Mesocyclone Evolution and Differences between Isolated and Embedded Supercells ()
1. Introduction
A supercell thunderstorm, which was originally defined by Browning [1,2], is a thunderstorm that is characterized by the presence of a mesocyclone. Supercell thunderstorms are perhaps the most violent of all thunderstorm types, and are capable of producing damaging winds, large hail, weak-to-violent tornadoes, and flash flooding with severe economic losses [3-5]. The definition of a mesocyclone was as given by Burgess [6] and included criteria on the shear value, the depth of the circulation, and its temporal continuity. A mesocyclone is a strong, rotating updraft found within a powerful thunderstorm. A typical mesocyclone is 2 to 6 miles across. A mesocyclone forms within a thunderstorm as a result of wind shear, or a change in wind speed and direction with altitude.
Since the advent of the WSR-88D Doppler radar, a suite of severe weather detection algorithms such as storm cell identification and tracking (SCIT) algorithm [7] and mesocyclone detection algorithm (MDA) [8,9], has had a positive impact on severe storm detecting and severe weather forecasting. A full description of the build 9 WSR-88D mesocyclone algorithm（B9MA） processing was described by Tipton et al. The B9MA searches for cyclonic azimuthal shear patterns (pattern vectors) in Doppler velocity data. Momentum and shear values are calculated for each pattern vector as follows:
, and
where Vin and Vout are maximum base velocity values toward and away from the radar, respectively. Momentum and shear values are compared to the following adaptable parameters: threshold high momentum (540 km2·h−1) and low shear (7.2 h−1), threshold high shear (14.4 h-1) and low momentum (180 km2·h−1). Pattern vectors that pass momentum and shear tests are expected to be characteristic of mesocyclones. Two-dimensional features are formed when a minimum number of pattern vectors (adaptable parameter threshold pattern vector (TPV)) are in close proximity to each other. Two-dimensional features are combined in the vertical to form three-dimensional features and mesocyclones are identified.
For this study, a set of 61 supercells from WSR-98D (the B9MA version of WSR-88D, manufactured by METSTAR, China) in southern part of North China Plain and the region of Jianghuai Plain, China, were chosen. Differences of severe weather, storm parameters, mesocyclone parameters were sought between different types of supercells and different regions.
2. Method
2.1. Radar Data Processing
To ensure radar product quality, WSR-98D data in Archive Level II format were post-processed by using Unit Control Position (UCP) and unified algorithm adaptable parameters. The components of meteorological algorithm in UCP most relevant to this study were WSR-98D SCIT and B9MA. All the default algorithm adaptable parameter values completely from WSR-88D system have not been adjusted in WSR-98D. The SCIT algorithm can automatically identify and track up to 100 storms each volume scan, and display some storm parameters such as the maximum reflectivity (DBZM), the maximum reflectivity height (HT), the cell-based vertically integrated liquid (C-VIL) and the top of storm cell (TOP) in storm structure product. The mesocyclone product fed by B9- MA algorithm can automatically identify storm-scale vortices and display mesocyclone parameters such as the mesocyclone base (M-BASE), the mesocyclone top (MTOP), the maximum shear of mesocyclones (M-SHEAR) and the maximum shear height (M-HT), and diameter of the core mesocyclone (DIAM).
2.2. Data Analysis Method
Storm parameters were analyzed and compared between different supercell types and different geographical area. All supercells in this study were divided into two categories (type I and type II) based upon the radar reflectivity and two regions (region I and region II) based upon different geographic regions. Type I (isolated supercells) was defined as being isolated from any quasi-linear region of ≥40 dBZ reflectivity. Type II (embedded supercells), was defined as a supercell storm embedded within a quasi-linear area of continuous reflectivity at the lowest volume scan ≥40 dBZ extending over a distance greater than 50 km [10]. The main research areas were plain areas included southern part of North China Plain (referred to as region I in this paper) and the region of Jianghuai Plain (referred to as region II in this paper). The WSR- 98D had lower height of radar antenna above sea level in plain areas, so can detect low-level vortices within the storm.
Mesocyclone parameters were analyzed and compared in three stages. The first stage was the first and second volume scan time accompanied by mesocyclone symbol (approximately 12 minutes).while the third stage was the last two volume scan time (approximately 12 minutes). The volume scan time between the first stage and the third stage was the second stage of mesocyclone (12 minutes at least).
The storm parameters were averaged from one volume scan associated with the first mesocyclone symbol (yellow solid circle）to the subsequent volume scan with the last mesocyclone symbol in one case, then averaged over all of the cases.
The mesocyclone parameters were averaged respectively in stage 1, stage 2 and stage 3 in one case, then averaged respectively in their three stages over all of the cases.
3. Samples and Examples
3.1. Supercell Samples
The set of 61 data cases, originated from 7 different radar sites (JINAN, SHIJIAZHUANG, TIANJIN, PUYANG, XUZHOU, HEFEI, NANJING) and associated with mesocyclones that lasted for at least 36 minutes (continuous six volume scans), were selected in this study from 2003 to 2012. The ranges of these storms were all 30 - 100 km from the radar sites, so the sample size and data quality were not much of a problem. 7 different radars mainly distributed in southern part of North China Plain and the region of Jianghuai Plain.
There were 32 isolated supercells and 29 embedded supercells among all the 61 supercells. 15/32 of isolated supercells and 13/29 of embedded supercells distributed in region I. 17/32 of isolated supercells and 16/29 of embedded supercells distributed in region II.
3.2. Examples
Below were two examples of parameter calculation and low-level reflectivity feature for isolated supercell and embedded supercell.
Figure 1 showed an isolated supercell in low-level reflectivity and the trend of mesocyclone. The supercell was discrete or isolated in radar reflectivity, associated with mesocyclone from 9:53 to 11:18 (continuous 15 volume scans). The storm parameters were averaged from 9:53 to 11:18. The averaged values of DBZM, C-VIL, HT and TOP for this supercell were 61.5 dBZ, 53.9 kg·m−2, 5.8 km and 9.9 km respectively. The first stage of mesocyclone was 9:53 and 9:59 volume scan time, while the third stage was 11:12 and 11:18 volume scan time, and the other volume scan time was the second stage. The averaged values of M-BASE and M-TOP in the three stages were 4.7, 3.2, 3.0 and 6.7, 7.2, 7.1 km respectively.
(a)
(b)
Figure 1. Low-level reflectivity (a) at 11:00 (GMT) and trend of mesocyclone (b) from 9:53 to 11:18 (GMT) on 26 August 2008, PUYANG radar.
Figure 2 showed an embedded supercell in low-level reflectivity and the trend of mesocyclone. The supercell was embedded in linear convective systems, associated with mesocyclone from 10:30 to 11:13 (continuous 8 volume scans). The storm parameters were averaged from 10:30 to 11:13. The averaged values of DBZM, C-VIL, HT and TOP for this supercell were 48.9 dBZ, 15 kg·m−2, 4.6 km and 10.0 km respectively. The first stage of mesocyclone was 10:30 and 10:36 volume scan time, while the third stage was 11:06 and 11:13 volume scan time, and the other volume scan time was the second stage. The averaged values of M-BASE and M-TOP in the three stages were 1.7, 0.9, 1.4, and 5.1, 6.0, 3.9 km respectively.
The averaged values of M-BASE and M-TOP in the three stages were 3.2, 2.1, 2.2 and 5.9, 6.6, 5.5 km respectively for two mesocyclones above.
4. Results
4.1. Severe Reports
There were 54 documented severe weather reports for the 61 supercells. The severe reports consisted of 7 tornado reports, 43 hail reports, 32 large hail reports, 41 damag-
(a)
(b)
Figure 2. Low-level reflectivity (a) at 11:06 (GMT) and trend of mesocyclone (b) from 10:30 to 11:13 (GMT) on 30 July 2003, JINAN radar.
ing wind gusts(≥17.2 m·s−1) reports and 13 heavy precipitation (≥50 mm·h−1 in this paper) reports. There were two embedded supercells in regions I, one isolated supercell and four embedded supercells in regions II with no severe weather reports.
Detailed severe weather reports for isolated and embedded supercells in different regions are shown in Tables 1 and 2. The probability of producing tornado for supercells in regions II was significantly higher than that in regions I. The probability of hail and large hail occurring for isolated supercells was significantly higher than that for embedded supercells. The probability of gust occurring for embedded supercells was obviously higher than that for isolated supercells. The probability of heavy rain occurring for isolated supercells was higher than that for embedded supercells.
4.2. Storm Parameters
Figure 3 showed the storm parameters of supercells and the differences of storm parameters for isolated supercells and embedded supercells in different regions.
The values of storm parameters were significantly higher for type I supercells compared to those for type II supercells. The average values of DBZM, C-VIL, HT, and TOP for type I supercells were 61.6 dBZ, 53 kg·m−2,
Table 1. Severe weather reports for isolated supercells (Type I) in different regions.
Table 2. Severe weather reports for embedded supercells (Type II) in different regions.
5.3 km, and 11.5 km respectively. The average values of DBZM, C-VIL, HT, and TOP for type II supercells were 58 dBZ, 39 kg·m−2, 3.9 km, and 10.1 km respectively.
The values of storm parameters were significantly higher for type I supercells compared to those for type II supercells in region I. The comparison of parameters between type I and type II supercells in region II was the same as that in region I.
The values of storm parameters were somewhat higher for type I supercells in region I compared to those in region II. The average values of storm parameters for type I supercells in region I were 62 dBZ, 54.7 kg·m−2, 5.7 km, and 11.8 km respectively. While those in region II were 61.2 dBZ, 51.2 kg·m−2, 4.9 km, and 11.2 km respectively. The values of storm parameters were somewhat higher for type II supercells in region I compared to those in region II.
4.3. Mesocyclone Parameters and Evolution
4.3.1. The Similarities and Differences between Type I Mesocyclone and Type II Mesocyclone
Figure 4 showed the mesocyclone parameters and the evolution trend during the three stages of mesocyclone for 32 isolated supercells and 29 embedded supercells. The mesocyclones of isolated supercells demonstrated the same evolution trend as the mesocyclones of embedded supercells. The mesocyclone formed in the middle and lower levels of an updraft. The M-BASE obviously