1. Introduction
The purpose of hail suppression is to eliminate hailstones or to make hailstorms hail in advance to protect crops or property in the protected area from hail damage. Since the launch of hail suppression, to explore more effective artificial hail-suppression methods, a large number of artificial hail-suppression experiments have been conducted to establish conceptual models of hail formation and hail suppression for various types of hailstorms (Browning 1977; Foote and Knight 1979; Pflaum 1980; Heymsfield et al. 1980; Ziegler et al. 1983; Miller and Fankhauser 1983; Xu and Duan 2001; Makitov et al. 2017). According to the theory of hail formation and the basic structure and potential dynamic characteristics of hailstorms, the hail-forming process can be influenced by introducing hygroscopic (salt aerosol) or ice-forming (silver iodide or dry ice) artificial nuclei into the cloud (Dennis 1977; Wieringa and Holleman 2006). Many hail-suppression hypotheses, such as “cloud water glaciation,” “beneficial competition,” “trajectory lowering,” “early rainout,” “promotion of coalescence,” “explosion effect,” and “dynamics effect,” have been proposed (WMO 1996; Fukuta et al. 2000). However, the formation process of hail is complex and easily affected by the environment. The effect of seeding is different not only in various storms but also in the same type of storms due to different microphysical processes and dynamic conditions in the cloud (Browning and Atlas 1977). Since modern detection methods are constantly developing, the hail-suppression hypothesis still has to be verified and refined by conducting scientific and objective artificial hail-suppression effect tests (Schickedanz and Changnon 1970).
Observation and verification of the seeding effect have always been an important and difficult part of hail-suppression projects. Early major hail-suppression projects revealed that seeding could increase hail days and 20%–40% of rainfall (Switzerland) and could reduce hail energy by 70% (the United States and Canada) or 50% (France) (Dessens et al. 2016). Currently, with increasingly detailed and comprehensive observation methods, radar parameters closely related to the physical characteristics of hailstorms are used to analyze the seeding effect. In Alberta’s airborne cloud seeding program, the radar observation results show that the strong echo area (radar reflectivity area > 60 dBZ) and maximum vertically integrated liquid water content (VIL) of seeded hailstorms are significantly smaller than that of unseeded hailstorms (Gilbert et al. 2016). In China, some results show that the echo top is significantly lower, as well as the echo intensity, and the value of VIL after seeding (Wang and Yu 1995; Liu et al. 2020). The decreasing rate of the maximum echo height in seeded hailstorms increased, and the structure of the hail cloud significantly changed (Li 1996). There are also studies that demonstrated the inhibitory effect of seeding on hailstorms (Goyer 1975; Sioutas 2017). However, some studies have not found any effect of seeding (Gavrilov et al. 2010; Rivera et al. 2020). Therefore, the seeding effect needs to be further examined through more work.
Since artificial seeding has an impact on storm dynamics and microphysical processes (Browning and Atlas 1977), Spain and Argentina examined the microphysical response of hailstorms after seeding via cloud physics instruments carried by aircraft. The results indicated that the supercooled water decreased and the ice concentration increased in the clouds after seeding (Krauss et al. 2000; Dessens et al. 2016). The hail-suppression project in North Dakota added an inert gaseous tracer and radar chaff to the catalyst and observed them with gas analyzers carried by aircraft and radar (Detwiler 2002).
Recently, dual-polarization radar has been available (Snyder et al. 2013). Because polarization parameters of dual-polarization radar are sensitive to changes in particle morphology, they can play an important role in revealing the changes in microphysics and thermodynamics in clouds after seeding (Heinselman and Ryzhkov 2006; Ventura et al. 2013; Homeyer and Kumjian 2015). According to the size and phase state of the hail and isotropy caused by its falling and rolling, the hail-forming region generally has typical polarization characteristics of high horizontal reflectivity ZH, low differential reflectivity ZDR, and a low correlation coefficient ρHV (Snyder et al. 2013; Kennedy et al. 2014). Since hail and large raindrops and large amounts of ice crystals (whose ZDR value and ρHV value are generally higher than graupel and hail) are simultaneously generated in the seeded area, the microphysical response of seeding indicates a decrease in the ZH value and an increase in the ZDR and ρHV values in the seeded area. The process of hail embryo growth into hail can be effectively inhibited by seeding (Chen et al. 2016). Like most studies, Chen et al. (2016) investigated the seeding effect of hailstorms based on the comparative analysis of seeded and unseeded hailstorms (Ćurić 1990; Gilbert et al. 2016; Sioutas 2017). However, there still exist differences in microphysical processes in different hailstorms.
In addition, the hail-suppression hypotheses behind the changes in physical parameters were verified while examining the seeding effect. Krauss (1999) revealed that the evolutionary characteristics of radar parameters of seeded hailstorms are consistent with early rainout and/or trajectory lowering. More studies support the view that among numerous hail-suppression hypotheses, the theories of beneficial competition and early rainout are the most promising basis for hail-suppression operations (WMO 1996; Wieringa and Holleman 2006).
The method of introducing competitive embryos into clouds, which is examined by Browning and Atlas (1977), may be suitable for hail formed by supercooled raindrops or graupel in multicell hailstorms. The effect of artificial seeding is affected by many factors, such as hailstorm type, hail formation mechanism, and environmental conditions, and one hail-suppression hypothesis is not universally applicable (Browning and Atlas 1977; Dessens et al. 2016). Among them, it is more difficult to find an effective hail-suppression method for multicell hailstorms because of the interaction between the cells, the more complex hail formation mechanism, and the hail embryo trajectory, when compared with ordinary unicellular storms.
Therefore, this study aims to explore the response characteristics of multicell hailstorms before and after seeding using dual-polarization parameters of the X-band dual-polarization radar. To further study the cloud microphysical processes, we adopted an improved X-pol hydrometeor identification (HID) method (Dolan and Rutledge 2009; Zhao et al. 2020) to identify hydrometeors in hailstorms. We divided the seeded hailstorm into three regions based on the seeding agent diffusion. Additionally, the evolution of the seeded hailstorm was divided into four periods (see section 3a for details). The response of the hailstorm after seeding was investigated by comparing changes in the radar parameters and hydrometeors in different regions and periods. Based on the variation characteristics of hydrometeors in the seeded area before and after seeding, we intended to explore how the artificial ice nuclei can influence the microphysical processes in the key hail-forming region.
The remaining parts of the paper are as follows. Section 2 first introduces the data, the data processing method, the construction of the HID method, and the theoretical basis of hail suppression. Section 3 describes the characteristics of the reflectivity, polarization parameters, and hydrometeors in the hail cloud after seeding. The discussion on problems of hail-suppression tests and future research direction is in section 4, followed by the principal findings of our study that are summarized in section 5.
2. Data and method
a. Study area and case
Weining Yi, Hui, and Miao Autonomous County (103°36′–104°45′E, 26°30′–27°25′N) is located in the Yunnan–Guizhou Plateau and slope transition region in northwestern Guizhou Province (Fig. 1) in China. Weining covers an area of 6296.3 km2 and governs 35 villages and towns, with an average altitude of 2200 m. The distance from the ground to the 0°C layer is low in Weining because of its high elevation, which is very conducive to hail formation in spring and summer. Also, it is a part of the “hail zone” in China and a frequent location of the Yunnan–Guizhou quasi-stationary front (D. Zhao et al. 2019) and southwest vortex (Wang and Tan 2014). The hail periods last long (from March to October) in Weining because of this unique complex terrain and weather system. Because cultivated and cash crops in Weining are mainly potatoes, buckwheat, and tobacco, hail disasters may induce severe economic losses. Since 2012, 37 antiaircraft gun and rocket operation stations have been set up in Weining to reduce hail (Fig. 1), with an average annual shell consumption of more than 18 000, protecting a crop area of 200 000 hectares. The Weining hailstorm is characterized by rapid generation and development, high hail frequency, and multiple types of hailstorms. The hail-suppression operations for different types of hailstorms mostly rely on the commander’s hail-suppression experience and field judgment. Because of the lack of quantitative and objective effect analyses, the effectiveness of hail suppression is controversial.
Location and elevation (m) of Weining in Guizhou Province. The color-coded symbols represent facilities.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Therefore, this paper explored the evidence of seeding effects by analyzing the evolution of the structure and microphysical characteristics of a multicell hailstorm after seeding in Weining. In Weining, from 2010 to 2017, hail mainly occurred from April to August, accounting for 98.36% of the total hail days (Zeng et al. 2018). Most hail disasters (65%) in Weining from 1997 to 2014 have been caused by local thunderstorms (Ke et al. 2016). This type of hailstorm is mainly unicellular and weak multicellular. Considering the quality of radar data and relatively small radar attenuation, we chose a multicell hail case in Weining on 27 April 2019, for analysis.
b. AWMO
With the extensive development of weather modification research and cloud physics experiments since the 1950s, China has gradually established a comprehensive system for artificial weather modification operations (AWMO) (She et al. 2019). The AWMO system mainly consists of weather radar, operational tools (aircraft, rockets, and antiaircraft guns), commanders, and operators. The weather radar is responsible for monitoring the echo development, and the commander instructs the operator to complete the operation through the Artificial Operation Directing System (AODS). In Weining, the AODS is based on X-band dual-polarization weather radar (local radar). In the season of frequent hail disasters, the C-band weather radar in Zhaotong (surveillance radar) automatically searches for the echo with hail according to the early recognition algorithm for hailstorms. The PPI scanning of the C-band radar has nine elevation angles with a period of 5 min. The X-band radar will enter the echo monitoring mode when the searching radar detects a risky echo (ZH ≥ 30 dBZ) and then automatically track the evolution of the target echo within a period of less than 1 min. When the echo develops to meet the operational conditions (the echo center of ZH ≥ 30 dBZ exceeds the −20°C layer), the commander accurately locates the operation target area according to the radar parameters, estimates the operation parameters, and instructs the operator to adjust the operating parameters of the antiaircraft gun through the communication system to complete the hail-suppression operation accurately and quickly. The reference for shells quantity is shown in Table 1.
No. of shells used for hail-suppression operations with antiaircraft guns and rockets (obtained from Weining Meteorological Bureau). The first row of the table head gives cell type, and the second row gives the volume (km3).
c. Radar and data
In this paper, the response of the multicell hailstorm after artificial hail suppression on 27 April 2019 over Weining was analyzed based on X-band dual-polarization radar (YLD1-D) data from Xueshan station (104.08°E, 27.05°N), as well as observation data, including weather station sounding data, weather modification operation records, and surface hail records. Fifth-generation hourly pressure-level reanalysis data obtained from the European Centre for Medium-Range Weather Forecasts (ECMWF) with a resolution of 0.25° × 0.25° (ERA5; Hersbach et al. 2020) were used to examine the atmospheric environment. The radar time was the start time of the volume scan, representing the data with a time range of approximately 6 min. Note that all times in the study are in coordinated universal time (UTC). The radar location is about 2200 m above sea level. The height shown in figures is the height above the ground (elevation above sea level minus 2200 m).
1) Attenuation correction
The radial data of the YLD1-D radar include reflectivity ZH, Doppler velocity V, spectrum width W, hydrometeor, differential reflectivity ZDR, differential phase ΦDP, specific differential phase KDP, and cross-correlation coefficient ρHV. For more information about the YLD1-D radar performance parameters, refer to Table 2. To ensure data quality, all data before analysis were subjected to quality control as follows. We first used the phase unfolding algorithm and the integrated filter method (C. H. Zhao et al. 2019) to handle the differential phase data and then adopted the least squares method to calculate the specific differential propagation phase. The self-consistent method with constraints (Park et al. 2005; Bi et al. 2012) was employed for the attenuation correction of horizontal reflectivity and differential reflectivity.
Characteristics of the YLD1-D radar. (The “range bin” represents the counting unit of radar resolution.)
2) Construction of the HID method
Based on the dual-polarization parameters (ZH, ZDR, KDP, and ρHV) and environmental temperature parameter T, an improved HID method was employed in our study, which includes seven types of hydrometeors: drizzle rain (DZ), rain (RN), aggregates (AG), ice crystals (CR), high-density graupel (HDG, growing in areas with high supercooled water content), low-density graupel (LDG, expected in relatively cold regions of storms, i.e., from −10° to −20°C), and rain–hail mixture (RH).
According to Park et al. (2009) and Huang et al. (2021), we adopted the trapezoidal membership function of HID based on fuzzy logic and set the weights of T, ZH, ZDR, KDP, and ρHV to 0.3, 0.3, 0.2, 0.1, and 0.1, respectively. For convective storms that can produce hail, liquid drops still exist above the 0°C layers because freezing is not instantaneous (Smith et al. 1999). To avoid misclassifications of the particle phase caused by the environmental temperature parameter, the identification of RN in the local mixed-phase region above the melting layer mainly depended on ρHV; therefore, the weights of ZH, ZDR, KDP, and ρHV were set to 0.4, 0.2, 0.3, and 0.1, respectively. This allows liquid water above the melting layer to be identified correctly, which is conducive to examining whether the supercooled water accumulation zone exists in the storm during its development and the variation in supercooled water (SWA) content after the ice nucleus was artificially induced. In the HID method, RN above the 0°C layer is considered SWA. The membership function threshold for the seven types of hydrometeors was set with reference to Huang et al. (2021). Based on the temperature profile, the temperature ranges for the five hydrometeors (DZ, HDG, LDG, AG and CR) are listed in Table 3. In this study, HID results show the distribution characteristics of hydrometeors in the cloud before and after seeding.
HID method temperature range.
In addition, to further investigate the change of hydrometeors, we quantified the hydrometeor. The YLD1-D radar has a range bin resolution of 1° in azimuth and 75 m in range. In the HID method, each range bin is assigned a hydrometeor category according to the characteristics of polarization parameters. The HID results represent the dominant hydrometeors in the range bin. The number of range bins is a proxy for the volume of the storm occupied by each hydrometeor. Considering that there were approximately 6 min for a radar volume scan, we chose the radar volume scan data closest to the operation time for analysis.
3) Verification of attenuation correction and HID results
Since the echo attenuation of C-band radar was less than that of X-band radar when passing through heavy rain areas or detecting long-distance precipitation, the maximum reflectivity in a column of the YLD1-D radar after attenuation correction was compared with that of the C-band radar (CINRAD/CC) to examine the attenuation correction effect. The CINRAD/CC radar is located in Zhaotong (103.72°E, 27.35°N) northwest of the YLD1-D radar, and its scanning range coincides with that of the YLD1-D radar in midwestern Weining. Because the echo still has attenuation at C band, we chose the cell at the edge of the strong echo zone as the verification cell (and there was no echo in front of the verification cell along the radial direction). In this paper, the volume scanning time of the CINRAD/CC radar is about 5 min, while that of the YLD1-D radar is about 6 min. The data at the time 1154 UTC were chosen to verify the attenuation correction because this time is closest to the hail time (1147 UTC) and both radars have data at this time. Figures 2 and 3 depict the different times of the same hail cloud (one radar volume scanning interval).
The composite reflectivity of (a) the CINRAD/CC radar and the YLD1-D radar (b) before and (c) after attenuation correction at 1154 UTC 27 Apr 2019. The dash-outlined box is the verification cell.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
The (a) ZH and (b) HID results at 0.5° elevation, and (c) vertical section of the HID results at 1147 UTC 27 Apr 2019. The red lines indicate the azimuth of the cross section, and black-outlined boxes represent the location of the hailstorm.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
As shown in Fig. 2, after the correction, strong echo areas (ZH ≥ 45 dBZ) became larger, and the strongest echo center was more obvious. As shown in Fig. 3a, the strongest echo center (ZH ≥ 55 dBZ) of the hailstorm appeared on the western boundary of Yangjie town, where the HID identified RH (Fig. 3b). The cross section of the HID results along the strongest echo center suggested that RH appeared near the surface approximately 18 km away from the radar (Fig. 3c), which was consistent with the time and place of hailfall recorded by the ground observation station. The above results show that HID has credibility for the identifications of RH.
Vertical distribution characteristics of hydrometeors in Fig. 3c show that RH, LDG, and HDG are distributed in the convective core of the hailstorm. The location of LDG is higher than that of HDG. AG is located at the edge of the hailstorm, and CR is mainly distributed in the top layer (above −10°C layer) in the hailstorm. DZ are distributed below the 0°C layer. These findings are consistent with the theoretical vertical distribution characteristics of hydrometeors in hailstorms and previous identification results (Dolan and Rutledge 2009; Zhao et al. 2020), indicating that HID also has credibility for the identifications of other hydrometeors.
d. Methods
1) Hail-suppression basis
The formation of hailstones goes through the embryo stage and hail stage at different locations in the storm (Young 1977). Because of the short updraft duration of a single cell in multicell storms, embryos cannot grow into hail by the mode of “microphysical recycling growth” as experienced by hailstones between the embryo curtain and the main updraft in supercell storms. Most hailstone embryos originate from feeder cells in front of the main storm and then enter the main updraft through turbulence and grow into hail (Heymsfield et al. 1980; Foote and Wade 1982; Ziegler et al. 1983; Cheng and Rogers 1988). Regarding the formation of the initial embryos, Dye et al. (1974) revealed that ice crystal and graupel mechanisms are the dominant mechanisms of hail formation in high-plain storms. Huang et al. (2021) also confirmed this point in their study on hailstone embryo characteristics in Weining, Guizhou Province. They found that low-density graupel contributed the most to hail formation in multicell hailstorms.
However, at the present time, embryos cannot be eliminated, and growth of graupel and hailstone is mainly in the liquid water accumulation zone within clouds in the last few minutes of its growth stage (Sulakvelidze et al. 1967; Zhao et al. 2020; Huang et al. 2021). Most hail-suppression operations are conducted on the basis of depleting supercooled water in storms (Dennis 1977). In addition, it will take longer for particles seeded into the feeder cell to be involved in precipitation progress. This seeding method is more suitable for multicell storms with a longer cycle and a larger scale. The weak multicell hailstorm in Weining develops rapidly and has a short cycle and a small scale. Therefore, the hail-suppression concept in Weining is to reduce the available supercooled water in the later stage of hail growth. The target area for seeding mainly ranges in height from −5° to −20°C of the main updraft.
The seeding agent is introduced by rockets or antiaircraft guns into the developing updraft and then released in the target region via flare combustion or explosion. There are both a supercooled water accumulation area and a high cloud water content zone. The supercooled water accumulation zone is mainly composed of supercooled raindrops and drizzle and has a relatively large velocity of the updraft. In the cloud water glaciation hypothesis, the collection efficiency of hail on ice crystals is lower than that on supercooled water droplets. Seeding materials can stimulate the freezing of supercooled raindrops. Seeding materials in areas with high cloud water content will result in the nucleation of many tiny ice particles and will be lifted upward by updrafts. The updraft in the plateau multicell hailstorm is generally weaker than that in the plain multicell hailstorm (or the supercell storm), so ice nuclei have enough time to grow into ice particles that can serve as artificial hail embryos, and then “compete” with natural embryos to consume supercooled raindrops and cloud droplets. For the beneficial competition hypothesis, the supercooled water content in hailstorms is the main limiting factor of hail growth. The introduction of enough competitive embryos can reduce the size of hail reaching the ground. In addition, seeding also promotes precipitation by accelerating the growth of ice-phase hydrometeors, making them fall and melt in the early updraft and finally fall to the ground as rain instead of hail (Gilbert et al. 2016). This early rainout process can influence the updraft in the convection cell or weaken strong storms by inducing rainfall nearby.
2) Seeding agent diffusion and analysis method
In addition to the seeding location, the spread range, particle concentration, and spatiotemporal distribution of the catalyst are directly related to the results when investigating macro- and microresponses of hailstorms to seeding (Zhou and Zhu 2014). Therefore, determining the extent of the seeding materials in light of the operation information, the environmental wind, and diffusion helps to determine the impact of seeding and make the analysis more reliable.
The tool for this hail-suppression operation is the WR-98 rocket, which carries the BR-91-Y silver iodide (AgI) catalyst and uses the seeding method of flare burning. Detailed information on the WR-98 rocket and AgI are listed in Tables 4 and 5, respectively. Previous studies revealed that the formation cycle of natural hail is approximately 15 min (Kennedy et al. 2001), and that approximately 90% of the nuclei can be nucleated within approximately 5 min after the catalyst is injected into the hailstorm (Feng et al. 1995). In addition, the artificial ice nuclei can grow to the same size as the natural hail embryos within 10 min. Consequently, to examine the influence of artificial seeding on the microphysical processes of hail growth, the radar parameters 15 min after the operation were chosen to study microresponse characteristics.
Technical parameters of WR-98 antihail rocket.
The catalyst diffusion pattern in hailstorms during this period is assumed to follow the simulation results proposed by Zhou and Zhu (2014). That is, seeding could be regarded as an instantaneous line source, and the extent of the significant effective region (the ice nucleation concentration was 104 per cubic meter) was approximately 3 and 4.8 km wide after 10 and 30 min, respectively. In practice, the horizontal airflow in the convective storm made the seeded zone diffuse and move downwind. Therefore, it is reasonable to employ multiline source seeding to ensure full coverage within the target region. With diffusion, turbulence, and updraft transport, the entire target area can be covered by the seeding agent. Additionally, the diffusion of the multiline source seeding is generally the same as that of the instantaneous line source seeding, with little difference in the diffusion within 30 min. In summary, the effective region diameter exceeded 3 km within 15 min after seeding and satisfied the target concentration (more than 104 per cubic meter) of the seeding agent within 3 km.
Research on the physical response of hailstorms after hail suppression is conducted by comparing a seeded cell with another unseeded cell with a similar structure (Fukuta et al. 2000). The research cell (CELL1) and the control cell (CELL2) selected in this study were two cells in the same multicell hailstorm. CELL2 was generated northeast of CELL1, and both cells had the same environmental background and similar hailstorm structure. Since the airspace allocation of CELL2 was not approved in time, the best seeding time of CELL2 (the development stage of hail cloud) had been missed. The distance between the two cell centers was approximately 13 km during the hail-suppression operation of CELL1; therefore, the development process of CELL2 could be regarded as a natural process and was compared with the macroresponse of the CELL1 operation.
Since hail-growth processes are complex and vulnerable to the environment, there exists a large error in using two similar cells to explore the microphysical changes within a hailstorm. Therefore, the variation in radar parameters of the same cell after seeding was used to analyze the microscopic physical response characteristics. It is difficult to totally separate the natural evolution of hailstorms from possible changes induced by seeding (Rivera et al. 2020). To explore the seeding response based on the natural change in hail cloud as much as possible, we classified the research cell into three regions (see section 3a). Then we analyzed the changes in radar macro- and microparameters after seeding in each region to provide a physical basis for the effect of hail suppression.
Figures shown in section 3 are the main results of this paper. Section 3 includes the macrophysical echo characteristics of CELL1 before seeding (Figs. 4 and 5), the macrophysical echo evolution of CELL1 (Fig. 6), the evolution of the dual-polarization parameters of CELL1 (Fig. 7), rocket operation information (Fig. 8), the schematic diagram of the study areas (Fig. 9), the comparison of the echo height characteristics of CELL1 and CELL2 before and after seeding (Fig. 10), the changing characteristics of the averaged dual-polarization parameters at each region of CELL1 after seeding (Fig. 11), the evolution of hydrometeors at each elevation of CELL1 after seeding (Fig. 12), the evolution of hydrometeors of CELL1 in the vertical direction after seeding (Fig. 13), and comparisons of hydrometeors between the whole hailstorm and the seeded region before and after seeding (Fig. 14).
The evolution of ZH cross section (along the strongest reflectivity) from 1013 to 1027 UTC.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
The ZH PPIs at (a) 6.2° elevation and (b) 8.7° elevation, and (c) the cross section of ZH at 1047 UTC. The red lines indicate the azimuth of the section.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Characteristics of echo evolution of the hailstorm after seeding from 1054 to 1127 UTC: (a1)–(a6) composite reflectivity and (b1)–(b6) cross sections of ZH. The red lines indicate the azimuth of the section, and the black-outlined boxes represent the locations of C0, C1, and C2.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
The (a1)–(a3) ZH, (b1)–(b3) ZDR, (c1)–(c3) KDP, and (d1)–(d3) ρHV evolutionary characteristics of CELL1 in T1a and T1b. Times (UTC) and elevations are indicated. The black-outlined boxes represent the locations of C0, C1, and C2.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Diagram of the rocket trajectory in the (a) horizontal and (b) vertical directions. The black circle represents the average effective range of the rocket, and the black arrow indicates the launch azimuth. Points A and B are the starting and ending positions of seeding, respectively.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Schematic diagram of the study area in (a) horizontal and (b) vertical perspectives.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Evolution of the hailstorm echo height before and after seeding. The shaded area is the operation time.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
The variations of mean (a) ZH, (b) ZDR, (c) KDP, and (d) ρHV, and (e) the maximum ZH, in the C0, C1, and C2 regions in T1a and T1b.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
HID results of each elevation at (a1)–(a3) 1054, (b1)–(b3) 1100, (c1)–(c3) 1107, and (d1)–(d3) 1114 UTC. Elevations are indicated. The black-outlined boxes represent the locations of C0, C1, and C2.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Vertical cross sections of HID results at (a) 1054, (b) 1100, (c) 1107, and (d) 1114 UTC. The azimuth of the section is shown in Fig. 6, and the black-outlined boxes represent the locations of C0, C1, and C2.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
Quantitative change characteristics of hydrometeors in the hailstorm before and after seeding. In the biaxial line chart, the solid line corresponds to the left axis and the dotted line corresponds to the right axis. The shaded area is the operation time.
Citation: Journal of Applied Meteorology and Climatology 62, 9; 10.1175/JAMC-D-22-0129.1
3. Results
a. Preseeding hailstorm radar characteristics and seeding information
In the hail-suppression procedure, the location of the target seeded region was determined according to the hailstorm criteria, then the design of the hail-suppression operation was decided. In this case, CELL1 was approximately 25 km away from the detection radar before operation. To clarify the macro- and microphysical characteristics in the hail-forming region, an elevation angle of 8.7° was selected for analysis.
The radar showed that the initial echo (ZH ≥ 30 dBZ) of CELL1 appeared near the −20°C layer at 1013 UTC and then gradually extended downward with the development of the hailstorm; the echo extended to the ground at 1027 UTC (Fig. 4). At 1047 UTC (Fig. 5), the echo of CELL1 was in a cluster shape, the ZH values in the strong echo center exceeded 40 dBZ, and the echo top extended above the −20°C level, indicating that the hailstorm was in the development stage. The feeder cloud was located 3–5 km in front of the main cloud, and the ZH of the strongest echo was more than 45 dBZ.
With CELL1 developing, the feeder cloud gradually merged into the main cloud. The initial hail embryos from the feeder cloud merged with the main updraft for further growth. After merging, the main updraft strengthened rapidly, and a strong echo center (ZH ≥ 45 dBZ) appeared between the −10° and −20°C layers (Fig. 6b1), indicating the presence of hail embryos in the upper layer of the multicell hailstorm.
The radar showed that the high-reflectivity region of hail clouds had typical polarization characteristics of hail-forming regions (−0.5 ≤ ZDR ≤ 1.5 dB, KDP ≤ 0.5° km−1, and 0.70 ≤ ρHV ≤ 0.96), suggesting that this region was mainly composed of small hailstones and low-density graupel (Figs. 7a1–d1) (Picca and Ryzhkov 2012; Kennedy et al. 2014; Yang et al. 2020; Cao et al. 2021).
The regions with large ZH and small ρHV also include some areas with large KDP and ZDR, which may mainly be composed of a mixture of rain and hailstone. In addition, the middle of the hail cloud updraft contained a supercooled water accumulation zone with ZDR greater than 0 and ρHV approximately 0.92. This resembled the type I mixed-phase region in the hail cloud studied by Zhao et al. (2020). The type I mixed-phase region mostly had large rain droplets and mixtures of rain and hail, and the wet growth of HDG occurred in this region. According to the historical hail statistics of Weining and the hail observation of the same type of hailstorm on that day, we estimated that CELL1 would produce hail with a size of about 8–10 mm without hail suppression. After seeding, we estimated that the hail size would decrease to about 1–5 mm or only rainfall. Therefore, according to the radar detection, the performance of WR-98 rockets, operational technical points, and the high-altitude wind conditions of Weining, the operators determined the azimuth, elevation angle, shell amount, and other information of this hail-suppression operation, as shown in Table 6. Figure 8 shows the approximate horizontal and vertical trajectories of the rocket in the target area. The AgI was injected and diffused along the rocket trajectory, and the seeding length was approximately 5 km, passing through a 3-km height of the cloud.
Artificial hail-suppression operation information of the Yangjie station.
According to the diffusion pattern of the seeding agent in section 2d(2) and the seeding trajectory, we define the core of the storm where seeding occurred as C0, the region where seeding agent diffuses as C1, and the peripheral region (not affected by seeding) as C2. The C0, C1, and C2 regions in CELL1, as shown in Fig. 9, have horizontal and vertical widths of 1.5, 3, and 6 km and heights of 2, 3, and 4 km, respectively. Note that the C1 region did not contain C0, and the C2 region did not contain C0 and C1. We defined the operation time, early seeding period (1054–1100 UTC), late seeding period (1100–1107 UTC), and postseeding period (after 1107 UTC) as T0, T1a, T1b, and T2, respectively.
b. Macrophysical response characteristics of the hailstorm after seeding
The radar parameters such as echo top (ET; ZH ≥ 18 dBZ), echo extension heights of 30 dBZ (TOP) and 45 dBZ (45ET), and hail layer thickness (TOP0) can indicate the development of hail. When CELL1 began convection, the strong echo center (ZH ≥ 45 dBZ) was above the 0°C layer and then extended downward. The 45ET exhibited rapid growth from 1027 to 1047 UTC after the echo extended to the ground (Fig. 10).
As shown in Fig. 6b1, the cloud developed with a tilt to approximately 7.5 km, and the echo height with ZH values higher than 45 dBZ extended to near the −20°C layer, indicating a great hailfall risk. The hail-suppression operation to the target area was conducted in Yangjie at 1056 UTC (T0). After seeding, the echo area (ZH > 30 dBZ) significantly decreased, and the strong echo center weakened and gradually dispersed (Figs. 6a2–a6).
As shown in Fig. 10, within T1a and T1b, the ET and TOP of CELL1 lowered by 1.47 and 1.36 km, respectively. The 45ET decreased rapidly from 5.33 km to approximately 1.27 km, and TOP0 decreased from 3.52 to 2.16 km. The declining rate of the 45ET was 338 m·min−1, which was the most obvious change in the macroscopic response of the hailstorm after seeding. In contrast, the declining rates of TOP and ET were relatively small, approximately 113 and 123 m·min−1, respectively. The declining rate of 45ET decreased significantly in T2 (it was only 71 m·min−1), while the declining rates of TOP and ET increased to 214 and 186 m·min−1, respectively. The changing rate of the echo height in the three periods after seeding was obviously different. The 45ET mainly lowered in T1a and T1b, while ET and TOP rapidly decreased in T2.
In comparison with CELL1, CELL2 had a similar development process in the early stage. However, different from the change in echo height after seeding in CELL1, the height declining rates of ET, TOP, and 45ET in CELL2 during two volume-scanning times (1054–1107 UTC) were slightly different, and were 148, 156, and 283 m·min−1, respectively. In addition, the 45ET and TOP of CELL2 remained at a relatively high height, which developed again in a favorable synoptic background.
In contrast with CELL2, the strong echo height of CELL1 decreased faster during T1a and T1b, while ET and TOP decreased more slowly. After natural precipitation, ET, TOP, and 45ET of CELL2 decreased rapidly (Fig. 10). The results show that the echo height variation of the seeded hailstorm was different from that of the natural hailstorm. However, there are differences in the natural evolution process of different hailstorms. The difference in the macroscale variation of the hailstorm mentioned above may be due to more than just seeding. Whether the seeding material will have an impact on the hail cloud cannot be confirmed only by macroscale observation. Therefore, in sections 3c and 3d, we further explore the seeding effect of the hailstorm through the variation in radar polarization parameters and particle classification.
c. Microphysical response characteristics of the hailstorm after seeding
The dual-polarization parameters can represent the phase state, size, and shape of hydrometeors in hailstorms to a certain degree; thus, the microphysical changes in hailstorms after seeding can be explored by analyzing the changes in these parameters. Previous studies on the dual-polarization characteristics of natural hailstorms showed that the combination of high ZH, near-zero ZDR, low ρHV, and a negative ZDR area (from −2 to 0 dB) above the ZDR column can be used to identify the hail-forming region in the cloud (Kumjian 2013a,b; Kumjian et al. 2014). The height of the hailstorm’s convective center gradually decreases with the weakening of the hailstorm. To clearly show the changes in radar parameters of the hailstorm’s convective center before and after seeding, the radar elevation angle selected for analysis also decreased with the decrease in the hailstorm’s height. According to the distance between CELL1 and the radar, we chose the observation data of 8.7°, 7.5°, and 5.25° elevation angles to show changes in dual-polarization characteristics in the cloud after seeding.
In T1a, the ET and hailstorm centroid decreased in height. At the 7.5° elevation, ZDR and KDP in the center of the hailstorm significantly increased, and ρHV ranged from 0.94 to 0.99 (Figs. 7a2–d2). As shown in Fig. 11, the results of dual-polarization parameters showed that the mean ZH and maximum ZH in C0, C1, and C2 decreased, among which C2 decreased the most. The mean ZDR of all three regions increased, and the amplitude was basically the same. The mean KDP decreased in C0 and increased in C1 and C2, while the mean ρHV increased in C0 and C1 and decreased in C2. In T1b, the hailstorm echo height decreased continually.
At a 5.25° elevation, the maximum ZH was lower than 40 dBZ, the range of ZDR and KDP areas of high values simultaneously increased, and ρHV exceeded 0.97 (Figs. 7a3–d3). As shown in Fig. 11, the decrease in the mean ZH and maximum ZH in C0 exceeded that in C1 and C2. The increasing amplitude of ZDR in the three regions was similar to that of the previous period, of which C1 had the largest increasing amplitude. The trend of the mean KDP in C0 remained unchanged, but the amplitude weakened. The trends of the mean KDP in C1 and C2 decreased and were opposite to those of the previous period. The mean ρHV in C0 and C1 increased continuously, with a larger increase in C0. The mean ρHV in C2 decreased continuously.
Therefore, in comparison with C0 and C1, the decreasing amplitude of the mean and maximum ZH and the increasing amplitude of the mean ZDR in the seeded period (T1a and T1b) in C2 remained basically unchanged. Additionally, the changing trend of the mean ρHV in C0 and C1 was the opposite of that in C2. However, the parameter changes of C0 and C1 were similar, which in T1b were more pronounced than those in T1a.
Because of the explosion effect caused by the burning flare and the artificially introduced ice nuclei, the development of the updraft was disrupted. The reflectivity of C0 and C1 decreased slightly in T1a and decreased sharply in T1b. Generally, the explosion effect is a faster reaction than beneficial competition (Xu 2001), which is difficult to capture by radar with a volume scanning period of about 6 min. In addition, KDP is a good indicator of the change in cloud water content (Homeyer and Kumjian 2015; Wilson and Van Den Broeke 2021). The mean KDP of C0 and C1 continuously decreased during T1a and T1b, indicating a further decrease in water content there. However, the reflectivity did not increase, suggesting that artificial ice nuclei consumed liquid cloud droplets and raindrops to produce more small ice particles, thereby inhibiting the continuous growth of hailstones.
At the later stage of seeding, a large number of ice particles fell from the upper part of the hailstorm. Some of the ice particles with a small size melted completely into raindrops, and the others with a larger size may have only partially melted. The surface of these partially melted particles was covered with water film, leading to the increase in ZDR and ρHV in C0 and C1. At this time, there were mostly rain areas containing small ice particles near the 0°C layer. The drag effect and the evaporative cooling of precipitation particles strengthened the sinking movement in the cloud, and rainfall was observed on the ground.
d. Variation characteristics of hydrometeors in the hailstorm after seeding
To intuitively analyze the changes in hydrometeors with time in C0, C1, and C2 after seeding, as well as the differences between these regions, we used the HID method to give a classified display of hydrometeors. In addition, we chose the region that was centered at the convective center of the hailstorm and had the same size as C1 at 1047 and 1114 UTC to highlight the change in the number of range bins dominated by each hydrometeor before and after seeding.
The HID results showed that at the beginning of the operation (Figs. 12a1–a3 and 13a), AG was mainly distributed in the weak echo area at the edge of the 2–5-km hail cloud. The convective core of the hailstorm was mainly composed of LDG and HDG, while RN and DZ were below the 0°C layer. In addition, RN above the 0°C layer was considered supercooled water in the HID results (Zhao et al. 2020). At this time, a small amount of RH existed in the middle and upper parts of the hail-forming region, which was consistent with the echo position of ZH greater than 45 dBZ in the vertical profile (Fig. 6b1). In T1a, RH fell from 5 km into the supercooled water accumulation zone in the layers from −5° to −10°C (Fig. 13b). In T1b, RH fell below the study region (Fig. 13c).
Table 7 shows the change in the number of range bins dominated by each hydrometeor in C0, C1, and C2 (boldface and italic fonts indicate decrease and increase, respectively). As shown in Table 7 for C0, in T1a, RH, HDG, and AG decreased, and LDG and SWA increased. In T1b, RH continued to decrease and LDG and SWA decreased sharply. AG and CR increased sharply. In the early stage of T2 (1107–1114 UTC), SWA and LDG increased and AG and CR decreased. By comparison, the change of hydrometeors in T1b is significantly different from the other two periods.
Variation in the number of range bins dominated by each hydrometeor in C0, C1, and C2 (along the strongest reflectivity profile) during periods T1a, T1b, and the early stage of T2 (1107–1114 UTC). The unit is gates per·6 minutes; boldface and italic fonts indicate decrease and increase, respectively.
As shown in Table 7 for C1, in T1a, RH and SWA increased. LDG, HDG, AG, and CR decreased. In T1b, RH, SWA, LDG, and HDG decreased and AG and CR increased sharply. Note that the change in SWA, LDG, AG, and CR in C1 is consistent with that in C0. In the early stage of T2 (1107–1114 UTC), SWA, AG, and CR decreased, and LDG increased. As shown in Table 7 for C2, SWA and LDG both decreased in the three periods. In T1b, the decreasing amplitudes of SWA and LDG in C2 were significantly less than those in C0 and C1. The increasing amplitudes of AG and CR in C2 were also significantly less than those in C0 and C1.
Table 7 indicates that the changes in hydrometeors in T1a and T1b were consistent with the polarization parameters in section 3c. The change in the number of range bins dominated by each hydrometeor in the seeded regions mainly occurred within T1b, which was consistent with the time required for the AgI to complete nucleation and start to take effect. Artificial ice nuclei consumed SWA in the seeded region, inhibited the growth of RH and LDG, and generated a large number of CR and AG with small particle sizes. RH and LDG with a larger size fell into the accumulation zone before other particles, causing a significant increase in the number of range bins dominated by HDG near the 0°C layer. Some LDG was converted to HDG at this time (Dye et al. 1974). With most particles reaching precipitation size, the sedimentation of a large number of ice particles weakened the updraft (Farley 1987). Moreover, the number of range bins dominated by each hydrometeor in C2 as compared with C0 and C1 had an opposite trend (or had a trend with smaller amplitude), which further showed that the artificial ice nuclei had an influence on microphysical processes in the hailstorm seeded region.
In the early stage of T2 (1107–1114 UTC), the artificial ice nuclei were gradually exhausted, and large amounts of CR were blown to the cloud top or formed AG. AG further grew into LDG, and the consumption rate of SWA decreased (Table 7; Fig. 13d). These findings explain that the ET and TOP in CELL1 decrease more slowly than in CELL2, as mentioned in section 3b, because a large number of ice crystals and aggregates formed by seeding maintain a high cloud-top height. The changes in the microphysical processes and thermodynamics in clouds lead to changes in clouds at the macrolevel.
To better show the evolution of hydrometeors with time after seeding, we compared the change in hydrometeors in the seeded regions (C0 and C1) with that in the whole hailstorm before and after seeding, as shown in Fig. 14. In T1a and T1b, the number of range bins dominated by RH, SWA, LDG, and HDG in C0 and C1 decreased by 100%, 53.4%, 41.1%, and 76.6% at rates of 43, 42.8, 112.7, and 6 gate·min−1, respectively. The number of range bins dominated by AG and CR increased by 53% and 119.4%, at rates of approximately 60.2 and 60.9 gate·min−1, respectively. Different from the changes in hydrometeors in the seeded region, the number of range bins dominated by every hydrometeor in the whole hailstorm decreased after seeding. However, in C0 and C1, there were marked changes in the number of range bins dominated by LDG, HDG, and SWA caused by the introduction of artificial ice nuclei, especially the increase in the number of range bins dominated by AG and CR. Therefore, we believe that after seeding, the AgI caused the nucleation of a large number of small ice crystals, which inhibited the growth of hailstones and graupel embryos by competing with natural hail embryos for supercooled liquid water and cloud droplets and generated large amounts of ice crystals and aggregates. The hailstones and other large ice particles fell and melted before further growth, leading to early rainout. Hydrometeors consumed supercooled water in the accumulation zone and avoided the occurrence of hail.
The above results are the observations and preliminary analysis of the hail-suppression effect of a multicell hailstorm in Weining. As the observation and analysis approaches become more refined, more case studies will help to deepen the understanding of hail growth and its suppression.
4. Discussion
For ordinary multicell hailstorms (viz., the case in this paper), the updraft is weaker and has a shorter duration than supercell hailstorms (or the stronger multicell hailstorms), and the time for ice particles to grow into hail embryos in the updraft is limited. Therefore, most hail embryos form at the front edge of the weak echo region of the main cell with a weak updraft and are then transported through turbulence to the main updraft for further growth into hail. Surrounded by low-density graupel and supercooled cloud droplets, hail embryos suspended near the −20°C layer grow into hailstones by collision freezing. When the hailstone grows, it will fall into the supercooled water accumulation zone and then further grow into large hail that can fall to the ground.
Thus, hail suppression for this kind of hailstorm mainly follows the concepts of cloud water glaciation and beneficial competition. The location for introducing artificial ice nuclei should be the high-level hail embryo growth area and the low-level supercooled water accumulation zone in the updraft. After entering the cloud, the artificial ice nuclei nucleate rapidly, freezing the cloud droplets to form ice crystals and frozen droplets. The collection efficiency of hail on ice crystals is lower than that of supercooled water droplets (Dennis 1977), which first limits the growth of hailstones in the high layer. The artificial ice nuclei seeded in the supercooled water accumulation zone can quickly freeze the raindrops and generate millimeter-sized frozen embryos, which can act as competitive embryos to consume for the available supercooled water. Hail production is reduced due to cloud seeding. However, radar observation is limited by time resolution and space resolution. The interpretation of cloud microphysical processes based on radar observation still lacks support from direct observation. In situ microphysical observations and numerical simulation should also be included to further validate the seeding effect.
For the seeding method of direct injection in the updraft, as in this paper, it is crucial to correctly identify the hail area in the mixed-phase precipitation cloud for hail warning and hail suppression. Because of the different hydrometeor properties (phase state, shape, size, orientation, and fall behavior) of particles between hail and rain areas, high ZH and low ZDR are used to identify hail areas using dual polarization (Bringi et al. 1984; Aydin et al. 1986). Additionally, based on ZH, ZDR, KDP, LDR, and ρHV, the fuzzy-logic hydrometeor classification method is utilized to identify and classify the ice particles in the hail area in a more detailed way, which has been validated by the aircraft observation results (Vivekanandan et al. 1999; Liu and Chandrasekar 2000; Lim et al. 2005). Due to a shorter wavelength, the X-band radar is more sensitive to the shape and arrangement of particles in the cloud. The HID method of X-band dual-polarization radar was proposed over the past decade (Dolan and Rutledge 2009; Snyder et al. 2010). In this paper, both the improved HID method and polarization parameters are employed to identify the location of hail in the mixed-phase region. They help to accurately identify the location of the target operation area in the updraft and a hail-suppression effect.
Polarization parameters are different at different wavelengths (Kumjian and Ryzhkov 2008). Therefore, the polarization characteristics are retained in this paper for comparison with previous studies. The HID method makes the particle distribution and the particle changes more physically intuitive to interpret. However, it is important to set the weight coefficients properly. The weights are generally determined subjectively (Vivekanandan et al. 1999). Thus, more cloud microphysical observations are needed to optimize the identification.
In addition, hail growth is a finely balanced mechanism controlled by cloud dynamics and microphysics and their interactions (Browning 1977). Due to synergistic effects between microphysical mechanisms and the thermodynamics of hailstorms, the latent heat released by the process of the artificial ice nucleus freezing the supercooled droplets leads to the enhancement of the vertical velocity of the updraft and then further increases the accumulation of supercooled water (Ilotoviz et al. 2018). This process may create more uncertainty for hail suppression if the mechanism of hail growth is not fully understood or the optimal operation opportunity is not seized. In multicell hailstorms, the cells usually interact with each other and easily split after hail suppression. In comparison with the unseeded cells, the seeded cells are obviously weakened after seeding, but a large number of artificially introduced ice embryos may provide embryonic sources for the surrounding cells when the main body of the storm weakens, which results in a new hailfall risk. Therefore, the dynamic impacts induced by microphysical changes in hail clouds should also be considered in the future. Results presented in this study are not necessarily definitive, but as more and more cases are studied with more refined techniques, better understanding will result. Based on a further study of the hail-growth mechanism and hail-suppression mode, as well as comprehensive summaries obtained from the analyses in actual operations, we can further explore a more effective method for multicell hailstorm suppression.
5. Conclusions
In this study, an improved HID method was used for hydrometeor identification and classification after data preprocessing from the YLD1-D radar. Based on the above work, the macro- and microphysical response characteristics of the multicell hailstorm after seeding on 7 April 2019 were analyzed.
The weak updraft in the early stage of the hailstorm (CELL1) allowed enough time for artificial ice nuclei to play their role, which improved the seeding effect. At the beginning of the operation in the target area, ZDR ranged from −0.5 to 2.5 dB, ρHV ranged from 0.70 to 0.97, and KDP was less than 1° km−1, indicating that the target area may have been a mixed area of small hailstones and supercooled raindrops. The identification results showed hail, supercooled water, and low-density graupel. After seeding, ZH decreases, and ZDR and ρHV increase in the seeded region, which suggests that hydrometeors changed into graupel, ice crystals, and aggregates. At this time, the identification results show that large amounts of ice crystals and aggregates were generated in the original location. Hail and some low-density graupel fell into the supercooled water accumulation zone and formed more high-density graupel via riming, inducing a temporary increase in the number of range bins dominated by HDG. When hail and high-density graupel grow to a terminal fall velocity greater than the updraft velocity, they fall along the back of the updraft, and rainfall occurs on the ground. Precipitation reduces the total water content of the clouds and the supercooled water volume that can reach the hail-forming region, thereby eliminating hailstones or lessening the size of hailstones that can reach the ground.
By comparing the polarization parameters and particle changes in different regions and periods, we find that the decrease in low-density graupel and supercooled water, and the increase in ice crystal and aggregate, mainly occur in T1a and T1b, especially within T1b. This period is consistent with the time required for the beneficial competition after the AgI completes nucleation. For different regions during T1a and T1b, change in polarization characteristics and hydrometeors in C0 and C1 is consistent, which further suggests that seeding influences the microphysical process of the hail cloud.
In addition, there are differences in the macrodevelopment between the seeded hailstorm and the unseeded hailstorm. After the natural development and precipitation of hail in the unseeded hailstorm, both the echo top and the height of the strong echo decrease rapidly. In the seeded hailstorm, seeding caused a large number of light ice crystals to be brought to the high layer by the updraft, which is shown in the macroscopic detection as a relatively high cloud top that lasted for some time after seeding. Macro- and microevolution and differences between different regions and periods of the hailstorm were consistent with theories such as cloud water glaciation, beneficial competition, and early rainout for hail suppression of multicell hailstorms with a large number of supercooled raindrops.
Acknowledgments.
This research was jointly supported by National Natural Science Foundation of China Project (41875169), the Second Qinghai-Tibet Plateau Comprehensive Scientific Research Project (2019QZKK0104), Sichuan Science and Technology Program Project (2022YFS0545), Guizhou Science and Technology Program Project Qiankehe Support [(2023) general item 193], Guizhou Science and Technology Program Project Qiankehe Support [(2022) 206], Yunnan Science and Technology Program Project (202203AC100006), and Key Laboratory for Cloud Physics of China Meteorological Administration (2020Z007).
Data availability statement.
The data are available by contacting the corresponding author.
REFERENCES
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.
Bi, Y. H., J. L. Liu, S. Duan, D. R. Lv, D. B. Su, and Y. C. Chen, 2012: Attenuation correlation of reflectivity for X-band dual-polarization radar (in Chinese with English abstract). Chin. J. Atmos. Sci., 36, 495–506.
Bringi, V. N., T. A. Seliga, and W. A. Cooper, 1984: Analysis of aircraft hydrometeor spectra and differential reflectivity (ZDR) radar measurements during the Cooperative Convective Precipitation Experiment. Radio Sci., 19, 157–167, https://doi.org/10.1029/RS019i001p00157.
Browning, K. A., 1977: The structure and mechanisms of hailstorms. Hail: A Review of Hail Science and Hail Suppression, Meteor. Monogr., No. 16, Amer. Meteor. Soc., 1–47.
Browning, K. A., and D. Atlas, 1977: Some new approaches in hail suppression experiments. J. Appl. Meteor., 16, 327–332, https://doi.org/10.1175/1520-0450(1977)016<0327:SNAIHS>2.0.CO;2.
Cao, S. Y., W. Sun, F. F. Wei, P. F. Shen, and Y. He, 2021: Study of “7.6” hail event in Jiangsu based on dual-polarization weather radar observations (in Chinese with English abstract). Daqi Kexue Xuebao, 44, 549–557.
Chen, Y. C., L. B. Zhang, Y. L. Jin, J. L. Ma, L. J, and R. Qin, 2016: A case study of effectiveness of artificial hail suppression based on dual polarization radar data (in Chinese with English abstract). Mater. Sci. Technol., 44, 479–488.
Cheng, L., and D. C. Rogers, 1988: Hailfalls and hailstorm feeder clouds—An Alberta case study. J. Atmos. Sci., 45, 3533–3545, https://doi.org/10.1175/1520-0469(1988)045<3533:HAHFCA>2.0.CO;2.
Ćurić, M., 1990: One evidence of the hail suppression efficiency derived from radar measurements. J. Wea. Modif., 22, 79–81.
Dennis, A. S., 1977: Hail suppression concepts and seeding methods. Hail: A Review of Hail Science and Hail Suppression, Meteor. Monogr., No 16, Amer. Meteor. Soc., 181–193.
Dessens, J., J. L. Sánchez, C. Berthet, L. Hermida, and A. Merino, 2016: Hail prevention by ground-based silver iodide generators: Results of historical and modern field projects. Atmos. Res., 170, 98–111, https://doi.org/10.1016/j.atmosres.2015.11.008.
Detwiler, A. G., 2002: Some reflections on hailstorms and hail suppression. J. Wea. Modif., 34, 64–72, https://doi.org/10.54782/jwm.v34i1.454.
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.
Dye, J. E., C. A. Knight, V. Toutenhoofd, and T. W. Cannon, 1974: The mechanism of precipitation formation in northeastern Colorado cumulus III. Coordinated microphysical and radar observations and summary. J. Atmos. Sci., 31, 2152–2159, https://doi.org/10.1175/1520-0469(1974)031<2152:TMOPFI>2.0.CO;2.
Farley, R. D., 1987: Numerical modeling of hailstorms and hailstone growth. Part III: Simulation of an Alberta hailstorm—Natural and seeded cases. J. Climate Appl. Meteor., 26, 789–812, https://doi.org/10.1175/1520-0450(1987)026<0789:NMOHAH>2.0.CO;2.
Feng, D. X., R. Z. Chen, G. E. Jiang, B. H. Luo, and Y. S. Cui, 1995: The high efficient AgI pyrotechnics and their ice nucleating properties (in Chinese with English abstract). Acta Meteor. Sin., 53, 568–579.
Foote, G. B., and C. A. Knight, 1979: Results of a randomized hail suppression experiment in northeast Colorado. Part I: Design and conduct of the experiment. J. Appl. Meteor., 18, 1526–1537, https://doi.org/10.1175/1520-0450(1979)018<1526:ROARHS>2.0.CO;2.
Foote, G. B., and C. G. Wade, 1982: Case study of a hailstorm in Colorado. Part I: Radar echo structure and evolution. J. Atmos. Sci., 39, 2828–2846, https://doi.org/10.1175/1520-0469(1982)039<2828:CSOAHI>2.0.CO;2.
Fukuta, N., K. Wakimizu, K. Nishiyama, Y. Suzuki, and H. Yoshikoshi, 2000: Large, unique radar echoes in a new, self-enhancing cloud seeding. Atmos. Res., 55, 271–273, https://doi.org/10.1016/S0169-8095(00)00065-X.
Gavrilov, M. B., L. Lazić, A. Pešic, M. Milutinović, D. Marković, A. Stanković, and M. M. Gavrilov, 2010: Influence of hail suppression on the hail trend in Serbia. Phys. Geogr., 31, 441–454, https://doi.org/10.2747/0272-3646.31.5.441.
Gilbert, D. B., B. A. Boe, and T. W. Krauss, 2016: Twenty seasons of airborne hail suppression in Alberta, Canada. J. Wea. Modif., 48, 68–92, https://doi.org/10.54782/jwm.v48i1.551.
Goyer, G. G., 1975: Time-integrated radar echo tops as a measure of cloud seeding effects. J. Appl. Meteor., 14, 1362–1365, https://doi.org/10.1175/1520-0450(1975)014<1362:TIRETA>2.0.CO;2.
Heinselman, P. L., and A. V. Ryzhkov, 2006: Validation of polarimetric hail detection. Wea. Forecasting, 21, 839–850, https://doi.org/10.1175/WAF956.1.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Heymsfield, A. J., A. R. Jameson, and H. W. Frank, 1980: Hail growth mechanisms in a Colorado storm: Part II: Hail formation processes. J. Atmos. Sci., 37, 1779–1807, https://doi.org/10.1175/1520-0469(1980)037<1779:HGMIAC>2.0.CO;2.
Homeyer, C. R., and M. R. Kumjian, 2015: Microphysical characteristics of overshooting convection from polarimetric radar observations. J. Atmos. Sci., 72, 870–891, https://doi.org/10.1175/JAS-D-13-0388.1.
Huang, H. X., Y. J. Zhou, Y. Zeng, S. P. Zou, and Z. Yang, 2021: Study on the evolution characteristics of hail embryos in Weining, Guizhou, based on X-band dual linear polarization radar (in Chinese with English abstract). Chin. J. Atmos. Sci., 45, 539–557, https://doi.org/10.3878/j.issn.1006-9895.2005.20105.
Ilotoviz, E., A. Khain, A. V. Ryzhkov, and J. C. Snyder, 2018: Relationship between aerosols, hail microphysics, and ZDR columns. J. Atmos. Sci., 75, 1755–1781, https://doi.org/10.1175/JAS-D-17-0127.1.
Ke, L. P., J. Liu, M. Xie, and Y. Zhang, 2016: Research on forecast index of Weining hail weather (in Chinese with English abstract). J. Guizhou Meteor., 40, 14–19.
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.
Kennedy, P. C., S. A. Rutledge, B. Dolan, and E. Thaler, 2014: Observations of the 14 July 2011 Fort Collins hailstorm: Implications for WSR-88D-based hail detection and warnings. Wea. Forecasting, 29, 623–638, https://doi.org/10.1175/WAF-D-13-00075.1.
Krauss, T. W., 1999: Radar characteristics of seeded and non-seeded hailstorms in Alberta, Canada. Seventh WMO Scientific Conf. on Weather Modification, Chiang Mai, Thailand, WMO, 17–22.
Krauss, T. W., R. T. Bruintjes, and H. Martinez, 2000: A new hail suppression project using aircraft seeding in Argentina. J. Wea. Modif., 32, 73–80.
Kumjian, M. R., 2013a: 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., 2013b: Principles and applications of dual-polarization weather radar. Part II: Warm- and cold-season applications. J. Oper. Meteor., 1, 243–264, https://doi.org/10.15191/nwajom.2013.0120.
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., 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.
Li, L. Y., 1996: The effect analysis on hail suppression by antihail gun using radar echo variations (in Chinese with English abstract). Meteor. Mon., 22, 26–30.
Lim, S., V. Chandrasekar, and V. N. Bringi, 2005: Hydrometeor classification system using dual-polarization radar measurements: Model improvements and in situ verification. IEEE Trans. Geosci. Remote Sens., 43, 792–801, https://doi.org/10.1109/TGRS.2004.843077.
Liu, H., and V. Chandrasekar, 2000: Classification of hydrometeors based on polarimetric radar measurements: Development of fuzzy logic and neuro-fuzzy systems, and in situ verification. J. Atmos. Oceanic Technol., 17, 140–164, https://doi.org/10.1175/1520-0426(2000)017<0140:COHBOP>2.0.CO;2.
Liu, Z. W., S. Q. Tian, F. J. Wang, and D. L. Gong, 2020: Identification of convection cells and the effect of hail suppression during a hail event (in Chinese with English abstract). J. Meteor. Environ, 36, 10–16.
Lu, P. Y., 2011: Technical requirements of rockets on hail suppression (in Chinese with English abstract). Anhui Nongye Kexue, 39, 14 960–14 962.
Makitov, V. S., V. S. Inyukhin, H. M. Kalov, and R. H. Kalov, 2017: Radar research of hailstorm formation and development over the central part of Northern Caucasus (Russia). Organization and main results of the regional hail suppression projects. Nat. Hazards, 88, 253–272, https://doi.org/10.1007/s11069-016-2433-7.
Miller, L. J., and J. C. Fankhauser, 1983: Radar echo structure, air motion and hail formation in a large stationary multicellular thunderstorm. J. Atmos. Sci., 40, 2399–2418, https://doi.org/10.1175/1520-0469(1983)040<2399:RESAMA>2.0.CO;2.
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.
Park, S. G., V. N. Bringi, V. Chandrasekar, M. Maki, and K. Iwanami, 2005: Correction of radar reflectivity and differential reflectivity for rain attenuation at X band. Part I: Theoretical and empirical basis. J. Atmos. Oceanic Technol., 22, 1621–1632, https://doi.org/10.1175/JTECH1803.1.
Pflaum, J. C., 1980: Hail formation via microphysical recycling. J. Atmos. Sci., 37, 160–173, https://doi.org/10.1175/1520-0469(1980)037<0160:HFVMR>2.0.CO;2.
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.
Rivera, J. A., F. Otero, E. N. Tamayo, and M. Silva, 2020: Sixty years of hail suppression activities in Mendoza, Argentina: Uncertainties, gaps in knowledge and future perspectives. Front. Environ. Sci., 8, 45, https://doi.org/10.3389/fenvs.2020.00045.
Schickedanz, P. T., and S. A. Changnon, 1970: The design and evaluation of hail suppression experiments. Mon. Wea. Rev., 98, 242–251, https://doi.org/10.1175/1520-0493(1970)098<0242:TDAEOH>2.3.CO;2.
She, Y., L. Chen, and Y. Zhang, 2019: Design of the artificial operation directing-system based on weather radar and antiaircraft-gun & rocket for hail suppression. 2019 Int. Conf. on Meteorology Observations (ICMO), Chengdu, China, Institute of Electrical and Electronics Engineers, 1–4, https://doi.org/10.1109/ICMO49322.2019.9026038.
Sioutas, M., 2017: Hail characteristics and cloud seeding effect for hail suppression in central Macedonia, Greece. Perspectives on Atmospheric Sciences, T. Karacostas, A. Bais, and P. T. Nastos, Eds., Springer Atmospheric Sciences, Springer, 271–277, https://doi.org/10.1007/978-3-319-35095-0_38.
Smith, P. L., D. J. Musil, A. G. Detwiler, and R. Ramachandran, 1999: Observations of mixed-phase precipitation within a CaPE thunderstorm. J. Appl. Meteor., 38, 145–155, https://doi.org/10.1175/1520-0450(1999)038<0145:OOMPPW>2.0.CO;2.
Snyder, J. C., H. B. Bluestein, G. F. Zhang, and S. J. Frasier, 2010: Attenuation correction and hydrometeor classification of high-resolution, X-band, dual-polarized mobile radar measurements in severe convective storms. J. Atmos. Oceanic Technol., 27, 1979–2001, https://doi.org/10.1175/2010JTECHA1356.1.
Snyder, J. C., H. B. Bluestein, V. Venkatesh, and S. J. Frasier, 2013: Observations of polarimetric signatures in supercells by an X-band mobile Doppler radar. Mon. Wea. Rev., 141, 3–29, https://doi.org/10.1175/MWR-D-12-00068.1.
Sulakvelidze, G. K., N. S. Bibilashvili, and V. F. Lapcheva, 1967: Formation of Precipitation and Modification of Hail Processes. Israel Program for Scientific Translations, 208 pp.
Ventura, J. F. I., F. Honoré, and P. Tabary, 2013: X-band polarimetric weather radar observations of a hailstorm. J. Atmos. Oceanic Technol., 30, 2143–2151, https://doi.org/10.1175/JTECH-D-12-00243.1.
Vivekanandan, J., D. S. Zrnic, S. M. Ellis, R. Oye, A. V. Ryzhkov, and J. Straka, 1999: Cloud microphysics retrieval using S-band dual-polarization radar measurements. Bull. Amer. Meteor. Soc., 80, 381–388, https://doi.org/10.1175/1520-0477(1999)080<0381:CMRUSB>2.0.CO;2.
Wang, Q. W., and Z. M. Tan, 2014: Multi‐scale topographic control of southwest vortex formation in Tibetan Plateau region in an idealized simulation. J. Geophys. Res. Atmos., 119, 11 543–11 561, https://doi.org/10.1002/2014JD021898.
Wang, Y. Z., and Q. Yu, 1995: Research on multi-physical parameters examining hail suppression results (in Chinese with English abstract). Meteor. Mon., 21, 3–8.
Wieringa, J., and I. Holleman, 2006: If cannons cannot fight hail, what else? Meteor. Z., 15, 659–669, https://doi.org/10.1127/0941-2948/2006/0147.
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.
WMO, 1996: Meeting of experts to review the present status of hail suppression. WMO/TD-764, WMP Rep. 26, 39 pp., https://library.wmo.int/index.php?lvl=notice_display&id=11939.
Xu, H. B., 2001: The possible dynamic mechanism of explosion in hail suppression (in Chinese with English abstract). Acta Meteor. Sin., 59, 66–76.
Xu, H. B., and Y. Duan, 2001: The mechanism of hailstone’s formation and the hail-suppression hypothesis: “Beneficial competition” (in Chinese with English abstract). Chin. J. Atmos. Sci., 25, 277–288, https://doi.org/10.3878/j.issn.1006-9895.2001.02.14.
Yang, J., Y. Y. Zheng, and F. Xu, 2020: An analysis of a hail case over the Yangtze and Huai River basin based on dual-polarization radar observations (in Chinese with English abstract). Acta Meteor. Sin., 78, 568–579, https://doi.org/10.11676/qxxb2020.031.
Young, K. C., 1977: A numerical examination of some hail suppression concepts. Hail: A Review of Hail Science and Hail Suppression, Meteor. Monogr., No. 16, Amer. Meteor. Soc., 195–214.
Zeng, Y., S. P. Zou, S. Cao, L. Chen, and Y. Huang, 2018: Analysis of spatial and temporal variation characteristics of hail from 1997 to 2017 in Weining of Guizhou province (in Chinese with English abstract). Plateau Mt. Meteor. Res., 38, 23–27 + 96.
Zhao, C. H., Y. J. Zhou, H. Xiao, P. G. Zhao, X. L. Zhang, and T. J. Hu, 2019: A study of method for filtering copolar differential phase of X-band dual-polarimetric Doppler weather radar (in Chinese with English abstract). Chin. J. Atmos. Sci., 43, 285–296, https://doi.org/10.3878/j.issn.1006-9895.1805.17289.
Zhao, C. H., Y. J. Zhang, D. Zheng, Y. J. Zhou, H. Xiao, and X. L. Zhang, 2020: An improved hydrometeor identification method for X-band dual-polarization radar and its application for one summer hailstorm over northern China. Atmos. Res., 245, 105075, https://doi.org/10.1016/j.atmosres.2020.105075.
Zhao, D., R. W. Yang, Y. Tao, W. K. Zhang, and X. C. He, 2019: Objective detection of the Kunming quasi-stationary front. Theor. Appl. Climatol., 138, 1405–1418, https://doi.org/10.1007/s00704-019-02894-w.
Zhou, Y. Q., and B. Zhu, 2014: Study on diffusion regularity and operation design of antiaircraft-gun, rocket and plane cloud seeding (in Chinese with English abstract). Meteor. Mon., 40, 965–980.
Ziegler, C. L., P. S. Ray, and N. C. Knight, 1983: Hail growth in an Oklahoma multicell storm. J. Atmos. Sci., 40, 1768–1791, https://doi.org/10.1175/1520-0469(1983)040<1768:HGIAOM>2.0.CO;2.