1. Introduction
Supercell thunderstorms produce a large proportion of high-end hail and tornado reports in the United States (Smith et al. 2012; Blair et al. 2017). They may be long-lived and responsible for powerful, long-track tornadoes that are often disproportionately responsible for tornado fatalities (Strader and Ashley 2014). One such long-lived supercell tracked from central Arkansas to northern Kentucky during 10–11 December 2021, producing two long-track tornadoes (one with a pathlength of 267 km) and 63 fatalities. The longest-track tornado from this storm had a pathlength comparable to the longest-identifiable continuous track segment produced by the 1925 Tri-State tornado (approximately 243 km; Johns et al. 2013).
Long-lived supercells persist for 4 h or more (Bunkers et al. 2006a) and may be long track depending on speed of motion. They are most common from 2300 to 0200 UTC and from April to July, and they are comparatively uncommon during the cool season (about 10% occur in September–February; 3% in December). Long-lived supercells are more likely than other supercells to be isolated and produce strong-violent tornadoes, and they less commonly experience storm mergers (Bunkers et al. 2006a). Among supercells that live for at least 5.9 h, many occur along a boundary (51%) or in the warm sector of a low pressure system (18%; Bunkers et al. 2006b). Long-lived supercell environments are characterized by large instability (Ziegler et al. 2010; Finch and Bikos 2010; Belo-Pereira et al. 2017) and strong surface–8-km bulk shear (Bunkers et al. 2006b), with very long-lived supercells (lifetime ≥ 7 h) characterized by the largest 0–8-km shear values (Davenport 2021). They often have relatively low lifted condensation levels (LCLs) and large 0–1-km storm-relative helicity (SRH), both attributes that have been associated with tornadic storms (Thompson et al. 2003). A rapid increase in SRH has been associated with rapid storm intensification (Scheffknecht et al. 2017), often concurrent with a decrease in CAPE (Davenport 2021).
Environments associated with long-track tornadoes are generally similar (Garner 2007). Long-track tornadoes [pathlength greater than 40.2 km (Garner 2007) or 48.3 km (Straka and Kanak 2022)] often appear to be surrounded by a broad circulation (Bluestein 2009). In a more recent study with a larger sample size, tornado pathlength was maximized in a corridor from Mississippi to northern Alabama most commonly from November to April (Garner et al. 2021). Their upper 25% of tornado pathlengths (pathlength ≥ 17.9 km) were characterized by fast-moving parent storms, large storm-top divergence, and moderate mixed-layer CAPE (MLCAPE 349–2885 J Kg−1; median 1196 J kg−1). Median MLCAPE for the top 50 pathlengths (pathlength ≥ 68 km and longevity ≥ 60 min) was 1236 J kg−1. Simulations indicate that their intensity and duration increase with environmental 0–3-km SRH (Naylor and Gilmore 2012). Long-track tornadoes often form near a boundary and travel into a more buoyant warm sector, without a boundary association during the majority of the tornado lifetime (Garner et al. 2021). Tornadoes with relatively steady intensity may exist near the inflection point between the forward (FFD) and rear-flank (RFD) downdrafts in a zone of optimal vorticity accumulation (Orf et al. 2017).
Since the upgrade of the Weather Surveillance Radar-1988 Doppler (WSR-88D) network in the United States from 2011 to 2013, additional radar variables are available that allow inferences about storm structure and microphysics (e.g., Kumjian 2013). Several of these variables will be discussed in this paper. Differential reflectivity ZDR is the difference in reflectivity between the return from a horizontally and vertically polarized wave, such that scatterers with their major axis oriented horizontally typically have positive ZDR and scatterers with major axis oriented vertically have negative ZDR. Thus, larger raindrops (which are progressively more oblate) typically have larger (increasingly positive) ZDR values. Given the sensitivity of ZDR to drop size, it is a valuable tool for studying size-sorting processes. In supercells, ZDR arcs (e.g., Kumjian and Ryzhkov 2008; Crowe et al. 2012) are areas of enhanced ZDR along the forward flank indicating size sorting of hydrometeors in the storm-relative inflow (Dawson et al. 2014). In addition to arcs, ZDR columns (e.g., Illingworth et al. 1987; Brandes et al. 1995; Kumjian et al. 2010) indicate regions of liquid drops being lofted or wet hail located above the ambient 0°C level within updrafts. Thus, they can serve as a proxy for updraft location, size, and intensity in convection. Tumbling objects such as falling hailstones tend to appear spherical when averaged through time and so typically have ZDR near 0 dB. Near-zero ZDR collocated with high radar reflectivity ZHH values often indicates regions of hail (e.g., Aydin et al. 1986; Straka et al. 2000; Kumjian and Ryzhkov 2008).
Copolar correlation coefficient ρhv is a measure of the variability of scatterers in a radar sample volume, including differences in particle phase, size, shape or aspect ratio, and orientation. Larger particle variability generally results in lower ρhv values (e.g., pure homogeneous raindrops will have a value near 1.0 whereas water-coated tumbling hail of various sizes may have a value of 0.75–0.95; Park et al. 2009). When nonmeteorological scatterers (e.g., birds, leaves, and chaff) are present the ρhv may be much lower, depending on the proportion and diversity of those scatterers. In supercells, this variable is used as a marker of the tornadic debris signature (TDS) when low ρhv values are collocated with rotation near the surface (e.g., Ryzhkov et al. 2005; Van Den Broeke and Jauernic 2014). Characteristics of the TDS may be related to characteristics of the parent tornado and its debris field (e.g., Bodine et al. 2013; Van Den Broeke 2015; Griffin et al. 2017; Umeyama et al. 2018) and/or underlying surface (e.g., Van Den Broeke 2015; Wienhoff et al. 2020).
Specific differential phase KDP is a local measure (range derivative) of the accumulated phase shift between a horizontally and vertically polarized wave. Values increase with drop size since larger drops have more oblate axis ratios, and scatterer number concentration. Thus, areas with high liquid water content and a large number of particles are characterized by high KDP values. This well-describes the region of many supercell storms that is under the FFD, known as the KDP foot (e.g., Romine et al. 2008) and indicates locally heavy precipitation. Many recent studies have focused on the angle and distance of separation between low-level ZDR arc and KDP foot centroids (e.g., Crowe et al. 2012; Loeffler and Kumjian 2018; Loeffler et al. 2020; Homeyer et al. 2020), which indicate magnitude of size sorting and potentially the degree of tornado threat (Loeffler et al. 2020).
In this paper we present the longest-duration high-quality polarimetric dataset for any supercell in the literature, from a 10–11 December 2021 storm that tracked from Arkansas to Kentucky. This dataset includes the full life cycles of two long-track tornadoes, one of which was among the longest-tracked in U.S. history, and it provides a rare opportunity to examine the structure of a relatively steady-state tornadic supercell. The radar signatures are integrated into a coherent model of the temporal evolution of inflow, updraft, hailfall, and tornadic debris within the 10–11 December 2021 supercell.
2. Data and methods
The 10–11 December supercell tracked through the domains of several WSR-88D instruments, including KLZK (Little Rock, Arkansas), KNQA (Memphis, Tennessee), KPAH (Paducah, Kentucky), KHPX (Fort Campbell, Kentucky), and KLVX (Louisville, Kentucky). For most of its track the storm was ≤ 100 km from the nearest radar (Fig. 1); therefore, the lowest radar tilt sampled the storm ≤ 1500 m above radar level (ARL). There were two brief exceptions: early in the storm’s life cycle between KLZK and KNQA in northeast Arkansas (2357–0042 UTC; maximum distance to a radar of 125 km) and toward the end of the first EF4 tornado along the Missouri–Tennessee border between KNQA and KPAH (0207–0230 UTC; maximum distance to a radar of 116 km).
Map of the 10–11 Dec 2021 supercell track (gray line) relative to WSR-88D sites. The supercell track was created by combining storm centroid tracks for this cell from all radars on this map. Tornado tracks from the NCEI Storm Events database (https://www.ncdc.noaa.gov/stormevents/) are also plotted, color-coded by EF-scale rating.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
The ZDR from the WSR-88D network has known calibration challenges (e.g., Richardson et al. 2017), so this variable was manually calibrated to ensure optimal values. Calibration followed Picca and Ryzhkov (2012) and assumed a true ZDR value of 0.15 dB in thunderstorm anvil ice crystals approximately 1.5 km above the ambient 0°C level [estimated from the spatiotemporally nearest uncontaminated Rapid Refresh (RAP) initialization sounding]. Two ZDR calibration factors separated by ≥1 h were averaged for each radar dataset to reduce bias in any one estimate (Table 1).
Manual ZDR calibration factors (dB) for each radar dataset. A positive value indicates that ZDR values are biased high.
The Supercell Polarimetric Observation Research Kit (SPORK; Wilson and Van Den Broeke 2022) was used to objectively identify and quantify supercell polarimetric features in this storm. SPORK is an algorithm that ingests radar data, detects and tracks storm objects, and for each storm object identifies and quantifies several polarimetric signatures common to supercells including ZDR arcs and columns, hail area, and separation angle and distance between the ZDR arc and KDP foot. The version of SPORK used here was modified slightly from that presented by Wilson and Van Den Broeke (2022)—rather than requiring a user-provided 0°C level and estimate of the supercell’s forward-flank edge orientation, it automatically estimates these fields from the most recent RAP analysis.
The ZDR column area requires knowledge of the environmental 0°C level to estimate. In the version of SPORK used here, a RAP-provided 0°C level was used at the location of the ZDR column for each sample volume. This yielded results closer to manual values than using a single 0°C level from a single RAP inflow sounding. The supercell forward-flank reflectivity gradient orientation is generally perpendicular to the deep-layer shear and was estimated using the RAP surface–400-hPa shear vector direction minus 87° (value determined for this dataset). This shear vector performed well among a sample of 8 related wind variables when plotted against forward-flank reflectivity gradient orientation for the 206-supercell dataset of Wilson and Van Den Broeke (2021) (Fig. 2; Pearson’s correlation = 0.80). Agreement is sufficient to indicate that forward-flank reflectivity gradient orientation can be estimated using this method, especially since forward-flank orientation serves as an estimate to help SPORK detect the ZDR arc and SPORK is not particularly sensitive to small changes in this estimate. Time series of ZDR column area and depth, ZDR arc area, hailfall area, and KDP–ZDR separation angle were produced by SPORK. For more details of SPORK, see Wilson and Van Den Broeke (2022).
Surface–400-hPa shear direction (°; x axis) vs forward flank reflectivity gradient orientation (°; y axis) for the 206-supercell sample of Wilson and Van Den Broeke (2021; blue dots; Pearson’s correlation = 0.80). Points were also added for the 11 Dec 2021 supercell at 0100 (red star), 0300 (purple triangle), and 0500 (yellow triangle) UTC, showing agreement with the method developed using prior storms.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
Manual time series of ZDR column area and depth, ZDR arc area, and hailfall area were also calculated following prior work (e.g., Van Den Broeke 2020) for comparison with SPORK output. The ZDR column depth was calculated as the top of the column minus the environmental 0°C level from a RAP sounding at the nearest hour, and column area was the manually enclosed area of ZDR ≥ 0.5 dB at 1 km above the same environmental 0°C level. The ZDR arc area was the manually enclosed area of ZDR ≥ 3.5 dB along the forward flank. Hail area was the manually enclosed collocated area of ZHH > 55 dBZ and −0.5 ≤ ZDR ≤ 1.0 dB. TDS area was calculated manually following Van Den Broeke and Jauernic (2014)—ZHH ≥ 20 dBZ was required to be collocated with rotation and ρhv ≤ 0.8 (≤0.9 if a local minimum). For more details and discussion of the comparison of manual and SPORK values, see section 4 (below).
Several metrics were manually calculated to quantify characteristics of the supercell’s primary low-level vortex—that with the largest maximum radial velocity Vr difference (maximum outbound minus maximum inbound velocity) at base scan. Maximum Vr difference (m s−1) was recorded. Then, Vr shear (s−1) was calculated by dividing Vr difference by the distance between the center points of the maximum inbound and outbound pixels [defined as Vr_Dist (m); similar to Stumpf et al. 1998]. The Vr shear accounts for temporal variations in rotational velocity and vortex size; thus, it is useful for identifying vortex intensification and weakening. These metrics were calculated regardless of base-scan altitude, which may have led to some biases since the metrics were moderately correlated to range, e.g., altitude of observation [Pearson’s correlation with range: Vr difference (−0.49), Vr shear (+0.45), Vr_Dist (−0.58)].
Environmental datasets were obtained from the Storm Prediction Center severe thunderstorm event archive (https://www.spc.noaa.gov/exper/archive/events/) including surface observations, archived mesoanalyses, upper-air charts, and archived radiosonde data. Soundings were only obtained from Little Rock since all others were greater than 200 km from the supercell. Additional environmental data were obtained from a RAP sounding collected in the supercell inflow at each hour from convection initiation (2000 UTC) to supercell dissipation (0700 UTC). If environmental data were needed at a particular time, the temporally nearest RAP sounding was used. For discussions of bias in RAP output see Coniglio (2012), Benjamin et al. (2016), and Coniglio and Jewell (2022). The RAP sounding location was chosen by following a line 40 km (Potvin et al. 2010) directly to the southeast of the tight reflectivity gradient on the supercell’s inflow side. No convective contamination was observed in the RAP soundings.
3. Environmental overview
Initiation of the 10–11 December supercell occurred just before 2000 UTC in southwestern Arkansas in the warm sector of a low pressure system over central Kansas (a common scenario described by Bunkers et al. 2006b). Fog was reported across Arkansas earlier in the day, with overcast skies afterward [similar to observations by Geerts et al. (2009)]. Forcing by synoptic-scale boundaries apparently played a minimal role in the initial stages of the convection of interest according to surface, satellite, and radar observations, which did not indicate the presence of any linear features, and warm sector convection remained relatively discrete. Low-level southerlies advected substantial moisture across Arkansas, with dewpoints around 20°C. This moisture advection helped erode an earlier capping inversion and surface-based CAPE reached approximately 2800 J kg−1 in the Little Rock 1900 UTC special sounding (Fig. 3).
Little Rock sounding from 1900 UTC 10 Dec 2021. Ambient temperature is the red solid line, and dewpoint is the green solid line. The path of a lifted surface parcel is indicated as the gold line. Wind is indicated to the right of the sounding, where barbs have conventional values, with the associated hodograph to the far right [height (km) is indicated as labeled dots].
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
Aloft, a 49 m s−1 southwesterly jet maximum at 500 hPa was located west of the initiation region associated with a deep, progressive upper-level trough. The 1900 UTC KLZK sounding (Fig. 3) indicated 40 m s−1 of 0–8-km bulk wind difference (BWD) and 25 m s−1 of 0–6-km BWD. RAP output from the initiation region suggested that by 2000 UTC 0–6-km BWD had increased to approximately 35 m s−1 (not shown), and effective BWD was greater than 35 m s−1. Bunkers et al. (2006b) suggest that long-lived supercells become more likely as 0–8-km BWD exceeds 25 m s−1. Therefore, the environment prior to storm initiation was conducive to long-lived supercells and potentially very long-lived supercells (Davenport 2021).
Through the time when a base-scan mesocyclone became evident around 0020 UTC, moisture remained ample and MLCAPE approached 2000 J kg−1 (Fig. 4) with LCL heights of <800 m. The 0000 UTC KLZK sounding indicated that 0–8-km BWD had increased to approximately 45 m s−1, and SRH values of 360 m2 s−2 (0–1 km) and 420 m2 s−2 (0–3 km) were observed. The right entrance region of a jet streak provided upper-level divergence over most of Arkansas (not shown). Substantial deep-layer wind shear, cyclonic curvature in the low-level hodograph, and sufficient instability for convective maintenance promoted rapid supercell intensification (Figs. 4 and 5).
MLCAPE (J kg−1; contoured) and 0–6-km BWD (kt) from the RAP at (a) 2200, (b) 0000, (c) 0200, (d) 0400, and (e) 0600 UTC 10–11 Dec 2021. The white star indicates the approximate location of the supercell centroid at plot time. BWD is only plotted where MLCAPE > 250 J Kg−1.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
As in Fig. 4, but for surface−1-km SRH (m2 s−2; shaded) and RAP hodographs. Purple hodographs lines represent the 0–1-km layer, red represents 1–2 km, orange represents 2–3 km, gold represents 3–6 km, green represents 6–9 km, and blue represents 9–12 km. The white star indicates approximate supercell centroid location at plot time.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
Genesis of the first long-lived EF4 tornado (hereinafter EF4A) was reported at 0107 UTC near the Missouri border. According to a RAP sounding, supercell inflow was characterized by 250 m2 s−2 of 0–1-km SRH and an LCL height of about 815 m, placing it in the upper 75% of the distribution 0–1-km SRH and lower 25% of the distribution of LCL height for significant tornadic environments (Thompson et al. 2003). Between 0100 and 0600 UTC 850-hPa winds increased from 50 to 70 kt (1 kt ≈ 0.51 m s−1) across the region, associated with a strengthening low-level jet (LLJ). This led to a large increase in SRH, particularly from 0100 to 0300 UTC when 0–3-km SRH increased from 300 m2 s−2 to greater than 500 m2 s−2 (Fig. 5) and 0–1-km SRH increased to 440 m2 s−2. During the same period, MLCAPE decreased from approximately 1900 to 1400 J kg−1. EF4A dissipated at 0236 UTC, and a second long-track EF4 tornado (hereinafter EF4B) began at 0249 UTC. LCL height remained less than 815 m from 0200 to 0500 UTC. This combined with increasing shear from the LLJ may have promoted weaker storm outflow, allowing for a long-lived, steady-state low-level mesocyclone and tornado (McCaul and Cohen 2002; Bunkers et al. 2006b). With strong upper-level divergence and storm motion of approximately 22 m s−1, this historic event is consistent with the climatology of maximized tornado pathlengths (Garner et al. 2021).
4. Radar observations
The online supplemental material includes a storm-centered animation of ZDR column depth and base-scan ZHH, storm-relative velocity, spectrum width, correlation coefficient ρhv, KDP, and ZDR from 2223 to 0748 UTC; we refer the reader to that animation for evidence of many of the features described in this section. Only nonsupplemental adaptive intravolume low level scans (SAILS; Chrisman 2014) are shown in the animation to remove jerkiness and limit the file to a reasonable size. Thus, times discussed in the text but not shown in the animation are times when a SAILS scan was completed.
a. Radar reflectivity (ZHH)
Storm initiation occurred southwest of KLZK, with the first 35+ dBZ echoes at 1954 UTC. Convection started as a loosely associated, multicellular cluster but became more linear as it crossed Little Rock around 2200 UTC. The line still had a strongly multicellular appearance by 2259 UTC, with multiple ZHH maxima aloft accompanied by distinct ZDR columns. By 2338 UTC, a cell had developed supercellular characteristics in both ZHH and velocity, including a sharp low-level forward-flank ZHH gradient, overhanging area of high ZHH aloft, and midlevel mesocyclone. New convection merged into the storm’s inflow region until 0028 UTC (not shown).
The storm continued to organize, evidenced by the appearance of a bounded weak echo region (BWER) by 0022 UTC (not shown). It was difficult to characterize low-level supercell structure during this time because of storm-radar distance, but when an echo appendage could be discerned, it had increasing cyclonic curvature through approximately 16 min after formation of EF4A (0123 UTC). The storm then maintained a strongly cyclonically curved appendage, but the curvature lessened starting about 7 min prior to dissipation of the aforementioned tornado (after 0229 UTC).
Intensity of the base-scan ZHH core was relatively steady until 0225 UTC (11 min prior to dissipation of EF4A), when there was an abrupt increase in storm-core area with ZHH > 60 dBZ associated with hail fallout (section 4f). This was followed by another apparent hail fallout near the storm’s rear flank from 0248 to 0305 UTC (the first ∼15 min of EF4B). Verifying this signature, a report of 4.4 cm (1.75-in.) hail was received at 0255 UTC. Immediately following the initial hailfall, the character of the supercell’s ZHH structure changed rapidly. The hook echo surged forward from 0248 UTC (1 min before genesis of EF4B) through 0259 UTC, becoming strongly cyclonically curved. An area of high ZHH with tornadic debris evident in ρhv appeared at the tip of the appendage at 0259 UTC. Several weak convective cells merged into the supercell’s forward flank between 0301 and 0322 UTC, but this did not substantially change low-level storm structure. Thereafter, the forward-flank ZHH gradient sharpened, and until 0425 UTC (not shown) the inflow was clear of extraneous convection.
Supercell reflectivity structure was still apparent when new convection formed to the south and east of the main storm at 0458 UTC (approximately 50 min prior to dissipation of EF4B), though a series of storm mergers began to obscure it thereafter. While the storm retained a BWER and deep mesocyclone, by dissipation of EF4B (approximately 0550 UTC) it had become embedded in a larger convective cluster. Despite being embedded, the storm retained its midlevel mesocyclone until finally becoming incorporated into the convective cluster and weakening after 0717 UTC (not shown).
b. Velocity and spectrum width
During early stages of development, the storm’s linear organization was reflected in the nearly linearly oriented convergence on the convection’s leading edge. By 2338 UTC, the strongest cell had developed supercellular characteristics, including a well-defined midlevel (2–4 km ARL) mesocyclone and broad low-level (<1 km ARL) cyclonic shear. Thereafter, a low-level mesocyclone rapidly strengthened.
Base-scan spectrum width generally remained less than 10 m s−1 within the developing storm until 2347 UTC (not shown), when it began to increase near the low-level mesocyclone. As the supercell continued to organize, base-scan spectrum width continued to increase until there was a broad area of spectrum width 10–15 m s−1 generally collocated with the mesocyclone by 2357 UTC.
The low-level mesocyclone quickly intensified through 0020 UTC, with convergence appearing by 0038 UTC. The most concentrated area of high spectrum width through this intensification was on the interface between inbound and outbound velocities, partially marking the leading edge of the RFD. By 14 min after formation of EF4A (0121 UTC), the region of high spectrum width had become closely associated with the mesocyclone. The strong mesocyclone and accompanying high spectrum width remained steady until approximately 20 min prior to dissipation of EF4A (0219 UTC), when the spectrum width field became less orderly and the values near the updraft decreased. Thereafter they occasionally increased but were not immediately associated with the mesocyclone as prior.
The storm began to show signs of cycling at 0229 UTC, when two areas of rotation became evident. The first, associated with EF4A, slowly lagged relative to a second, broad mesocyclone centered just ahead of the persistent hook echo along the forward flank. The original circulation, however, remained closely connected with the leading edge of the RFD. Eventually it grew closer to the storm’s inflow on the nose of an RFD pulse and merged into the main updraft region by 0302 UTC (13 min after genesis of EF4B; not shown), apparently becoming an important vorticity source during intensification of that tornado (similar to observations by Houser et al. 2015; French et al. 2015).
After genesis of EF4B, the storm entered another extended period with a strong, deep mesocyclone collocated with a region of high spectrum width. This state lasted nearly 2 h until the region of high spectrum width became lower-magnitude and spatially discontinuous by 0440 UTC. The mesocyclone retained intense midlevel azimuthal shear through 0501 UTC but then began to weaken. Two areas of rotation were evident by 0523 UTC (24 min prior to EF4B’s dissipation), with the older, tornadic mesocyclone to the southwest of a younger, broader mesocyclone. The oldest mesocyclone continued to recede south and west of the main storm track, becoming indistinct by several minutes after tornado dissipation (0551 UTC). This is consistent with the nonoccluding cyclic mesocyclogenesis model of Adlerman and Droegemeier (2005).
By 0550 UTC the storm had become part of a cluster of convection, and it no longer appeared as distinct in Vr as during the tornadic portion of its life cycle. The newest mesocyclone never became as strong or well-defined as earlier mesocyclones. Though the storm produced another tornado at 0606 UTC, low-level rotation continued to become less defined during a series of storm mergers from 0633 to 0706 UTC. Inflow was cut off by 0717 UTC and the original storm rapidly dissipated.
c. ZDR-inferred updraft characteristics
Prior agreement between SPORK-derived and manual ZDR column area has been high (e.g., Wilson and Van Den Broeke (2021) found Spearman’s correlation of 0.817 over 56 storms) but was poorer for the 10–11 December 2021 supercell (Spearman’s correlation = 0.538 over 66 sample volumes). While the two series generally changed in tandem (Fig. 6), SPORK undercalculated ZDR column area in association with first appearance of EF4A’s TDS and with a maximum in TDS size. Larger amounts of debris in the updraft column generally overlapped smaller SPORK-calculated ZDR column areas, likely because of the typically low ZDR values in tornadic debris. For all times with TDS area greater than 5 km2, Spearman’s correlation was −0.119 between the manual and SPORK time series. The presence of hail also strongly decreased agreement between ZDR column area estimation methods, with Spearman’s correlation −0.624 when times with hail area greater than 5 km2 were compared. Thus, given the presence of substantial debris and some hail in this storm, we opted to use manual time series for both ZDR column area and depth.
Manually calculated time series of ZDR column area (km2; blue) and depth (km; orange), and SPORK time series of ZDR column area (km2; purple dashed). Tornadoes are indicated by colored lines at the bottom of the plot, with EFU and EF0 tornadoes in light blue, EF1 tornadoes in green, and EF4 tornadoes in red.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
The ZDR column depth (orange, Fig. 6) increased to approximately 2 km as the storm organized. The column was consistently 1.5–2.5 km deep from 0000 to 0500 UTC. This contrasts with most supercells studied prior, which often had variable and cyclic updraft column behavior (Beck et al. 2006; Van Den Broeke 2020). Genesis of EF4B was associated with a decline in ZDR column depth to less than 1.5 km at 0259 UTC (and with a hailfall area maximum; section 4f), possibly indicating updraft weakening around tornadogenesis. If it occurred, this would agree with prior-observed declines in updraft vertical extent around tornadogenesis (Lemon and Doswell 1979; Satrio et al. 2021).
The ZDR column area (blue, Fig. 6) was smaller than 50 km2 when the storm was young and relatively disorganized, then grew to 80+ km2 starting at approximately 0020 UTC. Column area remained large from 0020 to 0202 UTC, except for 0116–0139 UTC, which corresponds to genesis of EF4A. This decline in column area may indicate updraft weakening coincident with low-level vortex intensification. After 0202 UTC, the ZDR column became smaller (typically 20–40 km2; Fig. 6), though its size was still consistent with significantly tornadic storms (French and Kingfield 2021). Occurrence of a smaller ZDR column corresponds to the end of EF4A, the time between long-track tornadoes, and much of the lifetime of EF4B. Shortly after genesis of EF4B (by 0259 UTC) the column again became relatively large (40–70 km2; Fig. 6), even though an abundance of tornadic debris within the updraft likely decreased detectable column size. We speculate that the storm’s inflow may still have been characterized by relatively large CAPE since an axis of high CAPE values was located just south of the storm (Fig. 4d). SRH values (Figs. 5c,d) and low-level shear were increasing, potentially leading to a larger shear-induced vertical motion component in the updraft. Combined, these environmental factors may be partially responsible for the large observed ZDR column during this time.
d. ZDR arc characteristics
High ZDR values in a supercell’s forward flank are attributed to hydrometeor size sorting from vertical wind shear in the storm’s environment (Ryzhkov et al. 2005; Kumjian and Ryzhkov 2008; Dawson et al. 2014). The ZDR arc region is dominated by large, oblate drops with high ZDR values (Seliga and Bringi 1976) as the storm-relative winds have less time to act on these drops (Kumjian and Ryzhkov 2008, 2009). The agreement between SPORK-derived and manual ZDR arc size was strong (Spearman’s correlation = 0.826 over 46 sample volumes; Fig. 7). Manual inspection showed that the two series generally changed in tandem, but SPORK values were too small throughout most of the lifetime of EF4B. The SPORK underestimate appears to be due to the large amount of debris present that reduced ZDR below the algorithm’s threshold to be considered part of the arc. Some of this same area would, however, be included in the manual arc area if it was clear that the same size sorting process was acting to locally increase ZDR. Thus, we opted to use the manually calculated ZDR arc size. Underestimation of true ZDR arc size may persist since the ZDR arc is a low-level feature (Kumjian and Ryzhkov 2008), and radar-storm distance sometimes precluded a good low-level sample.
SPORK (blue) and manual (green) time series of ZDR arc area for the 10–11 Dec supercell from 2300 UTC 10 Dec 2021 to 0700 UTC 11 Dec 2021. Tornadoes are indicated by colored lines at the top of the plot, with EFU and EF0 tornadoes in light blue, EF1 tornadoes in green, and EF4 tornadoes in red.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
Lack of an organized ZDR arc and storm-radar distance (greater than 100 km) before 0030 UTC prohibits an analysis in the supercell’s organizing phase. By 0042 UTC, a robust ZDR arc with a steady area of 80–100 km2 developed that remained throughout the lifetime of EF4A (Fig. 7). Arc area remained greater than 40–50 km2 from 0042 to 0202 UTC, except for 0115–0138 UTC when the ZDR arc area decreased during genesis of EF4A (and ZDR column area concurrently decreased). A decrease in ZDR arc area after 0202 UTC corresponded with disorganization of the supercell’s forward flank and demise of EF4A.
The ZDR arc signature reappeared approximately 30 min after genesis of EF4B (at 0323 UTC). Arc area remained consistently smaller during EF4B than during EF4A, with an area of 60 − 85 km2 from 0324 to 0506 UTC. Arc area decreased between 0440 and 0456 UTC, possibly due to a small hailfall event that lowered ZDR values. The ZDR arc weakened toward the end of EF4B and became intermittent after 0535 UTC (Fig. 7).
e. ZDR–KDP separation
Separation between regions of enhanced ZDR and KDP in a supercell indicates size sorting by the storm-relative wind. Characteristics of this signature can indicate whether a storm’s low-level wind profile is favorable for tornado production (Crowe et al. 2012; Homeyer et al. 2020; Loeffler et al. 2020; Wilson and Van Den Broeke 2022). Tornado production is favored when the angle between the storm motion vector and a vector drawn from the centroid of the region of enhanced KDP to the centroid of the ZDR arc (referred to here as the separation angle) is large. Loeffler et al. (2020) found that tornadic storms had separation angles more orthogonal than nontornadic storms, and Wilson and Van Den Broeke (2022) found that most tornadic [nontornadic] supercells had separation angles greater than 40° [less than 40°]. The 10–11 December supercell consistently displayed a large separation angle (Fig. 8), with values greater than 40° throughout the period just before and during EF4A and during much of EF4B.
As in Fig. 7, but for ZDR–KDP separation angle.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
As the supercell organized in central and northeastern Arkansas from 2300 to 0000 UTC, a robust analysis of separation angle characteristics was precluded due to storm-radar distance (greater than 100 km from just after 0000 UTC to just before 0045 UTC) and the lack of an organized ZDR arc before 0000 UTC. Once a robust ZDR arc developed, the separation angle consistently remained 40°–60° through the genesis and first part of EF4A’s lifetime (0042–0133 UTC). Separation angles gradually increased from around 50° to 80° between 0133 and 0144 UTC. This corresponds to when Vr difference and Vr shear became large in magnitude as the vortex decreased in size (Fig. 9). Between 0213 and 0230 UTC, the supercell’s forward flank was mostly greater than 100 km from a radar, so SPORK’s ZDR arc and separation angle estimates are of reduced quality.
Time series of maximum Vr difference divided by 10 to fit on same primary axis (label Vr_Difference; green; m s−1), distance between maximum inbound and outbound velocities (Vr_Dist; double blue; km), and Vr difference divided by Vr_Dist (label Vr_Shear; orange; s−1) for low-level mesocyclone of the 10–11 Dec 2021 supercell. Tornadoes are indicated by colored lines at the bottom of the plot, with EFU and EF0 tornadoes in light blue, EF1 tornadoes in green, and EF4 tornadoes in red. Top bars qualitatively indicate when the vortex was relatively intense or well structured: “compact vortex” = Vr_Dist consistently less than 2 km, “large Vr difference” = Vr difference consistently greater than 60 m s−1, and “large shear” = Vr shear consistently larger than 0.05 s−1.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
After 0230 UTC (6 min prior to demise of EF4A), the separation angle rapidly decreased from 83° to 51°, coincident with the ZDR arc moving downshear along the storm’s forward flank and disorganizing before dissipating at 0253 UTC. A robust ZDR arc signature did not reappear until 0324 UTC, 35 min into EF4B’s lifetime. Separation angle values during EF4B were more erratic than during EF4A but still mostly greater than 40° (Fig. 8). The more-erratic nature of the separation angle during this period was likely due to the consistently weaker and smaller ZDR arc. Within 15 min of EF4B’s demise, the ZDR arc weakened and a robust separation angle was not identified after 0535 UTC.
f. Hailfall
Hailfall extent at 1 km ARL was calculated manually and by SPORK throughout the supercell’s lifetime. Both methods showed similar trends: the two analyses had Spearman’s correlation of 0.879 over all 67 sample volumes (most of which had 0-km2 hail area) and Spearman’s correlation of 0.655 over the 12 sample volumes where at least one method returned a nonzero value. Differences in the SPORK/manual results could be due to several factors. Manual analysis was conducted at each SAILS time while SPORK calculated hailfall area at the beginning of each volume scan. Higher variability in the manual analysis is attributable to the use of more frequent radar scans. In addition, the presence of debris may have made the SPORK hailfall area less accurate, since debris may have been misclassified as hail given its relatively low ρhv, relatively high ZHH, and locally variable ZDR. Radar-storm distance may also be responsible for some artificially large hail areas due to downward interpolation in construction of 1-km constant altitude plan position indicator scans. Thus, manual hail area calculations were used.
Substantial hailfall was not produced during the supercell’s organizing phase. A three-body scatter spike was noted aloft from 0126 to 0138 UTC (not shown), though manual analysis only indicated 0.84–2.08 km2 of hailfall area during this time (Fig. 10). Both KPAH and KNQA indicated a rapid increase in hailfall area just after 0200 UTC that peaked between 68 km2 (KPAH) and 79 km2 (KNQA) at 0230 UTC (6 min prior to dissipation of EF4A). The hailfall area quickly decreased after the peak, reaching a minimum of 3.4 km2 (KPAH) at 0253 UTC (4 min after genesis of EF4B). Shortly after this minimum, a 4.4-cm (1.75-in.) hail report was received. After another brief increase to approximately 12 km2 around 0300 UTC, it decreased to 0 km2 by 0315 UTC (Fig. 10). A second report of 4.4-cm (1.75-in.) hail was received from the storm’s forward-flank ZHH maximum at 0324 UTC, when hail was not expected given the manual and SPORK analyses. At this time, however, ρhv and ZDR were locally lower in this area, supportive of hail (not shown). A minor hailfall (maximum area 23.51 km2) occurred between 0410 and 0420 UTC, but hail area was no longer detected by 0423 UTC and remained absent thereafter (Fig. 10).
As in Fig. 7, but for hailfall area.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
The initial sharp peak in hailfall area occurred during the last ∼30 min of EF4A as it crossed northwest Tennessee. This was concurrent with an increase in Vr_Dist (diameter of the circulation; Fig. 11), suggesting that hail fallout may have been related to updraft broadening. Additionally, hail embryos added to the updraft during cell mergers around this time may have contributed to more prolific hail formation (see section 5c). Hail area steadily decreased until reaching a minimum at the start of EF4B. Several periods of hailfall were indicated in the time of updraft disruption between the two significant tornadoes, when several weak tornadoes occurred. A secondary hailfall maximum occurred 10–15 min after genesis of EF4B and then rapidly decreased to 0 km2 (Fig. 11).
Time series of dual-pol and circulation metrics. Circle sizes for each metric are relative to the largest value of each metric. Separation angle was obtained from SPORK, and all other metrics are manually calculated (Vr_Dist = circulation diameter). Gray-shaded regions represent times when the supercell was poorly sampled at <1500 m ARL. Tornadoes are indicated by colored lines at the bottom of the plot, with EFU and EF0 tornadoes in light blue, EF1 tornadoes in green, and EF4 tornadoes in red.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
g. TDS characteristics
A well-defined TDS existed through much of the supercell’s life (blue, Fig. 12). Three TDS instances prior to 0111 UTC were small, relatively shallow, and had ρhv values of 0.80–0.90. These TDSs were loosely associated with several weak, short-lived tornadoes reported during this time. Appearance of a TDS prior to the storm’s first tornado report indicates but does not necessitate (Van Den Broeke 2015; Houser et al. 2016) that a tornado may have occurred prior to the report. Other inconsistencies between TDS and tornado report timing may be due to the time required for debris to be lofted to radar beam height and/or to slight temporal errors in tornado report(s).
TDS characteristics from 2341 UTC 10 to 0609 UTC 11 Dec 2021. The blue line indicates base-scan TDS area (km2; left axis), the orange dots indicate maximum TDS altitude (km; left axis), and the green line indicates mean ρhv associated with the TDS (right axis). Tornadoes are indicated by colored lines at the bottom of the plot, with EFU and EF0 tornadoes in light blue, EF1 tornadoes in green, and EF4 tornadoes in red.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
A TDS first appeared at 0123 UTC (not shown), 16 min after reported genesis of EF4A. Since the tornado was 76.1 km from the radar (beam altitude 1.07 km ARL), the temporal delay may be partially explained by the time needed to loft debris to radar beam height. Since the tornado was reported to produce EF-2 damage within 4 min of tornadogenesis, however, we cannot provide reasonable speculation on why the TDS did not appear sooner. Once it had appeared the TDS signature rapidly deepened and broadened, with debris visible to 6–9 km ARL and TDS area 10–20 km2 through the tornado’s life (Fig. 12). After a brief period without a TDS (0247–0257 UTC), the TDS associated with EF4B appeared. Though its maximum altitude was comparable to that of EF4A (5–10 km ARL; Fig. 12), area of the TDS remained quite small (5–10 km2) through the first 2 h. This size is comparable to prior results—for example, Van Den Broeke and Jauernic (2014) found an average size of an EF4-associated TDS of 6.83 km2. TDS vertical extent was larger than has been commonly found in prior work—Van Den Broeke and Jauernic (2014) found that average vertical debris extent with EF4 tornadoes is 4.32 km. A sharp TDS size increase occurred at 0458 UTC (Fig. 12), and the TDS generally remained larger than 10 km2 from then until disappearing at 0604 UTC (not shown), approximately 20 min after demise of EF4B.
In storms with tornadoes that loft large quantities of debris, fallout of debris may be observed that is well separated from the tornado (e.g., Van Den Broeke 2015). This debris fallout region may be large and occur in the storm core, as observed in association with a 2019 EF4 tornado near Kansas City, Missouri (Wang et al. 2020). The debris fallout region in the core of the 10–11 December 2021 supercell was more extensive than has been observed in prior storms. Fallout was most notable with EF4B (0249–0547 UTC); an associated TDS was observed from 0259 to 0603 UTC (Fig. 12). Debris began to be observed outside the strongest base-scan Vr vortex at 0307 UTC coincident with an outflow surge on the south side of the vortex (not shown), and by 0309 UTC an extensive base-scan debris field was present northward through the western half of the storm core (Figs. 13a,b). A large area of debris was present south of the vortex by 0313 UTC (not shown), which thereafter was caught up into the main tornado updraft and transported northeast. This sequence resembles the model of debris ejections developed by Kurdzo et al. (2015) from the 2013 Moore, Oklahoma, tornado and Houser et al. (2016) from the 2011 El Reno, Oklahoma tornado, though in this case the ejection-like phenomenon is present for 6–7 min as opposed to the 1–2 min found by Kurdzo et al. (2015) or a few minutes as found by Houser et al. (2016).
Examples of the debris distribution (white annotations) in base-scan (left) ZHH and (right) ρhv from KPAH at (a),(b) At 0309 UTC (beam altitude at panel center is 0.49 km); (c),(d) at 0324 UTC (altitude is 0.33 km); and (e),(f) at 0347 UTC (altitude is 0.70 km).
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
By 0324 UTC much of the storm core contained debris at low levels (Figs. 13c,d). A large amount of debris reached the north flank of the storm where it may have fallen out (e.g., see Figs. 13e,f at 0347 UTC), though large portions of the storm core contained substantial debris and associated lowering of ρhv through 0600 UTC (not shown). We are not aware of prior documentation of such a long-lasting and spatially extensive tornado debris field, though this may be because polarimetric documentation of long-track tornadoes is uncommon. We speculate that the storm-relative shear being oriented along a line from the tornado toward the storm core allowed debris transport into the storm core (0–6-km shear was oriented roughly parallel to the storm path; Fig. 4). More complete insight would need to be provided by numerical simulations.
Given the long period over which a TDS was apparent and the relatively high-quality radar data through this time, this case provides an opportunity to compare TDS characteristics with radar-derived measures of low-level rotation and vortex size (Fig. 9). TDS area and mean ρhv were not strongly related to these radar measures. One exception is delta Vr in comparison with mean ρhv in the TDS (r = −0.42), indicating that when velocity difference across the tornado is large, more nonmeteorological scatterers may be lofted, decreasing the ρhv value. Maximum debris column height was associated more strongly with the radar measures (Fig. 14). Strong velocity difference in the base-scan mesocyclone was associated with a taller debris column (r = 0.72; Fig. 14a), in agreement with Van Den Broeke and Jauernic (2014) and Emmerson et al. (2019). Vr shear was similarly associated with debris height (r = 0.56; Fig. 14b), and most scans without a TDS also had low Vr shear. The Vr shear was generally larger for EF4B owing to the smaller vortex size and consistently large Vr difference (Fig. 9). Deep debris columns were generally associated with more compact vortices (r = −0.50; Fig. 14c).
Maximum TDS height (y axes; km) vs (a) maximum delta Vr within the base-scan vortex (m s−1), (b) Vr shear (s−1), and (c) Vr_Dist (m). Pearson’s correlation using all points is indicated. When zeroes are excluded, correlation decreases to 0.61 [in (a)], 0.35 [in (b)], and −0.25 [in (c)].
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
TDS characteristics for the two EF4 tornadoes were compared with landcover, which was taken from the Esri 2020 land cover dataset (Karra et al. 2021), a global, satellite-derived dataset with 10-m resolution and nine landcover classifications. EF4A tracked across a highly agricultural area of the Mississippi River valley (Fig. 15a), then into a forested area before dissipating over western Tennessee. TDS characteristics varied through the tornado life cycle and were not clearly a function of landcover (Fig. 12). EF4B tracked across more variable landcover, with different areas dominated by forest, crops, and urban areas (e.g., Mayfield, Kentucky; Fig. 15b). The notable increase in TDS size at 0458 UTC (Fig. 12) was associated with a landcover transition from mostly forest to mostly cropland, but this pattern was not repeatable across similar landcover transitions.
Landcover along the two EF4 TDS tracks: (a) EF4A (0107–0236 UTC) and (b) EF4B (0249–0547 UTC). Stars indicate the location of the strongest debris-associated vortex in base-scan data every 5 base scans (approximately 7–12 min between stars).
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
Portions of the TDS were associated with a feature similar to the three-body scatter spike (TBSS) seen with hail events (e.g., Lemon 1998; Lindley and Lemon 2007; Blair et al. 2011). Banacos et al. (2012) documented an example associated with debris from a Massachusetts tornado, which they attributed to “Mie scattering off large and/or wet debris in a heavily forested area.” Their debris spike lasted for approximately 19 min and for part of that time was concurrent with a hail-generated TBSS (Banacos et al. 2012). In the present supercell, this signature was seen with EF4B from 0318 to 0335 UTC (surrounding production of EF3/4 damage near Mayfield), from 0345 to 0349 UTC (following production of EF3 damage near Benton, Kentucky), and around 0359 UTC (following production of EF4 damage in Cambridge Shores, Kentucky). Debris spike appearance at 0330 UTC is presented as an example (Fig. 16). The debris spike was most visible in ρhv (values consistently less than 0.7) and typically extended 10–25 km down radial from the TDS center (Fig. 16a). ZDR was noisy within the debris spike (Fig. 16b), possibly partially because of low collocated ρhv. The debris spike was typically visible to an altitude of 1–1.5 km and had maximum ZHH of 12–20 dBZ (Fig. 16c), though at 0332 UTC (several min after impacting Mayfield) the debris spike was visible to 3.44-km altitude (not shown). Vr generally indicates weak outbound values (Fig. 16d). In both this and the study of Banacos et al. (2012), a debris spike was produced during and just after production of EF3 + damage, and when the area down radial of the TDS was relatively free from scatterers, allowing good visibility.
Debris spike at 0330 UTC from KPAH: (a) ρhv, (b) ZDR, (c) ZHH, and (d) Vr. The white oval is the TDS, and yellow annotation indicates the location of the debris spike in ρhv. The TDS was 37.6 km from KPAH (altitude 0.43 km); the ρhv debris spike extends 25.5 km downradial from the TDS center.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
5. Integration of radar observations
a. Supercell organization and intensification
Evolution of the radar signatures from storm initiation through development of EF4A indicated development and strengthening of the mid and low-level mesocyclones and associated main updraft, and possibly the development of enhanced storm-relative inflow as evidenced by increasing ZDR arc area (Fig. 11). The ZDR column area markedly increased between 0000 and 0030 UTC according to manual and SPORK ZDR column area time series (Fig. 6), and the manual time series of rotational velocity shows an initial increase from below 40 to 60 m s−1 between 2345 and 0014 UTC followed by a continued increase to greater than 80 m s−1 by 0120 UTC (Fig. 9). A tighter circulation accompanied these changes. Meanwhile, a marked increase in ZDR arc area occurred between 0030 and 0100 UTC (Fig. 7). Proximity soundings from 0000 to 0100 UTC do not show a notable increase in low-level shear or SRH to explain this increase in forward-flank size sorting. Thus, this increase in arc area may indicate an increase in size sorting due to stronger storm-relative flow induced by the storm modifying its environment, from the presence of an unresolved mesoscale heterogeneity, and/or from a deepening of the sorting layer leading to enhanced particle size sorting.
By the time the supercell produced its first tornado (0015 UTC), it had a well-defined updraft and mesocyclone. Before formation of the storm’s first violent tornado (EF4A; 0107 UTC), continued expansion of the already-large updraft was evident through growth of the ZDR column area, which typically exceeded 70 km2 between 0020 and 0107 UTC (for context, the 75th percentile of pretornadic ZDR column area was 54 km2; Wilson and Van Den Broeke 2022). Updraft expansion may be indirectly inferred through the aforementioned enhancement of storm-relative inflow indicated by rapid strengthening of the ZDR arc from 0030 to 0100 UTC.
b. First steady-state period
Between 0100 and 0200 UTC, the supercell’s dual-pol signatures were distinguished by remarkable steadiness as EF4A tracked from northeast Arkansas to western Tennessee. ZDR column area remained large (consistent with French and Kingfield 2021) except for a decrease centered at 0122 UTC. The separation angle and ZDR arc area remained large throughout this period (Fig. 7), while hailfall area was negligible. The consistency of the ZDR column and lack of a hailfall signature indicate that the storm maintained a large, robust updraft with no notable cycles or weakening. The persistently large separation angle is consistent with favorable low-level shear profiles, while the steady and large ZDR arc indicates that the storm likely did not experience consequential inflow disruption. After a period of strengthening rotational velocity and tightening circulation (0107–0120 UTC) coincident with EF4A’s development and rapid intensification, circulation metrics were also remarkably consistent, with rotational velocity remaining near or above 80 m s−1 and circulation diameter remaining below 2 km through 0200 UTC (Fig. 9).
c. Cell interaction and hail event
Just after genesis of EF4A (0107 UTC), 5-km ARL ZHH and ZDR column depth at 0115 UTC show a large, isolated supercell with one dominant updraft (Fig. 17a). New updrafts developed along the supercell’s flanking line around 0200 UTC, producing small precipitation cores that impinged on the supercell from the west (Fig. 17b). By 0201 UTC, these new updrafts moved slightly south relative to the main updraft and intensified, with associated precipitation apparently falling into the supercell updraft (Fig. 17c). Some of these hydrometeors from the new updrafts may have served as hail embryos, possibly increasing supercell hail production and leading to an updraft disruption. Indeed, by 0223 UTC the supercell’s ZDR column was reduced to a fraction of its former area (Fig. 6), and its precipitation core aloft was larger with higher reflectivity values (Fig. 17d). The area of polarimetrically inferred hailfall at 1 km ARL rapidly increased from 0213 to 0230 UTC leading up to demise of EF4A. The ZDR arc area also quickly decreased from 0200 to 0230 UTC, possibly due to disruption by hail and to reduced size sorting as inflow into the weakening updraft decreased. Meanwhile, rotational velocity briefly exceeded 90 m s−1 at 0220 UTC before decreasing coincident with dissipation of EF4A (0236 UTC). The low-level circulation diameter broadened from 3 km (0200 UTC) to greater than 5 km (0236 UTC) and reached 8 km by 0243 UTC (Fig. 9). The association between tornado dissipation, decreasing ZDR column area, and decreasing strength of the ZDR arc is consistent with Segall et al. (2021). This case is also consistent with the finding that intense low-level mesocyclones are more likely to be weakened during cell mergers (Flournoy et al. 2022).
Storm-centered plots of ZHH at 5 km ARL and ZDR column depth from SPORK at (a) 0116, (b) 0138, (c) 0202, (d) 0224, (e) 0241, and (f) 0324 UTC.
Citation: Monthly Weather Review 151, 9; 10.1175/MWR-D-22-0330.1
Examining time series of these metrics together (Fig. 11), decreases in ZDR column and ZDR arc areas occur first around 0200 UTC (just after precipitation from the flanking-line cells began to affect the main updraft), followed by a hail area increase at 1 km ARL, followed by weakening and broadening of the low-level circulation and dissipation of EF4A. We hypothesize the following chain of events, acknowledging that several elements are quite speculative: first, precipitation from flanking line cells introduced hail embryos to the main supercell’s updraft, increasing hail production. This additional hail aloft weakened the ZDR column signature through precipitation loading and masking of the high-ZDR area. The weakening updraft may have facilitated ZDR arc weakening through reduced inflow, and ZDR arc weakening may also correspond to the arrival of hail at the low levels. The combination of a weaker midlevel updraft (and possibly weaker storm-relative inflow and colder outflow due to hail reaching low levels) disrupted the low-level mesocyclone and led to its weakening, and led to demise of EF4A. During its disrupted state (0236–0249 UTC) the supercell produced three short-lived, weak tornadoes (Fig. 11).
d. Second steady-state period and subsequent weakening
A first radar-based indication of strengthening near genesis of EF4B occurred between 0235 and 0246 UTC, when SPORK-detected ZDR column area rapidly increased (Fig. 6). The ZDR column area increase does not appear in the manual time series. This may be because the manual time series used a single observed 0°C level estimate across the radar domain, whereas SPORK used a spatiotemporally changing field of 0°C level from RAP. These values could be several degrees Celsius different, possibly allowing SPORK to identify ZDR column trends sooner. Genesis of EF4B immediately followed this spike in SPORK-calculated ZDR column area, apparently just before remnant vorticity associated with EF4A merged with the low-level mesocyclone. This vorticity interaction appears similar to behavior noted in an Oklahoma supercell that produced violent tornadoes, during which two cyclonic tornadoes appeared to merge and result in a stronger tornado (French et al. 2015; Houser et al. 2015). SPORK column area remained large until EF4B’s demise. Manually calculated ZDR column area increased steadily throughout this tornado’s life cycle until reaching values close to those calculated by SPORK around 0400 UTC (Fig. 6). This increase in ZDR column area occurred concurrently with increasing separation between the supercell and its flanking line convection (Figs. 17e,f). This may indicate that less precipitation from flanking line storms was entering the supercell’s updraft, reducing hail production and allowing updraft recovery. Indeed, hail area decreased from a peak near the end of EF4A to 0 km2 as EF4B was strengthening (Fig. 10). This reduced hail production may have indicated, along with increased ZDR column area, the updraft’s recovery from earlier disruption. Less hail falling into the outflow may have also resulted in warmer outflow, possibly allowing the storm to support another long-track tornado.
The ZDR arc area was the third radar metric to recover to values observed during EF4A, with a spike in arc area between 0315 and 0400 UTC (Fig. 7). The implied size sorting increase may indicate the recovery of strong low-level inflow as the strengthening mesocyclone modified its environment. Once the ZDR arc recovered, the supercell maintained a strong ZDR column, defined ZDR arc, and limited or nonexistent hailfall through EF4B’s life cycle until the ZDR arc began to weaken after 0500 UTC. These polarimetric signature trends during EF4B’s life are paralleled by low-level rotation metrics, as the low-level circulation steadily strengthened from 0245 UTC to extreme values around 0500 UTC, and circulation diameter decreased from 0245 to 0315 UTC with a tight couplet through the end of EF4B (Figs. 9 and 11).
The ZDR arc area markedly declined after 0500 UTC (Fig. 7), implying weakening low-level storm-relative inflow and/or updraft. ZDR column area decreased and hailfall area remained negligible through demise of EF4B. Rotational velocity weakened as the circulation broadened prior to and during this tornado’s dissipation (Fig. 9), and shortly after the storm became indistinguishable.
6. Conclusions
We present polarimetric radar observations from a long-lived supercell with two associated long-track EF4 tornadoes. Relatively high-quality polarimetric data over the storm’s life allowed an unprecedented examination of interrelationships between radar signatures related to the genesis, maintenance, and demise of these two long-track tornadoes. The storm did not occur at a climatologically favorable time for long-lived supercells (Bunkers et al. 2006a), although it was consistent with the long-path tornado climatology (Garner et al. 2021). The storm’s deep-layer shear environment was typical of very long-lived supercells (Davenport 2021), though the storm did not occur near a boundary as with many very long-lived supercells (Bunkers et al. 2006b). A large SRH increase and concurrent CAPE decrease through the storm’s life is consistent with prior findings (Craven and Brooks 2004; Davenport 2021; Garner et al. 2021).
After the storm displayed supercell characteristics, it had some features that may be useful for future operational analysis of supercell evolution. Its ZDR column and arc areas grew rapidly by genesis of EF4A to values well exceeding the 75th percentile of typical pretornadic storm values (Wilson and Van Den Broeke 2022). The period containing EF4A was characterized by a deep ZDR column, large ZDR arc, large KDP–ZDR separation angle, and absence of hailfall. This relatively steady state was interrupted by precipitation from upstream cells falling into the supercell updraft, soon followed by hailfall, reduction in the updraft and storm-relative inflow strength, and interruption of the low-level mesocyclone. Such a disruption pattern may be beneficial for operational meteorologists to anticipate during long-track tornadoes. It is also important to recognize that reorganization and subsequent tornado intensification can occur quickly, as in this storm, when a remnant circulation associated with EF4A merged into the updraft and apparently became a vorticity source for EF4B. It is unknown how often this sort of “vorticity recycling” may occur in supercell tornado sequences, but we speculate it is one mode contributing to supercell non-occluding cyclic mesocyclogenesis. In this storm, we speculate this was possible because of the storm-relative environmental shear direction and a hailfall-associated RFD surge that allowed the prior vortex to collocate with the low-level mesocyclone. Future numerical work examining the plausibility of such a scenario may be beneficial.
After a second steady period characterized by a large ZDR column and arc during which a long-lived EF4 tornado occurred, the inflow and updraft weakened and the tornado dissipated. The often large ZDR column, especially when a tornado was ongoing, reinforces the supposition of French and Kingfield (2021) that this metric may be operationally beneficial to determine potential tornado intensity or likelihood. After tornado dissipation, the supercell became markedly less organized and discrete (Fig. 1). Several new updrafts merged with the supercell, and its remnants dissipated before reaching southern Ohio.
Acknowledgments.
University of Nebraska–Lincoln is acknowledged for graduate student financial support and publication fee support. The authors thank Jim Kurdzo and two anonymous peer reviewers for their insightful comments that improved the paper.
Data availability statement.
The data for this paper are available online (radar data—https://s3.amazonaws.com/noaa-nexrad-level2/index.html; RAP output—https://www.nco.ncep.noaa.gov/pmb/products/rap/; surface, upper-air, and sounding observations—https://www.spc.noaa.gov/exper/archive/events/ and https://www.ncei.noaa.gov/products/land-based-station/automated-surface-weather-observing-systems; landcover data—https://livingatlas.arcgis.com/landcover/).
REFERENCES
Adlerman, E. J., and K. K. Droegemeier, 2005: The dependence of numerically simulated cyclic mesocyclogenesis upon environmental vertical wind shear. Mon. Wea. Rev., 133, 3595–3623, https://doi.org/10.1175/MWR3039.1.
Aydin, K., T. A. Seliga, and V. Balaji, 1986: Remote sensing of hail with a dual linear polarization radar. J. Climate Appl. Meteor., 25, 1475–1484, https://doi.org/10.1175/1520-0450(1986)025<1475:RSOHWA>2.0.CO;2.
Banacos, P. C., M. L. Ekster, J. W. Dellicarpini, and E. J. Lyons, 2012: A multiscale analysis of the 1 June 2011 northeast U.S. severe weather outbreak and associated Springfield, Massachusetts tornado. Electron. J. Severe Storms Meteor., 7 (7), https://ejssm.com/ojs/index.php/site/article/view/43.
Beck, J. R., J. L. Schroeder, and J. M. Wurman, 2006: High-resolution dual-Doppler analyses of the 29 May 2001 Kress, Texas, cyclic supercell. Mon. Wea. Rev., 134, 3125–3148, https://doi.org/10.1175/MWR3246.1.
Belo-Pereira, M., C. Andrade, and P. Pinto, 2017: A long-lived tornado on 7 December 2010 in mainland Portugal. Atmos. Res., 185, 202–215, https://doi.org/10.1016/j.atmosres.2016.11.002.
Benjamin, S. G., and Coauthors, 2016: A North American hourly assimilation and model forecast cycle: The Rapid Refresh. Mon. Wea. Rev., 144, 1669–1694, https://doi.org/10.1175/MWR-D-15-0242.1.
Blair, S. F., D. R. Deroche, J. M. Boustead, J. W. Leighton, B. L. Barjenbruch, and W. P. Gargan, 2011: A radar-based assessment of the detectability of giant hail. Electron. J. Severe Storms Meteor., 6 (7), https://ejssm.com/ojs/index.php/site/article/view/34.
Blair, S. F., and Coauthors, 2017: High-resolution hail observations: Implications for NWS warning operations. Wea. Forecasting, 32, 1101–1119, https://doi.org/10.1175/WAF-D-16-0203.1.
Bluestein, H. B., 2009: The formation and early evolution of the Greensburg, Kansas, tornadic supercell on 4 May 2007. Wea. Forecasting, 24, 899–920, https://doi.org/10.1175/2009WAF2222206.1.
Bodine, D. J., M. R. Kumjian, R. D. Palmer, P. L. Heinselman, and A. V. Ryzhkov, 2013: Tornado damage estimation using polarimetric radar. Wea. Forecasting, 28, 139–158, https://doi.org/10.1175/WAF-D-11-00158.1.
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.
Bunkers, M. J., M. R. Hjelmfelt, and P. L. Smith, 2006a: An observational examination of long-lived supercells. Part I: Characteristics, evolution, and demise. Wea. Forecasting, 21, 673–688, https://doi.org/10.1175/WAF949.1.
Bunkers, M. J., J. S. Johnson, L. J. Czepyha, J. M. Grzywacz, B. A. Klimowski, and M. R. Hjelmfelt, 2006b: An observational examination of long-lived supercells. Part II: Environmental conditions and forecasting. Wea. Forecasting, 21, 689–714, https://doi.org/10.1175/WAF952.1.
Chrisman, J. N., 2014: The continuing evolution of dynamic scanning. NEXRAD Now, No. 23, NOAA/NWS/Radar Operations Center, 8–13, https://www.roc.noaa.gov/wsr88d/PublicDocs/NNOW/NNow23a.pdf.
Coniglio, M. C., 2012: Verification of RUC 0–1-h forecasts and SPC mesoscale analyses using VORTEX2 soundings. Wea. Forecasting, 27, 667–683, https://doi.org/10.1175/WAF-D-11-00096.1.
Coniglio, M. C., and R. E. Jewell, 2022: SPC mesoscale analysis compared to field-project soundings: Implications for supercell environment studies. Mon. Wea. Rev., 150, 567–588, https://doi.org/10.1175/MWR-D-21-0222.1.
Craven, J. P., and H. E. Brooks, 2004: Baseline climatology of sounding derived parameters associated with deep moist convection. Natl. Wea. Dig., 28, 13–24.
Crowe, C. C., C. J. Schultz, M. Kumjian, L. D. Carey, and W. A. Petersen, 2012: Use of dual-polarization signatures in diagnosing tornadic potential. Electron. J. Oper. Meteor., 13, 57–78.
Davenport, C. E., 2021: Environmental evolution of long-lived supercell thunderstorms in the Great Plains. Wea. Forecasting, 36, 2187–2209, https://doi.org/10.1175/WAF-D-21-0042.1.
Dawson, D. T., E. R. Mansell, Y. Jung, L. J. Wicker, M. R. Kumjian, and M. Xue, 2014: Low-level ZDR signatures in supercell forward flanks: The role of size sorting and melting of hail. J. Atmos. Sci., 71, 276–299, https://doi.org/10.1175/JAS-D-13-0118.1.
Emmerson, S. W., S. E. Nelson, and A. K. Baker, 2019: A comprehensive analysis of tornado debris signatures associated with significant tornadoes from 2010–2017. 18th Annual Student Conf., Phoenix, AZ, Amer. Meteor. Soc., S210, https://ams.confex.com/ams/2019Annual/webprogram/Paper356019.html.
Finch, J. D., and D. Bikos, 2010: A long-lived tornadic supercell over Colorado and Wyoming, 22 May 2008. Electron. J. Severe Storms Meteor., 5 (5), https://ejssm.com/ojs/index.php/site/article/view/25.
Flournoy, M. D., A. W. Lyza, M. A. Satrio, M. R. Diedrichsen, M. C. Coniglio, and S. Waugh, 2022: A climatology of cell mergers with supercells and their association with mesocyclone evolution. Mon. Wea. Rev., 150, 451–461, https://doi.org/10.1175/MWR-D-21-0204.1.
French, M. M., and D. M. Kingfield, 2021: Tornado formation and intensity prediction using polarimetric radar estimates of updraft area. Wea. Forecasting, 36, 2211–2231, https://doi.org/10.1175/WAF-D-21-0087.1.
French, M. M., P. S. Skinner, L. J. Wicker, and H. B. Bluestein, 2015: Documenting a rare tornado merger observed in the 24 May 2011 El Reno-Piedmont, Oklahoma, supercell. Mon. Wea. Rev., 143, 3025–3043, https://doi.org/10.1175/MWR-D-14-00349.1.
Garner, J. M., 2007: A preliminary study on environmental parameters related to tornado path length. NWA Rep. 2007-EJ5, 29 pp., https://nwafiles.nwas.org/ej/pdf/2007-EJ5.pdf.
Garner, J. M., W. C. Iwasko, T. D. Jewel, R. L. Thompson, and B. T. Smith, 2021: An environmental study on tornado pathlength, longevity, and width. Wea. Forecasting, 36, 1471–1490, https://doi.org/10.1175/WAF-D-20-0230.1.
Geerts, B., T. Andretta, S. J. Luberda, J. Vogt, Y. Wang, L. D. Oolman, D. Bikos, and J. Finch, 2009: A case study of a long-lived tornadic mesocyclone in a low-CAPE complex-terrain environment. Electron. J. Severe Storms Meteor., 4 (3), https://ejssm.com/ojs/index.php/site/article/view/20.
Griffin, C. B., D. J. Bodine, and R. D. Palmer, 2017: Kinematic and polarimetric radar observations of the 10 May 2010, Moore–Choctaw, Oklahoma, tornadic debris signature. Mon. Wea. Rev., 145, 2723–2741, https://doi.org/10.1175/MWR-D-16-0344.1.
Homeyer, C. R., T. N. Sandmæl, C. K. Potvin, and A. M. Murphy, 2020: Distinguishing characteristics of tornadic and nontornadic supercell storms from composite mean analyses of radar observations. Mon. Wea. Rev., 148, 5015–5040, https://doi.org/10.1175/MWR-D-20-0136.1.
Houser, J. L., H. B. Bluestein, and J. C. Snyder, 2015: Rapid-scan, polarimetric, Doppler radar observations of tornadogenesis and tornado dissipation in a tornadic supercell: The “El Reno, Oklahoma” storm of 24 May 2011. Mon. Wea. Rev., 143, 2685–2710, https://doi.org/10.1175/MWR-D-14-00253.1.
Houser, J. L., H. B. Bluestein, and J. C. Snyder, 2016: A finescale radar examination of the tornadic debris signature and weak-echo reflectivity band associated with a large, violent tornado. Mon. Wea. Rev., 144, 4101–4130, https://doi.org/10.1175/MWR-D-15-0408.1.
Illingworth, A. J., 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/10.1002/qj.49711347604.
Johns, R. H., D. W. Burgess, C. A. Doswell III, M. S. Gilmore, J. A. Hart, and S. F. Piltz, 2013: The 1925 Tri-State tornado damage path and associated storm system. Electron. J. Severe Storms Meteor., 8 (2), https://ejssm.com/ojs/index.php/site/article/view/47.
Karra, K., C. Kontgis, Z. Statman-Weil, J. C. Mazzariello, M. Mathis, and S. P. Brumby, 2021: Global land use/land cover with Sentinel 2 and deep learning. 2021 IEEE Int. Geoscience and Remote Sensing Symp. IGARSS, Brussels, Belgium, Institute of Electrical and Electronics Engineers, 4704–4707, https://doi.org/10.1109/IGARSS47720.2021.9553499.
Kumjian, M. R., 2013: Principles and applications of dual-polarization weather radar. Part I: Description of the polarimetric radar variables. J. Oper. Meteor., 1, 226–242, https://doi.org/10.15191/nwajom.2013.0119.
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., and A. V. Ryzhkov, 2009: Storm-relative helicity revealed from polarimetric radar measurements. J. Atmos. Sci., 66, 667–685, https://doi.org/10.1175/2008JAS2815.1.
Kumjian, M. R., A. V. Ryzhkov, V. M. Melnikov, and T. J. Schuur, 2010: Rapid-scan super-resolution observations of a cyclic supercell with a dual-polarization WSR-88D. Mon. Wea. Rev., 138, 3762–3786, https://doi.org/10.1175/2010MWR3322.1.
Kurdzo, J. M., D. J. Bodine, B. L. Cheong, and R. D. Palmer, 2015: High-temporal resolution polarimetric X-band Doppler radar observations of the 20 May 2013 Moore, Oklahoma, tornado. Mon. Wea. Rev., 143, 2711–2735, https://doi.org/10.1175/MWR-D-14-00357.1.
Lemon, L. R., 1998: The radar “three-body scatter spike”: An operational large-hail signature. Wea. Forecasting, 13, 327–340, https://doi.org/10.1175/1520-0434(1998)013<0327:TRTBSS>2.0.CO;2.
Lemon, L. R., and C. A. Doswell III, 1979: Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Mon. Wea. Rev., 107, 1184–1197, https://doi.org/10.1175/1520-0493(1979)107<1184:STEAMS>2.0.CO;2.
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), https://ejssm.org/archives/wp-content/uploads/2021/09/vol2-3.pdf.
Loeffler, S. D., and M. R. Kumjian, 2018: Quantifying the separation of enhanced ZDR and KDP regions in nonsupercell tornadic storms. Wea. Forecasting, 33, 1143–1157, https://doi.org/10.1175/WAF-D-18-0011.1.
Loeffler, S. D., M. R. Kumjian, M. Jurewicz, and M. M. French, 2020: Differentiating between tornadic and nontornadic supercells using polarimetric radar signatures of hydrometeor size sorting. Geophys. Res. Lett., 47, e2020GL088242, https://doi.org/10.1029/2020GL088242.
McCaul, E. W., Jr., and C. Cohen, 2002: The impact on simulated storm structure and intensity of variations in the mixed layer and moist layer depths. Mon. Wea. Rev., 130, 1722–1748, https://doi.org/10.1175/1520-0493(2002)130<1722:TIOSSS>2.0.CO;2.
Naylor, J., and M. S. Gilmore, 2012: Environmental factors influential to the duration and intensity of tornadoes in simulated supercells. Geophys. Res. Lett., 39, L17802, https://doi.org/10.1029/2012GL053041.
Orf, L., R. Wilhelmson, B. Lee, C. Finley, and A. Houston, 2017: Evolution of a long-track violent tornado within a simulated supercell. Bull. Amer. Meteor. Soc., 98, 45–68, https://doi.org/10.1175/BAMS-D-15-00073.1.
Park, H. S., A. V. Ryzhkov, D. S. Zrnić, and K.-E. Kim, 2009: The hydrometeor classification algorithm for the polarimetric WSR-88D: Description and application to an MCS. Wea. Forecasting, 24, 730–748, https://doi.org/10.1175/2008WAF2222205.1.
Picca, J., and A. Ryzhkov, 2012: A dual-wavelength polarimetric analysis of the 16 May 2010 Oklahoma City extreme hailstorm. Mon. Wea. Rev., 140, 1385–1403, https://doi.org/10.1175/MWR-D-11-00112.1.
Potvin, C. K., K. L. Elmore, and S. J. Weiss, 2010: Assessing the impacts of proximity sounding criteria on the climatology of significant tornado environments. Wea. Forecasting, 25, 921–930, https://doi.org/10.1175/2010WAF2222368.1.
Richardson, L. M., W. D. Zittel, R. R. Lee, V. M. Melnikov, R. L. Ice, and J. G. Cunningham, 2017: Bragg scatter detection by the WSR-88D. Part II: Assessment of ZDR bias estimation. J. Atmos. Oceanic Technol., 34, 479–493, https://doi.org/10.1175/JTECH-D-16-0031.1.
Romine, G. S., D. W. Burgess, and R. B. Wilhelmson, 2008: A dual-polarization-radar-based assessment of the 8 May 2003 Oklahoma City area tornadic supercell. Mon. Wea. Rev., 136, 2849–2870, https://doi.org/10.1175/2008MWR2330.1.
Ryzhkov, A. V., T. J. Schuur, D. W. Burgess, and D. S. Zrnić, 2005: Polarimetric tornado detection. J. Appl. Meteor., 44, 557–570, https://doi.org/10.1175/JAM2235.1.
Satrio, C. N., D. J. Bodine, R. D. Palmer, and C. M. Kuster, 2021: Multi-radar analysis of the 20 May 2013 Moore, Oklahoma supercell through tornadogenesis and intensification. Atmosphere, 12, 313, https://doi.org/10.3390/atmos12030313.
Scheffknecht, P., S. Serafin, and V. Grubišić, 2017: A long-lived supercell over mountainous terrain. Quart. J. Roy. Meteor. Soc., 143, 2973–2986, https://doi.org/10.1002/qj.3127.
Segall, J. H., M. M. French, D. M. Kingfield, S. D. Loeffler, and M. R. Kumjian, 2021: Storm-scale polarimetric radar signatures associated with tornado dissipation in supercells. Wea. Forecasting, 37, 3–21, https://doi.org/10.1175/WAF-D-21-0067.1.
Seliga, T. A., and V. N. Bringi, 1976: Potential use of radar differential reflectivity measurements at orthogonal polarizations for measuring precipitation. J. Appl. Meteor., 15, 69–76, https://doi.org/10.1175/1520-0450(1976)015<0069:PUORDR>2.0.CO;2.
Smith, B. T., R. L. Thompson, J. S. Grams, C. Broyles, and H. E. Brooks, 2012: Convective modes for significant severe thunderstorms in the contiguous United States. Part I: Storm classification and climatology. Wea. Forecasting, 27, 1114–1135, https://doi.org/10.1175/WAF-D-11-00115.1.
Strader, S. M., and W. S. Ashley, 2014: Cloud-to-ground lightning signatures of long-lived tornadic supercells on 27–28 April 2011. Phys. Geogr., 35, 273–296, https://doi.org/10.1080/02723646.2014.918527.
Straka, J., and K. Kanak, 2022: A climatology of long-track tornadoes. Electron. J. Severe Storms Meteor., 17 (1), https://ejssm.com/ojs/index.php/site/article/view/82.
Straka, J., D. S. Zrnić, and A. V. Ryzhkov, 2000: Bulk hydrometeor classification and quantification using polarimetric radar data: Synthesis of relations. J. Appl. Meteor., 39, 1341–1372, https://doi.org/10.1175/1520-0450(2000)039<1341:BHCAQU>2.0.CO;2.
Stumpf, G. J., A. Witt, E. D. Mitchell, P. L. Spencer, J. T. Johnson, M. D. Eilts, K. W. Thomas, and D. W. Burgess, 1998: The National Severe Storms Laboratory mesocyclone detection algorithm for the WSR-88D. Wea. Forecasting, 13, 304–326, https://doi.org/10.1175/1520-0434(1998)013<0304:TNSSLM>2.0.CO;2.
Thompson, R. L., R. Edwards, J. A. Hart, K. L. Elmore, and P. Markowski, 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting, 18, 1243–1261, https://doi.org/10.1175/1520-0434(2003)018<1243:CPSWSE>2.0.CO;2.
Umeyama, A., B. L. Cheong, S. Torres, and D. Bodine, 2018: Orientation analysis of simulated tornadic debris. J. Atmos. Oceanic Technol., 35, 993–1010, https://doi.org/10.1175/JTECH-D-17-0140.1.
Van Den Broeke, M. S., 2015: Polarimetric tornadic debris signature variability and debris fallout signatures. J. Appl. Meteor. Climatol., 54, 2389–2405, https://doi.org/10.1175/JAMC-D-15-0077.1.
Van Den Broeke, M. S., 2020: A preliminary polarimetric radar comparison of pretornadic and nontornadic supercell storms. Mon. Wea. Rev., 148, 1567–1584, https://doi.org/10.1175/MWR-D-19-0296.1.
Van Den Broeke, M. S., and S. T. Jauernic, 2014: Spatial and temporal characteristics of polarimetric tornadic debris signatures. J. Appl. Meteor. Climatol., 53, 2217–2231, https://doi.org/10.1175/JAMC-D-14-0094.1.
Wang, E. Y., D. J. Bodine, J. M. Kurdzo, J. Barham, C. Bowman, and P. Pietrycha, 2020: Polarimetric characteristics of tornado debris fallout during the May 28 2019 Lawrence/Kansas City, KS tornado. Severe Local Storms Symp., Boston, MA, Amer. Meteor. Soc., 925, https://ams.confex.com/ams/2020Annual/mediafile/Manuscript/Paper363530/SLS_SYMP_2020.pdf.
Wienhoff, Z. B., H. B. Bluestein, D. W. Reif, R. M. Wakimoto, L. J. Wicker, and J. M. Kurdzo, 2020: Analysis of debris signature characteristics and evolution in the 24 May 2016 Dodge City, Kansas, tornadoes. Mon. Wea. Rev., 148, 5063–5086, https://doi.org/10.1175/MWR-D-20-0162.1.
Wilson, M. B., and M. S. Van Den Broeke, 2021: An automated Python algorithm to quantify ZDR arc and KDP–ZDR separation signatures in supercells. J. Atmos. Oceanic Technol., 38, 371–386, https://doi.org/10.1175/JTECH-D-20-0056.1.
Wilson, M. B., and M. S. Van Den Broeke, 2022: Using the Supercell Polarimetric Observation Research Kit (SPORK) to examine a large sample of pretornadic and nontornadic supercells. Electron. J. Severe Storms Meteor., 17 (2), https://ejssm.com/ojs/index.php/site/article/view/85.
Ziegler, C. L., E. R. Mansell, J. M. Straka, D. R. MacGorman, and D. W. Burgess, 2010: The impact of spatial variation of low-level stability on the life cycle of a simulated supercell storm. Mon. Wea. Rev., 138, 1738–1766, https://doi.org/10.1175/2009MWR3010.1.