected given the criteria used for column identification. Increasingly large ZDR values were more likely to occur closer to the 0°C level, consistent with the presence of the largest liquid drops near 0°C, with smaller supercooled drops and eventual freezing more likely at lower temperatures within the column (see also, e.g., Kumjian et al. 2014). Approximately 70% of the instances of ZDR ≥ 3 dB were confined to 500 m above the 0°C level.
The general characteristics of these distributions were broadly similar between the individual days, with the primary difference being that the deepest columns were more common in the stronger convection on 2 August, as discussed further below. The distributions were also relatively insensitive to the choice of minimum height allowed for columns extending above the 0°C level: similar frequency distributions were generated using a range of minimum column height thresholds from 250 m to 1 km above the 0°C level (i.e., one to four Cartesian grid points). While decreasing the height threshold allowed more columns to be identified, nearly all cells that developed a column structure reached 500 m or more above 0°C. The nature of the Cartesian interpolation used for this study likely had an impact on these analyses. Because several of the original data points were averaged together to create the value at a given Cartesian grid point, distinct features (e.g., local ZH and ZDR maxima) tend to be expanded somewhat in the Cartesian gridded output compared to the measurements in their original coordinate system. Because of this, as well as the range-dependent spatial width of the radar scans (up to ~1 km in the data used here), a local maximum in ZH or ZDR originally extending just above the 0°C level may be expanded to additional Cartesian grid points above that level. While such issues are inherent in this type of analysis, the focus on deeper, well-defined ZDR columns is expected to limit their effects.
Figure 6 summarizes the ZH and ZDR structure of the ZDR columns from another perspective, showing the frequency distribution of maximum ZH (Fig. 6a) and ZDR (Fig. 6b) values with respect to altitude above 0°C. Measurements are separated into 1-km height increments above 0°C, such that the measurements presented for a particular height interval include all columns reaching that height or greater. The maximum observed ZH values consistently decrease with altitude above 0°C, from ~60 dBZ near 0°C to ~50 dBZ at 3–4 km above 0°C. Similarly, with increasing height, the peak ZH value in each frequency distribution consistently decreases. The most frequent maximum ZH values near 0°C were 50–55 dBZ, decreasing to approximately 35 dBZ for the deepest columns (although it should be noted that relatively few cases contributed to this distribution, leading to the somewhat bimodal distribution of these ZH values).
Distribution of (a) maximum ZH and (b) maximum ZDR values within all ZDR columns, with respect to altitude above environmental 0°C level, in 1-km increments.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
Some corresponding trends are also evident in the distributions of maximum ZDR values shown in Fig. 6b. Most frequently, the largest ZDR values occur within the lowest kilometer above 0°C, confirming the trends shown in Fig. 5c. The requirement of a grid point with ZDR ≥ 2.5 dB near 0°C artificially constrains the lower end of this distribution, such that the maximum frequency (and lowest ZDR value) is in the 2.5–3-dB bin. Even so, the largest ZDR values (≥4 dB) are most commonly evident within the lowest kilometer above 0°C. A gradual decrease in maximum ZDR is also evident with increasing height above 0°C, with no values > 3 dB near the tops of the deepest columns. Maximum ZDR values near or below 2 dB were also most frequent for all columns exceeding 1 km in height, although again the minimum 2.5-dB constraint on measurements below this height complicates this trend.
Figure 7 presents scatterplots of ZH versus ZDR with respect to measurement height above the environmental 0°C level for all Cartesian grid points within the identified ZDR columns, more clearly indicating the microphysical characteristics described above. Measurements are shown separately in this figure for each of the three COPE cases analyzed here. Reflectivity values above 50 dBZ were present in each case. Similarly large ZH values were evident in other COPE cases in which the warm rain process dominated (e.g., in convection capped near the 0°C level; Leon et al. 2016). On 2 August (Fig. 7a), the ZH–ZDR pairs were most clustered within ~35–55 dBZ and ~2–4 dB for the shallower measurements (indicated by the color shading), particularly for columns extending to only ~1 km above the 0°C level. Both ZH and ZDR gradually trend to lower values for increasingly deep columns, although significant variability is present. This is consistent with the ZDR columns being associated with the largest liquid drops near the 0°C level, and combinations of smaller drops and ice at lower temperatures [see also Fulton and Heymsfield (1991) and Kumjian et al. (2014)]. Airborne measurements in the COPE convection also indicated a complex microphysical environment. Mixed-phase conditions were commonly observed, with graupel and liquid drops (some large) often coexisting above the 0°C level (Leon et al. 2016; Taylor et al. 2016; Jackson et al. 2018). The UWKA also sampled directly within multiple ZDR columns, supporting this interpretation of the radar signatures. Measurements in columns at temperatures near −6° and −9°C verified the presence of liquid drops in these clouds, with the largest drops were present at the warmer temperatures, as discussed above.
Scatterplot of ZH vs ZDR, with color shading indicating measurement height above 0°C for gridpoints within ZDR columns, for (a) 2 Aug, (b) 29 Jul, and (c) 3 Aug 2013.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
Additional information can also be gained from the context of prior studies examining ZH versus ZDR characteristics of subtropical and midlatitude convection over the United States (Zhang et al. 2006; Cao et al. 2008; Carr et al. 2017). While exact comparisons are complicated because of the increased potential for increased differential attenuation and non-Rayleigh effects in the NXPol measurements relative to the S-band measurements used in these prior studies (e.g., Ryzhkov and Zrnić 2005), there is overlap in the ZH–ZDR parameter space between the COPE measurements and the other studies. The COPE observations feature more variability and are generally distributed over larger ZDR values for a given ZH interval, relative to the past studies. This would be consistent with a general bias toward somewhat larger drop sizes in the ZDR columns sampled in the COPE dataset relative to past studies, noting that these measurements are biased toward measurements in convective updrafts. This is also suggested by the overlap in ZH–ZDR space with the subset of measurements containing rain dominated by “big drops” in Cao et al. (2008).
The deepest convection during COPE occurred on 2 August, along with the most frequent and deepest ZDR columns, as evident in the color shading in Fig. 7a. Although measurements from the other two days (Figs. 7b,c) consist almost entirely of data points within ZDR columns that extended only up to 2 km above 0°C, similar clustering of ZH and ZDR values was evident on these days compared to the measurements in the shallowest columns that developed on 2 August. From this perspective, the main difference on 2 August is evidently the convection’s strength, consistent with the increased instability relative to the other two cases. The microphysical development, at least as gauged by these radar variables, appears consistent among the three cases; that is, larger ZDR values do not necessarily occur at different heights relative to the ambient 0°C level despite differences in the strength of the convection. The spread of paired ZH–ZDR values is consistent with various mixtures of liquid and ice hydrometeors occurring above the environmental 0°C level among all the observed convective cells, consistent with airborne measurements at subfreezing temperatures.
Column area was also analyzed with respect to height above 0°C to examine whether the convection exhibited any distinct modes (e.g., shallower, larger areas of enhanced ZDR as compared with deeper but narrower columns). Figure 8 quantifies the ZDR column area with respect to height above the 0°C level, for all measurements in ZDR columns occurring on 2 August. In each panel, each line tracks the spatial area of an individual column with respect to height above 0°C, providing a more detailed representation of the individual column characteristics as compared to the preceding analyses. The panels are sorted by maximum column height, with Figs. 8a, 8b, and 8c showing columns extending to ≤1 km, 1–2 km, and ≥2 km above 0°C, respectively. The median area with respect to height above 0°C is indicated by the thick blue line overlaid in each panel, and the thinner blue lines represent the corresponding 25th- and 75th-percentile area values with height. The measurements from 2 August best illustrate the overall trends observed in all cases. The median and quartile area values immediately above the 0°C level increase somewhat from the shallowest columns (Fig. 8a) to those reaching 1–2 km above 0°C (Fig. 8b). However, the difference between each subset is relatively small in comparison with the overall range of area values at these heights, and the median and 75th-percentile area values near 0°C actually decrease somewhat for columns exceeding 2 km in height (Fig. 8c). In addition, the outliers with larger spatial extent near the 0°C level are not necessarily the deepest columns within each subset of data. Tracking the characteristics of individual columns, the cases that developed spatial areas close to the median value near 0°C more commonly extended to the largest heights in each case. In other words, there was not a strong association between area outliers (either large or small) and column height. This suggests that broader convective updrafts (and by inference, columns with larger area near 0°C) were less likely to grow to higher altitudes. However, this cannot be concluded definitively without corresponding in situ measurements of updraft size and velocity in most cells. Analyses of ZDR columns on 29 July and 3 August show similar signatures, again with no significant correlation between column area and height.
Area of individual ZDR columns on 2 Aug 2013 with respect to height above the environmental 0°C level, for maximum column heights of (a) 0–1, (b) 1–2, and (c) >2 km above 0°C. Individual lines correspond to each column, with the color shading indicating the maximum column height. The thick blue line corresponds to the median area, with the thin blue lines representing the 25th- and 75th-percentile area values.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
b. Rainfall properties of cells containing ZDR columns
Ultimately, the analyses of convection containing ZDR columns are designed to identify whether the distinct microphysical characteristics inferred from radar signatures (i.e., lofting of supercooled raindrops and, ultimately, enhanced hydrometeor growth via coalescence or accretion) are associated with quantifiable differences in rain production, compared to the regions of convection that do not feature enhanced ZDR above the 0°C level. Given the range of convective modes sampled during COPE, this could be identified in terms of a significant contribution to the total rainfall occurring in a given radar scan volume, or more locally, consistently producing the largest rain rates. To investigate this, rainfall rates were derived from the dual-polarization radar observations and statistically aggregated for both the convection with ZDR columns and the more general convective population not containing ZDR columns. Representative examples are shown here for two general modes of convection sampled during COPE: localized, relatively discrete convective cells (i.e., individual updraft structures) occurring over a small portion of the radar domain, and spatially larger complexes containing multiple embedded convective cells. Convective development in both modes occurred during the three days analyzed in this study, with the lifetime of individual cells and ZDR columns relatively short compared to that observed in studies focusing on deeper and stronger convection. The approach used for this analysis represents an instantaneous measurement of the precipitation characteristics of convective cells with ZDR columns at a given time. It is not intended to capture the impacts that warm rain may have on later precipitation development due to processes such as recirculation of hydrometeors that evolve over several tens of minutes. For example, an initial thermal may produce large supercooled drops (and a ZDR column structure), but subsequent convective development may also involve the recirculation of related hydrometeors as well as the increasing importance of ice processes. The resulting complex structure as measured by radar would evolve between subsequent radar volume scans, making a simple diagnosis of the microphysics and precipitation development beyond the scope of this study.
Rain-rate characteristics are shown in Fig. 9 for two radar scans in which relatively small convective cells developed. Figures 9a and 9b show ZH at 3.25-km altitude, along with the corresponding derived R at 1.25-km altitude, for measurements valid at 1401 UTC 2 August 2013. Figures 9c and 9d show corresponding ZH and R values from a later scan at 1415 UTC 2 August. Rain rates at 1.25-km altitude are used to minimize the impact of beam blockage, but they may not always capture the most intense precipitation below a particular ZDR column due to vertical changes in reflectivity below 1.25 km. Regardless, in both examples shown here, the most intense convection at the times shown encompasses a relatively small portion of the radar domain, with one cell in each case (outlined in red) producing enhanced positive ZDR values (not shown) above the 0°C level. The other, lower-reflectivity regions with no ZDR columns were mostly decaying cells and associated stratiform regions that may or may not have previously contained strong updrafts and ZDR columns. In addition to developing the largest ZDR values in each scan (not shown), the cells with ZDR columns are visually associated with the largest ZH values measured here. The largest derived rain rates are also evident in association with these cells. Maximum rain rates associated for the cells with ZDR columns in each case are ~30 mm h−1 (Fig. 9b) and ~20 mm h−1 (Fig. 9d). Nearly all rain-rate values elsewhere in the two domains are ≤ 5 mm h−1, aside from the decaying region due west of the intense rain in Fig. 9b that previously had reflectivity values of about 55 dBZ. These characteristics are similar to other times in which similarly small-scale, relatively discrete convection developed over the three COPE cases. At these times, the cells containing ZDR columns were typically visually associated with most of the enhanced rainfall rates occurring in each scan.
Cartesian-gridded NXPol measurements of (a),(c) radar reflectivity at 3.25-km altitude and (b),(d) derived rain rate at 1.25-km altitude from the scans valid at (top) 1401 UTC and (bottom) 1415 UTC 2 Aug 2013. Convective cells developing ZDR columns are circled in red. Note the difference in spatial scales between the top and bottom panels.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
This is better quantified in Fig. 10 for the two example cases using cumulative distribution functions (CDFs) of the estimated rain rates interpolated to 1.25-km altitude, over the entire domain within a given volume scan. The CDFs shown in dark gray represent the summed distribution of rain rates for the cells containing ZDR columns within each scan volume, and the light-gray CDFs show the rain rates for all other convection in the same volume. Similar signatures are evident in both radar volumes, with the main difference being fewer grid points containing precipitation at 1401 UTC (Fig. 10a) than at 1415 UTC (Fig. 10b), a direct result of the increasing spatial extent of the precipitation over the intervening time between radar scans. In both cases, the greater rainfall rates are associated with the cells containing ZDR columns. Correspondingly, the median rainfall rate for the cells with ZDR columns in each scan was ~15–20 mm h−1, as compared with less than 5 mm h−1 for those cells not containing a ZDR column. The cells with ZDR columns in these two examples also contributed a majority of the total rainfall occurring within the two scans. These characteristics are representative of discrete cells occurring at other times during the three days from COPE analyzed here. When the convection was predominantly characterized by relatively small, discrete cells (typically relatively early on each day), the cells featuring ZDR columns were consistently associated with locally enhanced precipitation rates, as well as being significant contributors to the total rainfall occurring over the entire radar domain. Thus for the discrete convective cells occurring during COPE, a strong warm rain process (as indicated by the ZDR column structure), appeared necessary for producing heavier rainfall rates (exceeding ~25 mm h−1).
Cumulative distribution functions of derived rain rate throughout the radar domain corresponding to the times shown in Fig. 9, with values for cells containing ZDR columns shown in dark gray and for the remaining precipitation in light gray. The measurements are valid at (a) 1401 UTC and (b) 1415 UTC 2 Aug 2013.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
Similar plots of radar reflectivity and derived rain rate are shown in Fig. 11 for two other radar scans from 2 August, for measurements from 1607 UTC (Figs. 11a,b) and 1617 UTC (Figs. 11c,d). These examples are representative of cases when the convection was more widespread (as the convergence line became more developed). Commonly, several cells embedded within a larger region of cloud had identifiable ZDR columns. Examples of these cells are again outlined in red. Relative to the discrete cells in Fig. 9, the convective structure (and in particular, the resulting rainfall) is much more complex. While the cells with ZDR columns are visually identifiable as being associated with local maxima in both reflectivity and rain rate, numerous other cells embedded within the complex also have enhanced values of both parameters. While much of the convective complex is associated with rainfall rates below ~5 mm h−1, many rates ≥ 30 mm h−1 are associated with the stronger embedded convective cells, regardless of whether or not they contained ZDR columns. Thus, for the more widespread convection considered here, a strong warm rain process (as indicated by a ZDR column) still produces heavy rainfall, but other factors such as ice processes or microphysical interactions among different convective cells (e.g., a seeder–feeder mechanism) can produce equally heavy rainfall. Although the much larger spatial extent of the convection decreases the overall rainfall contribution from the cells containing ZDR columns relative to the total rainfall, the cells with ZDR columns do produce most of the largest rain rates, approximately 80 mm h−1 at a few grid points in each case (Fig. 12). The median rainfall rate for the cells with ZDR columns was ~40 mm h−1 in both examples, as compared with ~15 mm h−1 for the more general convection. These examples are consistent with those occurring at other times during the three COPE days analyzed here when precipitation development was characterized by large convective complexes.
As in Fig. 9, but showing measurements from radar scans valid at (a),(b) 1607 UTC and (c),(d) 1617 UTC 2 Aug 2013.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
As in Fig. 10, but for measurements valid at (a) 1607 UTC and (b) 1617 UTC 2 Aug 2013.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
To determine whether the individual ZDR column properties might be good predictors of the associated rainfall rates, the mean and maximum rainfall rates and area-integrated rainfall for each of the cells containing ZDR columns were compared with the maximum ZH and ZDR values, the ZDR column height above 0°C, and the total column volume, both for the individual days and for the entire dataset. None of the individual column properties was a strong predictor of the resultant estimated rainfall, with the largest correlation coefficients on the order of ~0.5 for column volume and integrated rainfall. Because the rainfall rates represent instantaneous estimates from the radar perspective rather than cumulative values measured at the surface, the relation between ZDR column characteristics and associated precipitation may be obscured. Despite the enhancement in precipitation commonly observed simultaneously with the column structure, lag time between the columns’ initial development and the most enhanced precipitation may introduce some uncertainty here. In addition, enough variability exists between cases that any individual column property is limited as a predictor of heavy rainfall. On the other hand, the low correlations might also indicate that focusing on the strength and area of an active warm rain process, as indicated by the characteristics of the ZDR columns, misses the important influence of ice processes upon the precipitation falling to the ground. Overall, however, the results suggest that the presence of a ZDR column itself (i.e., following from a strong initial warm rain process) has the strongest association with the development of heavy rainfall in the COPE clouds.
To examine this association further, the overall characteristics of rainfall associated with cells containing ZDR columns were quantified by obtaining statistics from CDF analyses similar to those shown in Figs. 10 and 12, but drawn from the full set of radar measurements from all three days. These CDFs were first developed separately for each radar scan, summarizing the contribution to the total rainfall occurring in that scan separately for the cells with ZDR columns and the more general precipitation. Then, the 5th-, 25th-, 50th-, 75th-, and 95th-percentile rain-rate values were obtained from each CDF, giving these five percentile values for each radar scan, again separately for the cells with ZDR columns and for the general convective population. The distributions of these percentile values summarize the distribution of rainfall values over all radar scans from the three days.
The results of these analyses are shown as box-and-whisker plots in Fig. 13. Each box-and-whisker plot summarizes the distribution of a particular percentile rain-rate value over all radar scans. Distributions are again shown in dark gray for cells with ZDR columns and in light gray for the more general convection. Some similarly low rain rates (<5 mm h−1) occurred whether or not ZDR columns were present, but the range of rain rates is consistently largest for the cells with ZDR columns. In addition, the cells with ZDR columns are consistently associated with the largest rates in each case. The median and quartile values are the more statistically robust attributes of the distributions, particularly when (as for the precipitation shown in Figs. 9a,b and 10) fewer data points contributed to a particular distribution. The median and quartile rainfall rates for cells with ZDR columns are statistically larger than for the general population for all but the lower quartile values of the heaviest precipitation (the 95th-percentile rain rates in a given scan), emphasizing their importance in producing heavy rainfall. This was the case both for discrete cells developing in relative isolation and for clouds embedded in longer-lasting convective complexes. The statistical trends were also broadly similar when the measurements were separated by day, indicating some consistency for convection developing in a variety of meteorological environments.
Distribution of statistical percentile rain-rate values from all radar scans on 29 Jul, 2 Aug, and 3 Aug 2013, with values for cells containing ZDR columns shown in dark gray and for the remaining precipitation in light gray. Each box-and-whisker plot shows the 5th-, 25th-, 50th-, 75th-, and 95th-percentile values of the corresponding percentile rain-rate values for all radar scans, derived from distributions constructed as shown in Figs. 10 and 12.
Citation: Journal of Applied Meteorology and Climatology 57, 11; 10.1175/JAMC-D-17-0134.1
4. Discussion and conclusions
This study builds upon prior research involving the ZDR column structure as observed by radar in warm-based convection. Based upon early radar-based studies, and confirmed by subsequent airborne measurements and modeling studies, this signature is used as an indicator of convection in which the warm rain process (collision and coalescence producing liquid drops) is important for precipitation development, both directly and as a source of liquid for accretional growth. This study specifically investigates the relation between convection containing the ZDR column structure and the production of heavy precipitation, using radar measurements from the COPE field campaign (Leon et al. 2016). While the possibility of non-Rayleigh scattering complicating the polarimetric signatures used here is likely due to the use of an X-band radar (e.g., Ryzhkov and Zrnić 2005), the ensembles of particles sampled are expected to be associated with signatures consistent enough to be useful (e.g., Dolan and Rutledge 2009).
This work is distinct from prior studies in several key areas. The occurrence of ZDR columns was unexpectedly frequent in light of early studies (e.g., Hall et al. 1984; Caylor and Illingworth 1987; Illingworth et al. 1987) that sampled convection with ZDR columns over southern England; these studies were also more limited in terms of measurements and number of cases. More recent studies (e.g., Rowe et al. 2011; Kumjian et al. 2012, 2014; Van Den Broeke 2016; van Lier-Walqui et al. 2016) focus on convection in environments with higher instability and wind shear, analyzing ZDR columns in much deeper, longer-lived, and more organized convective cells. The observations presented in this study, then, are notable in that they show a consistent occurrence of the ZDR column structure in shallower and shorter-lived convection developing in relatively low-shear, low-instability environments (CAPE typically ~100–300 J kg−1, and ~600 J kg−1 at maximum). Furthermore, the number of cases sampled during COPE provides a large enough set of radar observations to be suitable for statistical analysis, with the radar observations also supplemented by airborne measurements of cloud properties.
After developing objective ZDR column identification criteria, the structural characteristics of convective cells featuring ZDR columns were investigated for three cases during the COPE campaign. The cases included ZDR columns both in individual, relatively discrete convective cells and embedded in larger-scale convective complexes. In contrast to deep convection producing ZDR columns in higher-instability environments, cloud tops were much shallower, commonly reaching 4–6 km with the deepest clouds reaching ~7–8 km. The mean peak updraft speeds were less than 10 m s−1 and the maximum observed updraft speed was ~18 m s−1. Cloud bases were typically near 10°C, with column height (identified using ZDR > 1 dB and ZH > 10 dBZ) averaging 1–1.25 km above the environmental 0°C level and extending to nearly 4 km above 0°C in a few cases. Correspondingly, in situ measurements near the environmental 0°C level commonly showed temperature perturbations of ~2°C or less when passing through updrafts, so subfreezing temperatures are expected through a majority of the columns. Columns featuring ZDR ≥ 3 dB were present, with the largest ZDR values most common in the lowest 500 m above 0°C. Maximum ZDR values decreased with increasing height above 0°C, consistent with glaciation and ice growth at colder temperatures. Airborne measurements that directly sampled the ZDR column structures indicated microphysically complex, mixed-phase conditions above the 0°C level, with large liquid drops observed in some instances (Leon et al. 2016; Taylor et al. 2016; Jackson et al. 2018). Airborne measurements in general were consistent with the ZDR columns consisting of a variety of liquid or (particularly at lower temperatures) mixed-phase hydrometeors (e.g., Kumjian et al. 2014). No correlation was found between column area near 0°C and total column height. Using the ZDR column structure as a proxy for the convective updraft in the absence of direct updraft measurements, broad updrafts were not deeper overall than narrow updrafts (Fig. 8). Other than the deepest cases generally occurring on 2 August, no substantial differences were identified in the ZDR columns developing in the varying thermodynamic environments on the other two days.
A major focus of the analyses of cells containing ZDR columns was to identify distinct characteristics of the precipitation they produced (as estimated from the radar measurements), and to contrast this against the precipitation characteristics of the more general convective population. While some uncertainty exists due to a potential lag between the initial appearance of the column signature and the most enhanced rainfall at the surface (e.g., Kumjian et al. 2014), enhanced rainfall was commonly apparent in this dataset by the time a given column signature was identified. When the convective mode was characterized by relatively discrete convective cells, the cells with ZDR columns were associated with higher derived rainfall rates at those locations. The examples in Figs. 2, 4, and 9 are typical, showing the most intense derived rainfall rates in association with these cells. Within larger-scale complexes of cloud and precipitation where some embedded convective cells also contained ZDR columns, these cells were also typically associated with enhanced derived rainfall rates (as evident in the examples shown in Fig. 11), but they were not the sole locations of higher rainfall rates. For both convective modes, CDFs were constructed for the estimated rainfall occurring over the entire radar domain (Figs. 10 and 12). These showed that the cells with ZDR columns consistently produced the largest estimated rainfall rates occurring within the radar domain (Figs. 9 and 11). Overall, the presence of the ZDR column signature (with its implications on the growth of liquid drops and their availability for subsequent accretional growth) was the strongest indicator of larger rainfall rates, rather than any individual characteristic of the columns.
To examine these signatures over the full dataset, the statistical characteristics of the derived rainfall were also analyzed for all radar measurements obtained during the three COPE cases. This further confirmed the importance of the cells with ZDR columns as contributors to heavy precipitation. As was visually evident for the individual radar scan volumes shown in Figs. 10 and 12, these cells and the convection in general produced a large range of estimated rainfall rates. However, after aggregating all of the radar measurements across all three days, the upper statistical percentile values for rainfall associated with the cells containing ZDR columns were consistently larger than for the other convection over the full set of radar measurements (Fig. 13). The results indicate the importance of these cells as producers of heavy precipitation, even when they did not dominate the total precipitation occurring over the radar domain (i.e., in the larger-scale convective complexes that developed over time). As indicated by the increased maximum derived rainfall rates and the statistically larger median and 75th-percentile values, the statistical analyses consistently showed the cells with ZDR columns to be associated with locally enhanced, potentially heavy precipitation, over the entire set of radar measurements used in this study. In addition, it is noted that the results may represent an underestimate of the association of cells containing ZDR columns with the production of heavier precipitation. Particularly in environments with stronger environmental wind shear (e.g., 2 August), some precipitation transported aloft in convective updrafts may be displaced farther horizontally from the core updraft. Even with this caveat, the association between convection containing ZDR columns and enhanced precipitation is consistently evident, both when aggregating all cases together and in the measurements made during each case.
The results presented here provide information about precipitation production in convective cells containing ZDR columns (compared to those without ZDR columns) captured at an instant in time, or at most, integrated over the time required to complete one radar volume scan. Therefore, any precipitation enhancement from the warm rain process that occurs relatively rapidly in time should be captured. However, influences on precipitation production that might occur over longer time scales, particularly after a ZDR column collapses, will not be captured by this methodology. Put another way, it is difficult to assess precipitation from cells that at one time, but no longer, contain ZDR columns. Further investigation into cells containing ZDR columns resulting from warm rain development and how that may impact overall precipitation production throughout a cell’s lifetime requires additional modeling studies. The work presented here can serve as a constraint to such studies. Ultimately, such modeling studies would then help to describe the evolving microphysical structure and processes that are not resolvable using observational measurements alone.
Acknowledgments
This work was funded by the U.S. National Science Foundation under Grants AGS-1230292 and AGS-1230203 for the U.S. investigators, with UWKA participation funded by Grant AGS-1441831. The work was also partly funded by the U.K. Natural Environment Research Council under Grant NE/J023507/1. We acknowledge the Centre for Environmental Data Analysis for storing archived COPE data and the NCAS Atmospheric Measurement Facility for use of the NXPol. We also acknowledge John Nicol for his significant role in assessing and calibrating the NXPol dataset and Mike Dixon for assistance working with the radar datasets and the Radx software.
REFERENCES
Adachi, A., T. Kobayashi, H. Yamauchi, and S. Onogi, 2013: Detection of potentially hazardous convective clouds with a dual-polarized C-band radar. Atmos. Meas. Tech., 6, 2741–2760, https://doi.org/10.5194/amt-6-2741-2013.
Beard, K. V., D. B. Johnson, and D. Baumgardner, 1986: Aircraft observations of large raindrops in warm, shallow, convective clouds. Geophys. Res. Lett., 13, 991–994, https://doi.org/10.1029/GL013i010p00991.
Bennett, L., 2017: MICROSCOPE: NCAS mobile X-band radar scan data from Davidstow Airfield. NCAS British Atmospheric Data Centre, accessed 12 May 2017, https://doi.org/10.5072/4bb383b7d6ca421bbedd57b8097d5664.
Blyth, A. M., and J. Latham, 1993: Development of ice and precipitation in New Mexican summertime cumulus clouds. Quart. J. Roy. Meteor. Soc., 119, 91–120, https://doi.org/10.1002/qj.49711950905.
Blyth, A. M., R. E. Benestad, P. R. Krehbiel, and J. Latham, 1997: Observations of supercooled raindrops in New Mexico summertime cumuli. J. Atmos. Sci., 54, 569–575, https://doi.org/10.1175/1520-0469(1997)054<0569:OOSRIN>2.0.CO;2.
Blyth, A. M., J. H. Lowenstein, Y. Huang, Z. Cui, S. Davies, and K. S. Carslaw, 2013: The production of warm rain in shallow maritime cumulus clouds. Quart. J. Roy. Meteor. Soc., 139, 20–31, https://doi.org/10.1002/qj.1972.
Braham, R. R., Jr., 1964: What is the role of ice in summer rain-showers? J. Atmos. Sci., 21, 640–645, https://doi.org/10.1175/1520-0469(1964)021<0640:WITROI>2.0.CO;2.
Braham, R. R., Jr., 1968: Meteorological bases for precipitation development. Bull. Amer. Meteor. Soc., 49, 343–353, https://doi.org/10.1175/1520-0477-49.4.343.
Brandes, E. A., J. Vivekanandan, J. D. Tuttle, and C. J. Kessinger, 1995: A study of thunderstorm microphysics with multiparameter radar and aircraft observations. Mon. Wea. Rev., 123, 3129–3143, https://doi.org/10.1175/1520-0493(1995)123<3129:ASOTMW>2.0.CO;2.
Bringi, V. N., D. A. Burrows, and S. M. Menon, 1991: Multiparameter radar and aircraft study of raindrop spectral evolution in warm-based clouds. J. Appl. Meteor., 30, 853–880, https://doi.org/10.1175/1520-0450(1991)030<0853:MRAASO>2.0.CO;2.
Bringi, V. N., L. Liu, P. C. Kennedy, V. Chandrasekar, and S. A. Rutledge, 1996: Dual multiparameter radar observations of intense convective storms: The 24 June 1992 case study. Meteor. Atmos. Phys., 59, 3–31, https://doi.org/10.1007/BF01031999.
Bringi, V. N., K. Knupp, A. Detwiler, L. Liu, I. J. Caylor, and R. A. Black, 1997: Evolution of Florida thunderstorms during the Convection and Precipitation Electrification Experiment: The case of 9 August 1991. Mon. Wea. Rev., 125, 2131–2160, https://doi.org/10.1175/1520-0493(1997)125<2131:EOAFTD>2.0.CO;2.
Brown, P. R. A., and H. A. Swann, 1997: Evaluation of key microphysical parameters in three-dimensional cloud-model simulations using aircraft and multiparameter radar data. Quart. J. Roy. Meteor. Soc., 123, 2245–2275, https://doi.org/10.1002/qj.49712354406.
Cao, Q., G. Zhang, E. Brandes, T. Schuur, A. Ryzhkov, and K. Ikeda, 2008: Analysis of video disdrometer and polarimetric radar data to characterize rain microphysics in Oklahoma. J. Appl. Meteor. Climatol., 47, 2238–2255, https://doi.org/10.1175/2008JAMC1732.1.
Carr, N., P. E. Kirstetter, J. J. Gourley, and Y. Hong, 2017: Polarimetric signatures of midlatitude warm-rain precipitation events. J. Appl. Meteor. Climatol., 56, 697–711, https://doi.org/10.1175/JAMC-D-16-0164.1.
Caylor, I. J., and A. J. Illingworth, 1987: Radar observations and modelling of warm rain initiation. Quart. J. Roy. Meteor. Soc., 113, 1171–1191, https://doi.org/10.1002/qj.49711347806.
Cooper, W. A., and R. P. Lawson, 1984: Physical interpretation of results from the HIPLEX-1 experiment. J. Climate Appl. Meteor., 23, 523–540, https://doi.org/10.1175/1520-0450(1984)023<0523:PIORFT>2.0.CO;2.
Diederich, M., A. Ryzhkov, C. Simmer, P. Zhang, and S. Trömel, 2015: Use of specific attenuation for rainfall measurement at X-band radar wavelengths. Part II: Rainfall estimates and comparison with rain gauges. J. Hydrometeor., 16, 503–516, https://doi.org/10.1175/JHM-D-14-0067.1.
Dolan, B., and S. A. Rutledge, 2009: A theory-based hydrometeor identification algorithm for X-band polarimetric radars. J. Atmos. Oceanic Technol., 26, 2071–2088, https://doi.org/10.1175/2009JTECHA1208.1.
Doviak, R. J., and D. S. Zrnić, 1993: Doppler Radar and Weather Observations. Academic Press, 562 pp.
Dufton, D. R. L., 2016: Quantifying uncertainty in radar rainfall estimates using an X-band dual polarisation weather radar. Ph.D. dissertation, University of Leeds, 207 pp., http://etheses.whiterose.ac.uk/15486/.
Dufton, D. R. L., and C. G. Collier, 2015: Fuzzy logic filtering of radar reflectivity to remove non-meteorological echoes using dual polarization radar moments. Atmos. Meas. Tech., 8, 3985–4000, https://doi.org/10.5194/amt-8-3985-2015.
French, J. R., J. H. Helsdon, A. G. Detwiler, and P. L. Smith, 1996: Microphysical and electrical evolution of a Florida thunderstorm. 1. Observations. J. Geophys. Res., 101, 18 961–18 977, https://doi.org/10.1029/96JD01625.
Fulton, R., and G. M. Heymsfield, 1991: Microphysical and radiative characteristics of convective clouds during COHMEX. J. Appl. Meteor., 30, 98–116, https://doi.org/10.1175/1520-0450(1991)030<0098:MARCOC>2.0.CO;2.
Golding, B. W., P. Clark, and B. May, 2005: Boscastle flood: Meteorological analysis of the conditions leading to flooding on 16 August 2004. Weather, 60, 230–235, https://doi.org/10.1256/wea.71.05.
Gourley, J. J., A. J. Illingworth, and P. Tabary, 2009: Absolute calibration of radar reflectivity using redundancy of the polarization observations and implied constraints on drop shapes. J. Atmos. Oceanic Technol., 26, 689–703, https://doi.org/10.1175/2008JTECHA1152.1.
Hall, M. P. M., J. W. F. Goddard, and S. M. Cherry, 1984: Identification of hydrometeors and other targets by dual-polarization radar. Radio Sci., 19, 132–140, https://doi.org/10.1029/RS019i001p00132.
Hallett, J., and S. C. Mossop, 1974: Production of secondary ice particles during the riming process. Nature, 249, 26–28, https://doi.org/10.1038/249026a0.
Helmus, J., and S. Collis, 2016: The Python ARM Radar Toolkit (Py-ART), a library for working with weather radar data in the Python programming language. J. Open Res. Softw., 4, e25, https://doi.org/10.5334/jors.119.
Herzegh, P. H., and A. R. Jameson, 1992: Observing precipitation through dual-polarization radar measurements. Bull. Amer. Meteor. Soc., 73, 1365–1376, https://doi.org/10.1175/1520-0477(1992)073<1365:OPTDPR>2.0.CO;2.
Houze, R. A., 1977: Structure and dynamics of a tropical squall-line system. Mon. Wea. Rev., 105, 1540–1567, https://doi.org/10.1175/1520-0493(1977)105<1540:SADOAT>2.0.CO;2.
Huang, Y., and Coauthors, 2008: The development of ice in a cumulus cloud over southwest England. New J. Phys., 10, 105021, https://doi.org/10.0188/1367-2630/10/10/105021.
Illingworth, A. J., W. F. Goddard, and S. M. Cherry, 1987: Polarization radar studies of precipitation development in convective storms. Quart. J. Roy. Meteor. Soc., 113, 469–489, https://doi.org/10.0112/qj.49711347604.
Jackson, R., J. R. French, D. C. Leon, D. M. Plummer, S. Lasher-Trapp, A. M. Blyth, and A. Korolev, 2018: Observations of the microphysical evolution of convective clouds in southwest United Kingdom. Atmos. Chem. Phys. Discuss., https://doi.org/10.5194/acp-2018-437, in press.
Jameson, A., M. Murphy, and E. Krider, 1996: Multiple-parameter radar observations of isolated Florida thunderstorms during the onset of electrification. J. Appl. Meteor., 35, 343–354, https://doi.org/10.1175/1520-0450(1996)035<0343:MPROOI>2.0.CO;2.
Johnson, D. B., 1982: The role of giant and ultragiant aerosol particles in warm rain initiation. J. Atmos. Sci., 39, 448–460, https://doi.org/10.1175/1520-0469(1982)039<0448:TROGAU>2.0.CO;2.
Johnson, D. B., 1987: On the relative efficiency of coalescence and riming. J. Atmos. Sci., 44, 1671–1680, https://doi.org/10.1175/1520-0469(1987)044<1671:OTREOC>2.0.CO;2.
Keenan, T., and Coauthors, 2000: The Maritime Continent Thunderstorm Experiment (MCTEX): Overview and some results. Bull. Amer. Meteor. Soc., 81, 2433–2456, https://doi.org/10.1175/1520-0477(2000)081<2433:TMCTEM>2.3.CO;2.
Kennedy, P. C., S. A. Rutledge, W. A. Petersen, and V. N. Bringi, 2001: Polarimetric radar observations of hail formation. J. Appl. Meteor., 40, 1347–1366, https://doi.org/10.1175/1520-0450(2001)040<1347:PROOHF>2.0.CO;2.
Knight, C. A., 2006: Very early formation of big, liquid drops revealed by ZDR in continental cumulus. J. Atmos. Sci., 63, 1939–1953, https://doi.org/10.1175/JAS3721.1.
Koenig, L., 1963: The glaciating behavior of small cumulonimbus clouds. J. Atmos. Sci., 20, 29–47, https://doi.org/10.1175/1520-0469(1963)020<0029:TGBOSC>2.0.CO;2.
Kumjian, M. R., and A. V. Ryzhkov, 2008: Polarimetric signatures in supercell thunderstorms. J. Appl. Meteor. Climatol., 47, 1940–1961, https://doi.org/10.1175/2007JAMC1874.1.
Kumjian, M. R., S. M. Ganson, and A. V. Ryzhkov, 2012: Freezing of raindrops in deep convective updrafts: A microphysical and polarimetric model. J. Atmos. Sci., 69, 3471–3490, https://doi.org/10.1175/JAS-D-12-067.1.
Kumjian, M. R., A. P. Khain, N. Benmoshe, E. Ilotoviz, A. V. Ryzhkov, and V. T. J. Phillips, 2014: The anatomy and physics of ZDR columns: Investigating a polarimetric radar signature with a spectral bin microphysical model. J. Appl. Meteor. Climatol., 53, 1820–1843, https://doi.org/10.1175/JAMC-D-13-0354.1.
Leon, D. C., and Coauthors, 2016: The Convective Precipitation Experiment (COPE): Investigating the origins of heavy precipitation in the southwestern United Kingdom. Bull. Amer. Meteor. Soc., 97, 1003–1020, https://doi.org/10.1175/BAMS-D-14-00157.1.
Lindley, T. T., and L. R. Lemon, 2007: Preliminary observations of weak three-body scatter spikes associated with low-end severe hail. Electron. J. Severe Storms Meteor., 2 (3), http://www.ejssm.org/ojs/index.php/ejssm/article/view/12.
Meischner, P. F., V. N. Bringi, D. Heimann, and H. Höller, 1991: A squall line in southern Germany: Kinematics and precipitation formation as deduced by advanced polarimetric and Doppler radar measurements. Mon. Wea. Rev., 119, 678–701, https://doi.org/10.1175/1520-0493(1991)119<0678:ASLISG>2.0.CO;2.
Neely, R. R., III, and Coauthors, 2018: The NCAS mobile dual-polarisation Doppler X-band weather radar (NXPol). Atmos. Meas. Tech., https://doi.org/10.5194/amt-2018-41, in press.
Petersen, W. A., and Coauthors, 1999: Mesoscale and radar observations of the Fort Collins flash flood of 28 July 1997. Bull. Amer. Meteor. Soc., 80, 191–216, https://doi.org/10.1175/1520-0477(1999)080<0191:MAROOT>2.0.CO;2.
Picca, J. C., and A. V. Ryzhkov, 2010: Polarimetric signatures of melting hail and S and C bands: Detection and short-term forecast. 26th Conf. on Interactive Information and Processing Systems (IIPS) for Meteorology, Oceanography, and Hydrology, Atlanta, GA, Amer. Meteor. Soc., 10B.4, https://ams.confex.com/ams/pdfpapers/161240.pdf.
Picca, J. C., M. R. Kumjian, and A. V. Ryzhkov, 2010: ZDR columns as a predictive tool for hail growth and storm evolution. 25th Conf. on Severe Local Storms, Denver, CO, Amer. Meteor. Soc., 11.3, https://ams.confex.com/ams/25SLS/techprogram/paper_175750.htm.
Ramachandran, R., A. Detwiler, J. Helsdon Jr., P. L. Smith, and V. N. Bringi, 1996: Precipitation development and electrification in Florida thunderstorm cells during Convection and Precipitation/Electrification Project. J. Geol. Res., 101, 1599–1619, https://doi.org/10.1029/95JD02931.
Rogers, R., 1967: Doppler radar investigation of Hawaiian rain. Tellus, 19, 432–454, https://doi.org/10.3402/tellusa.v19i3.9812.
Rowe, A. K., S. A. Rutledge, and T. J. Lang, 2011: Investigation of microphysical processes occurring in isolated convection during NAME. Mon. Wea. Rev., 139, 424–443, https://doi.org/10.1175/2010MWR3494.1.
Ryzhkov, A. V., and D. S. Zrnić, 2005: Radar polarimetry at S, C, and X bands: Comparative analysis and operational implications. Preprints, 32nd Conf. on Radar Meteorology, Albuquerque, NM, Amer. Meteor. Soc., 9R.3, http://ams.confex.com/ams/pdfpapers/95684.pdf.
Ryzhkov, A. V., M. Diederich, P. Zhang, and C. Simmer, 2014: Potential utilization of specific attenuation for rainfall estimation, mitigation of partial beam blockage, and radar networking. J. Atmos. Oceanic Technol., 31, 599–619, https://doi.org/10.1175/JTECH-D-13-00038.1.
Snyder, J. C., A. V. Ryzhkov, M. R. Kumjian, A. P. Khain, and J. Picca, 2015: A ZDR column detection algorithm to examine convective storm updrafts. Wea. Forecasting, 30, 1819–1844, https://doi.org/10.1175/WAF-D-15-0068.1.
Stith, J. L., J. E. Dye, A. R. Bansemer, A. J. Heymsfield, C. Grainger, W. Petersen, and R. Cifelli, 2002: Microphysical observations of tropical clouds. J. Appl. Meteor., 41, 97–117, https://doi.org/10.1175/1520-0450(2002)041<0097:MOOTC>2.0.CO;2.
Szumowski, M. J., R. M. Rauber, H. T. Ochs III, and L. J. Miller, 1997: The microphysical structure and evolution of Hawaiian rainband clouds. Part I: Radar observations of rainbands containing high reflectivity cores. J. Atmos. Sci., 54, 369–385, https://doi.org/10.1175/1520-0469(1997)054<0369:TMSAEO>2.0.CO;2.
Szumowski, M. J., R. M. Rauber, H. T. Ochs III, and K. V. Beard, 1998: The microphysical structure and evolution of Hawaiian rainband clouds. Part II: Aircraft measurements within rainbands containing high reflectivity cores. J. Atmos. Sci., 55, 208–226, https://doi.org/10.1175/1520-0469(1998)055<0208:TMSAEO>2.0.CO;2.
Taylor, J. W., and Coauthors, 2016: Observations of cloud microphysics and ice formation during COPE. Atmos. Chem. Phys., 16, 799–826, https://doi.org/10.5194/acp-16-799-2016.
Testud, J., E. Le Bouar, E. Obligis, and M. Ali-Mehenni, 2000: The rain profiling algorithm applied to polarimetric weather radar. J. Atmos. Oceanic Technol., 17, 332–356, https://doi.org/10.1175/1520-0426(2000)017<0332:TRPAAT>2.0.CO;2.
Trömel, S., M. Diederich, M. R. Kumjian, J. C. Picca, and C. Simmer, 2012: Columns of differential reflectivity: A precursor for storm evolution and convective rain. 2012 Fall Meeting, San Francisco, CA, Amer. Geophys. Union, Abstract A33G-0226.
Tuttle, J. D., V. N. Bringi, H. D. Orville, and F. J. Kopp, 1989: Multiparameter radar study of a microburst: Comparison with model results. J. Atmos. Sci., 46, 601–620, https://doi.org/10.1175/1520-0469(1989)046<0601:MRSOAM>2.0.CO;2.
Van Den Broeke, M., 2016: Polarimetric variability of classic supercell storms as a function of environment. J. Appl. Meteor. Climatol., 55, 1907–1925, https://doi.org/10.1175/JAMC-D-15-0346.1.
van Lier-Walqui, M., and Coauthors, 2016: On polarimetric radar signatures of deep convection for model evaluation: Columns of specific differential phase observed during MC3E. Mon. Wea. Rev., 144, 737–758, https://doi.org/10.1175/MWR-D-15-0100.1.
Wang, Y., P. Zhang, A. V. Ryzhkov, J. Zhang, and P. L. Chang, 2014: Utilization of specific attenuation for tropical rainfall estimation in complex terrain. J. Hydrometeor., 15, 2250–2266, https://doi.org/10.1175/JHM-D-14-0003.1.
Wilson, J. W., and D. Reum, 1988: The flare echo: Reflectivity and velocity signature. J. Atmos. Oceanic Technol., 5, 197–205, https://doi.org/10.1175/1520-0426(1988)005<0197:TFERAV>2.0.CO;2.
Yuter, S. E., and R. A. Houze Jr., 1995: Three-dimensional kinematic and microphysical evolution of Florida cumulonimbus. Part II: Frequency distributions of vertical velocity, reflectivity, and differential reflectivity. Mon. Wea. Rev., 123, 1941–1963, https://doi.org/10.1175/1520-0493(1995)123<1941:TDKAME>2.0.CO;2.
Zeng, Z., S. E. Yuter, R. A. Houze Jr., and D. E. Kingsmill, 2001: Microphysics of the rapid development of heavy convective precipitation. Mon. Wea. Rev., 129, 1882–1904, https://doi.org/10.1175/1520-0493(2001)129<1882:MOTRDO>2.0.CO;2.
Zhang, G., J. Sun, and E. A. Brandes, 2006: Improving parameterization of rain microphysics with disdrometer and radar observations. J. Atmos. Sci., 63, 1273–1290, https://doi.org/10.1175/JAS3680.1.
