Electrical Alignment Signatures of Ice Particles Before Intracloud Lightning Activity Detected by Dual‐Polarized Phased Array Weather Radar

The cloud electrification process has great significance in understanding the microphysical properties, electrical characteristics, and evolution of thunderstorms. This study employs an X‐band dual‐polarized multiparameter phased array weather radar (MP‐PAWR) to observe the electrical alignment signatures of ice particles before the first intracloud (IC) lightning flash, and to explore the evolution of the upper charge region in the early electrification stage of an isolated thunderstorm. Negative KDP signatures associated with vertically oriented ice particles by strong electric fields in the upper parts of the thunderstorm are analyzed by introducing composite KDP, which is defined as a minimum KDP value observed in a vertical column across all elevation scans at each specific horizontal grid point at and above a designated layer. About 7 min before the first IC lightning flash, the mean canting angle of ice particles in the upper parts of the cloud changed from horizontal to vertical by strong electric fields, and the concentration of vertically aligned ice particles on the top of the cloud reached the maximum 30 s before the first IC lightning flash. These signatures exhibit an early electrification process in the upper parts of the thunderstorm. These results indicate that with the high spatial and temporal resolution, MP‐PAWRs have the ability not only to detect the rapid evolution of microphysical structures but also to observe the early electrification of thunderstorms, which will facilitate forecasting IC lightning flash initiation combined with graupel presence signatures in the mixed‐phase region in normal operation.


Introduction
The cloud electrification process plays an important role in understanding microphysical, electrical properties and evolution of thunderstorms and lightning activity.According to the non-inductive charging theory, which is the most viable explanation for the initial electrification of thunderstorms, it is widely accepted that the transfer of electrical charge occurs during collisions between graupel and smaller ice crystals in the presence of supercooled water within the storm's mixed-phase region which typically ranges from about 0°C to 40°C (Takahashi, 1978;Takahashi & Miyawaki, 2002;Williams, 1989).The negatively charged graupel accounts for the main negative charge region distributes at approximately constant temperature levels between 10°C and 20°C, while positively charged ice particles move upward by updraft to form the upper positive charge region (Dye et al., 1986(Dye et al., , 1988;;Krehbiel, 1986;Saunders, 1993;Saunders et al., 2006;Workman & Reynolds, 1949).Numerous previous studies have explored the relationship between precipitation development and electrical characteristics in thunderstorms by weather radar (Bringi et al., 1997;Buechler & Goodman, 1990;Carey & Rutledge, 2000;Dye et al., 1989;Gremillion & Orville, 1999;Larsen & Stansbury, 1974;Lund et al., 2009;Williams et al., 1989).They utilized the reflectivity at different temperature levels in the mixed-phase region and the kinematic structure of clouds to investigate the initiation of thunderstorms, as well as the onset of electrification in thunderstorms and forecasting of lightning activity.Typically, in these studies, the formation of the main negative charge region within storms, characterized by the presence of graupel during the updraft, was considered a reliable indicator of initial electrification.In contrast, the features associated with the upper positive charge region which consists of charged ice particles, despite being an integral part of thunderstorm electrification, gained less attention in studying the initial electrification.
As the two primary charge separation theories proposed to explain electrification in the cloud, both the inductive and non-inductive charging mechanisms normally cause ice crystals in a thunderstorm to become positively charged and graupel to become negatively charged (Aufdermaur & Johnson, 1972;Brooks & Saunders, 1994;Reynolds et al., 1957;Saunders, 2008;Takahashi, 1978;Takahashi & Miyawaki, 2002).Due to the size differences, the charged ice crystals are carried aloft to accumulate the upper positive charge region, while the graupel tends to fall or be suspended by the updraft to form the main negative charge region below.The small ice crystals on the top of the cloud, such as plate, column, and needle shapes, could be oriented vertically with their principal axis by the strong electric fields.This sign was first observed as an unusual optical phenomenon in 1950, the vertically aligned ice crystals reflect the light from their horizontal surfaces like small mirrors.Vonnegut (1965) mentioned this phenomenon could have resulted from changes in the alignment of ice crystals due to the strong electric field within thunderstorms.An optical polarization measurement was made by Mendez (1969), and the results showed that "the electric field contained within the thunderstorm prior to and after a lightning discharge is imaged by the local degree of polarization produced by the sunlight reflected from oriented ice crystals on the surface of the storm." The changes of differential propagation and backscatter characteristics that arise from electrical vertically aligned ice crystals on the top of the cloud provide more polarized information to study the electrification and lightning activities in thunderstorms by weather radar observation.The earliest observations of the polarized parameter signatures related to such electrical alignment of ice particles were mainly performed with circularly polarized weather radars.Abrupt changes in the polarization parameters associated with lightning flashes were first observed in 1974 (Hendry & McCormick, 1976), and Hendry and McCormick (Hendry & McCormick, 1976;McCormick & Hendry, 1979) showed the existence of vertically canted ice particles in the upper parts of thunderstorms before lightning discharges by a Ku-band circular-polarization radar.They concluded that the abrupt changes in orientation and circular depolarization ratio were induced by the lightning discharges.Further research by Hendry (Hendry & Antar, 1982) showed that the depolarization effects in some cases could probably be caused by the propagation effects of radiowaves in the particles which are too small to be detected by radar.Similar observations were obtained by using S-band (Metcalf, 1992(Metcalf, , 1993(Metcalf, , 1995) ) and X-band (Krehbiel & Gray, 1993;Krehbiel et al., 1992Krehbiel et al., , 1996) ) circular polarimetric radar, respectively.These studies showed that the changes in propagation effects were associated with rapid changes in orientations of hydrometeors in the middle and upper parts of the cloud caused by electric fields.The mean canting angle of the hydrometeors is near 90°b efore lightning and subsequently recovers to the horizontal orientation following a lightning discharge.
The depolarization effects related to the electrical alignment of ice particles have also been observed by duallinear polarized weather radars (Caylor & Chandrasekar, 1996;Hubbert et al., 2014;i Ventura et al., 2013;Mattos et al., 2016Mattos et al., , 2017;;Krehbiel et al., 2005;Ryzhkov & Zrnić, 2007;Schvartzman et al., 2022;Scott et al., 2001;Zrnic & Ryzhkov, 1999).By observing with a dual-wavelength (S-band and X-band) dual-linear polarized weather radar, the first measurement of the specific differential phase (K DP ) related to the orientation of ice particles with in-cloud electric fields has been taken by Caylor and Chandrasekar (1996).The study implied that K DP and linear depolarization ratio (LDR) can be used to investigate the orientation of ice crystals in the upper parts of thunderstorms by in-cloud electric fields and roughly concluded the relation of LDR and K DP to ice crystal orientation angle.The negative K DP corresponds to the vertically aligned ice particles due to strong electric field, while positive K DP associates with horizontal alignment ice particles.The study showed that no orientation-related signatures were observed in the copolar correlation coefficient (ρ HV ) and differential reflectivity (Z DR ) of the S-band.It indicates that the polarimetric signatures are caused primarily by propagation effects.The period changes in K DP can be attributed to variations in the alignment of ice crystals, which tend to align parallel to the in-cloud electric fields.K DP decreases gradually on the top of the cloud as the electric field accumulates, then suddenly increases coincident with lightning discharges.Besides the negative K DP signatures associated with vertically aligned ice particles in the in-cloud electric fields, the Z DR streaks (radial signatures of Z DR that Z DR alternate between rather high values and rather low values: Ryzhkov & Zrnić, 2007) related to the depolarization effects of electrical canting ice particles (i Ventura et al., 2013;Hubbert et al., 2014;Schvartzman et al., 2022) were observed on the top of the thundercloud.
With the wide spread of dual-polarization weather radars, the signatures related to ice crystal orientation by the electric field have been well established.Based on the previous studies, the depolarization effects related to the electrical alignment of ice particles have been largely dominated by the propagation effects.Moreover, for the simultaneously transmitting and receiving horizontally and vertically polarized waves (SHV mode) scheme weather radar, which is the most widely used mode of dual-polarization weather radar, the measurements of Z DR , ρ HV , and the differential propagation phase shift (Φ DP ) are improved by eliminating inter-pulse variations and the Doppler effects (Scott et al., 2001).By using an X-band dual-polarization weather radar that transmitted circularly polarized waves and received simultaneously in horizontal and vertical polarization, Scott et al. (2001) showed that the electrical alignment directions of ice particles can be determined by changes in the polarization state between consecutive range bins caused by Φ DP of vertically aligned ice particles.However, weather radar in SHV mode would introduce measurement biases due to the cross-coupling between the horizontal and vertical waves, especially in the measurement of Z DR (Hubbert et al., 2010;Ryzhkov & Zrnić, 2007;Wang & Chandrasekar, 2006).Ryzhkov and Zrnić (2007) implied that the Z DR signature depends both on the canting angle of particles and system differential phase shift on transmission and propagation differential phase shift, while the K DP of aligned ice crystals is mainly dominated by the magnitude of the canting angle.Carey et al. (2009) explored the characteristics of K DP and Z DR related to ice particle orientation by modeling ice particles with the particle size distribution, mean canting angle, and different types of ice hydrometeor types.They concluded that electric fields and lightning potential are physically connected to the K DP alignment signature.K DP exhibits a linear relationship with particle concentration and a non-linear relationship with particle size.K DP is a good indicator of electrical alignment signatures compared to Z DR .Z DR of vertical alignment ice particles could be masked by larger horizontal aggregates while K DP is not.The K DP seems to be a more robust parameter in SHV mode weather radar to study the signatures related to depolarization effects associated with the electrical alignment of ice particles.
Although dual-polarization radar can observe the depolarization effects of electrically aligned ice crystals in strong in-cloud electric fields, it is still difficult to observe the electrical alignment signatures in the whole electrification process of thunderstorms, and to observe the evolution of electric fields related to the upper charge region in the usual operational application.The first reason is due to the fast motion and rapid development of thunderstorms and the lightning discharges being intermittent and short duration.The observations by traditional polarimetric weather radars with mechanically scanned parabolic antenna are limited by volume scanning time.
To study the depolarization effects related to the electrical alignment of ice crystals, weather radars need to point the antenna at fixed azimuth and either at a fixed elevation angle or within a small sector of elevation toward to suspected thunderstorm, then update the antenna position to track the electrically active cell.Most previous studies on electrical alignment signatures normally pointed the antenna in a fixed direction, or scanned on limited adjacent azimuth or elevation angles.The advantage of fixed location observation is that it can obtain sufficient dwell time to estimate polarized parameters and detect related signatures associated with each lightning activity.However, it is challenging to observe the electrical alignment signatures related to the upper charge region by traditional polarimetric weather radars within the evolution of thunderstorms in normal operation.On the other hand, the electrically active region of the thunderstorm was normally selected by taking the reflectivity of precipitation echoes for most previous researches, while the selected electrically active region was typically related to the main negative charge region.Or using other data as auxiliary indicators of an electrically active alignment region, such as the data of satellite communication links (Antar et al., 1980;Hendry & Antar, 1982), the lightning data (Metcalf, 1995) or data with another weather radar (Caylor & Chandrasekar, 1996).An observation from Krehbiel et al. (1984) showed a precise correlation between the first 15 negative IC lightning discharges of a localized storm and the updraft in the core of the storm, while a weaker radar echo extended up to and above the upper charge centers.This study implied that, during the early stages of electrification, the core of precipitation echo normally coincides with the center of the main negative charge region composited by graupel (Krehbiel, 1986;Krehbiel et al., 1984).It is worth noting that, in this study, a fast-scanning X-band vertical polarized radar using broad-band noise transmissions and a rapidly spinning antenna that can perform full volumetric hemispherical scans in 15-20 s was used to enable detailed observations of the rapid evolution processes in thunderstorms (Krehbiel & Brook, 1979).For well-developed thunderstorms, the discharge centers and paths are frequently (although not always) along the boundaries of high-reflectivity regions while tending to avoid the core (Dye et al., 1986;Proctor, 1983).These probably indicate that during the evolution of thunderstorms, selecting a region only by the reflectivity-based method to study the electrical alignment signatures may not fully capture the features in changing electric fields related to the upper charge region.
Phased array weather radars have gained significant attention in meteorological applications with the advantage of high temporal and spatial resolution, especially in short-term severe convective weather system monitoring and early warning, such as tornadoes, microbursts, and thunderstorms (Adachi et al., 2016;Heinselman et al., 2015;Palmer et al., 2022;Ushio et al., 2015;Yoshida et al., 2017).Phased array weather radars will help better understand the evolution and electrification development of thunderstorms by providing more specific information on cloud microphysical evolution processes.In this paper, the dual-polarized multiparameter phased-array weather radar (MP-PAWR), located at Saitama University, is used to observe an isolated thunderstorm and analyze polarimetric parameters signatures associated with electrically aligned ice particles.The MP-PAWR is an X-band dual-linear polarized multiparameter phased array weather radar developed by Toshiba Corporation, Osaka University, and the National Institute of Communication and Technology (NICT).It can perform a volume scan of the whole sky in 30 s with a maximum detection range of 60 km, and a range resolution of 75m.The high spatial and temporal resolution of the MP-PAWR provides an effective way to observe the evolution and electrification life cycle of thunderstorms in normal operational applications.By introducing the composite K DP , the electrical alignment signatures of ice particles before the first intracloud (IC) lightning flash are analyzed.The result allows us to explore the initial electrification process and vertical structure of the upper charge region in thunderstorms by the MP-PAWR.Section 2 is the introduction of data used in this paper and the methodology of data preprocessing and analysis.The observation and analysis results are presented in Section 3. A summary and discussions are shown in Section 4.

Lightning Data
Lightning data in this paper were obtained by the Lightning Detection Network (LIDEN) of the Japan Meteorological Agency (JMA).It consists of detection stations installed in 30 airports across the country receiving electromagnetic waves generated by lightning discharges, and of a central processing station that collects data from the detection station and determines the location of lightning discharges.The LIDEN system provides lightning data including types (intracloud or cloud-to-ground lightning), locations, and occurrence times of lightning discharges.We mainly employ IC flashes to focus on electrical alignment signatures of ice particles before the first IC lightning flash in the upper portion of the cloud.

Dual-Polarized Multiparameter Phase Array Weather Radar
The MP-PAWR is a dual-linear polarized multiparameter phased array radar operated at X band with SHV mode, which provides parameters such as horizontal and vertical radar reflectivity (Z H , Z V ), doppler velocity (V H ,V V ), doppler velocity spectrum width (W H ,W V ), and also polarization parameters such as differential reflectivity (Z DR ), differential propagation phase shift (Φ DP ), specific differential phase (K DP ), and correlation coefficient (ρ HV ).In a normal operational mode, the temporal resolution of a volume scan with 114 elevation angles is 30 s.
The MP-PAWR used in this paper is located at Saitama University in the metropolitan area of Japan (35.86158°N, 139.60908°E) at an altitude of 29.3 m above sea level.The capabilities and overall system description of the MP-PAWR were given by Takahashi et al. (2019) and Kikuchi et al. (2020).Also, comparison with a parabolic-type antenna X-band multi-parameter (X-MP) radar was performed by Asai et al. (2021).

Φ DP Filtering and K DP Estimation
To improve analysis accuracy, it is necessary to perform quality control of MP-PAWR polarization parameters first.In this paper, the quality control process is mainly applied to the differential propagation phase (Φ DP ), specific differential phase (K DP ), and horizontal reflectivity (Z H ).
There are many methods to filter Φ DP from noisy total differential phase shift by removing backscatter differential phase shift (δ) and smoothing fluctuation of Φ DP .Hubbert et al. (1993) proposed to use the method of Finite Impulse Response (FIR) and Infinite Impulse Response (IIR) low-pass filter to process total differential phase shift, and Hubbert and Bringi (1995) proposed an iterative filtering technique based on this method, which can effectively eliminate backscatter differential phase shift (δ) existing in total differential phase shift.Giangrande et al. (2013) used the linear programming theory to eliminate the noise and ensure the processed total differential phase shift meets the condition of monotonically increasing with radial distance under the melting layer.K DP could be estimated accurately from processed Φ DP .Matrosov et al. (2006) used a linear regression method to derive K DP from the slope of Φ DP .In this paper, Φ DP data provided by MP-PAWR are unwrapped already, we use the algorithm based on Hubbert and Bringi (1995) to filter Φ DP and estimate K DP by using the Colorado State University (CSU) RadarTools (Lang et al., 2007).
The algorithm of CSURadarTools to process raw Φ DP calculate the standard deviation of Φ DP at first, and only works when window length divided by gate spacing equals an even number (Lang et al., 2007).Considering that the depolarization effects of electrically aligned ice particles mainly depend on the differential propagation phase effects, the differential phase shift is accumulated along the radar radial.For weather radar with short wavelength, the backscatter differential phase δ is non-negligible for rain, whereas it can exceed 20°at X band for very large drops (Trömel et al., 2013).To better eliminate backscatter differential phase shift (δ) from the raw Φ DP and obtain a reliable estimation of K DP from smoothed Φ DP , we employ an FIR range filter with a window length of 4.5 km and a standard deviation of differential phase σ(Φ DP ) of 30°.This filtering process may result in the loss of the filtered Φ DP in the several beginning and ending range bins along the radial paths.However, in the region near the radar, Φ DP tends to be affected by the ground clutter, and Φ DP in far-range bins tends to fluctuate erratically because of the low signal-to-noise ratio.The missed data will not affect the analysis from the observation of electrical alignment signatures.Finally, K DP is estimated by the least squares method from the filtered Φ DP , and calculated over a window whose length is inversely proportional to the Z H values.
Figure 1 shows the radial profiles of rainfall rate, raw Φ DP , raw K DP , filtered Φ DP , and estimated K DP .The observation data was obtained by the MP-PAWR, at an azimuth angle of 330°and elevation angle of 4.99°, in a rain region below the melting layer at 16:00:24 JST on 20 August 2019.After quality-control processing, the filtered Φ DP monotonically increases and the estimated K DP keeps positive in the rain region below the melting layer.The estimated K DP reaches a maximum of ∼2.5°/km when the filtered Φ DP grows rapidly between 20 and 30 km in a heavy rain region, while the peak rainfall rate reaches about 50 mm/hr.The altitude of the melting layer is about 5.5 km, based on sounding data at the Tateno site on 20 August 2019.marked as black ovals in Figure 2, and K DP in the same region are negative values as Figure 2d shows.The minimum K DP at the 10 km height is less than 1°/km.Because small ice crystals tend to be more easily drifted up to the upper parts of the cloud top, and the K DP of smaller, vertically aligned ice crystals will dominate over the K DP of larger and horizontally aligned aggregates (Carey et al., 2009).These features should be attributed to the vertical alignment of ice crystals in strong electric fields.
Figure 3 shows constant-altitude plan-position indicator (CAPPI) images of horizontal reflectivity Z H after attenuation correction and the estimated K DP at an altitude of 10 km.The vertical cross-section results along the azimuth angle of 248.4°(black line in Figure 3) are shown in Figure 2. The IC lightning activities, as the black points marked, were detected in this region by the LIDEN system.The maximum Z H is above 50 dBZ.The negative K DP region in Figure 3b is associated with the lightning activities.The negative values of K DP are largely due to the vertical alignment of ice crystals by electric fields.As Figures 2 and 3 show, the area of estimated K DP is smaller than the area of raw K DP and reflectivity, it is mainly caused by the process of filtering the raw Φ DP as mentioned in above section.In the following part of this paper, if there is no special explanation, Φ DP and K DP are the filtered ones by the quality control process.

Attenuation Correction
X-band weather radars experience larger attenuation in rain than S/C-band radars, thus rain attenuation correction is needed, especially in convective rain events.The attenuation-correction procedures on MP-PAWR were introduced by using a method with K DP (Asai et al., 2021), and the correction results showed that the MP-PAWR provides highly accurate precipitation observations developed in a short time, comparing with X-band multiparameter (X-MP) radar in Japan.In this paper, the rain attenuation is corrected by using the ZPHI method (Bringi et al., 2001;Testud et al., 2000).The measured horizontal reflectivity Z′ H (mm 6 /m 3 ) at a range r are related to corrected horizontal reflectivity Z H (mm 6 /m 3 ) as Equation 1, and the specific attenuation A H (dB/km) is calculated according to Equation 2. A H (r) = [Z′ H (r)] b 10 0.1bα∆Φ DP 1) I(r 1 ,r 0 ) + 10 0.1bα∆Φ DP 1) I(r,r 0 ) ( 2) ∆Φ DP = Φ DP (r 0 ) Φ DP (r 1 ) , where r 1 and r 0 are represent the starting and ending range of a rain cell (r 1 < r<r 0 ).The coefficients α, and b of the A H -Z H relation are 0.254 and 1.143 respectively, which were used in attenuation correction with the X-MP radar (Park, Bringi, et al., 2005;Park, Maki, et al., 2005) and MP-PAWR (Asai et al., 2021).
Plan-position indicator (PPI) scan results of attenuation-corrected Z H at 16:00:24 JST, 20 August 2019, are shown in Figure 4.The PPI scan was conducted at an elevation angle of 4.99°.Figure 5 shows the radial profiles of measured Z H and attenuation-corrected Z H at an azimuth angle of 330°(marked as the black line in Figure 4) and an elevation angle of 4.99°.The corrected Z H closely matches the measured Z H within a range of 20 km from the radar.Between the 20-km and 30-km range away from the radar, the difference between measured Z H and corrected Z H increases.At distances greater than 30 km, the corrected Z H is more than 10 dB higher than the raw Z H .It means that Z H experiences significant attenuation between the 20-km and 30-km range, where the measured Z H is larger than 40 dBZ.The trend of Z H is consistent with the increasing Φ DP in this region as Figure 1 shows, the overall reflectivity is recovered in the further region after the attenuation correction.

Analysis Method for Electrical Alignment Signatures of Ice Particles
Small ice particles tend to be more easily uplifted to the top parts of the cloud to form the upper positive charge region and vertically aligned by strong electric fields.Conversely, graupel, being typically larger and denser, is less likely to become vertically aligned.Moreover, the K DP of larger horizontally oriented aggregates does not mask the K DP of small vertically aligned ice crystals (Carey et al., 2009;Hubbert et al., 2014).These characteristics make the K DP a good indicator of weather radar in SHV mode to observe the electrical alignment signatures of ice particles by strong electric fields in thunderstorms.
To explore the characteristics of the negative K DP region associated with the upper charge region, including its distribution on azimuth and height, as well as the concentration of vertically aligned ice crystals, composite K DP is introduced in this paper.It is defined as the minimum K DP value in a vertical column of all elevation scans at each horizontal two-dimension grid point, similar to the definition of composite reflectivity (Witt et al., 1998).
Considering that the main negative charge region distributes at approximately constant temperature levels between 10°C and 20°C (Dye et al., 1986(Dye et al., , 1988;;Krehbiel, 1986;Saunders, 1993;Saunders et al., 2006;Workman & Reynolds, 1949), we calculate the composite K DP above the 10°C layer, representing the minimum K DP in the vertical column at and above the 10°C layer, to investigate the features of upper positive charge region in early electrification stage.
Given that the negative K DP values have a linear relationship with the concentration of vertically aligned ice particles by the strong electric field (Carey et al., 2009), it is evident that a higher concentration of electrically aligned ice particles corresponds to a stronger electric field.Moreover, the gross electric fields structure in thunderstorms undergoes slower changes compared to the changes in local electric field magnitudes causing lightning initiation.By calculating the average composite K DP above different temperature layers, we can utilize the mean values of minimum K DP within different regions in the upper parts of the isolated thunderstorm, specifically at and above different ambient temperature levels starting from the 10°C layer, to evaluate the vertical structure of electrically aligned ice particles with relatively high concentration during the electrification process in the thunderstorm.Consequently, a rough qualitative analysis of the vertical electrical structure of the dense upper charge region can be conducted based on these results.Additionally, as a comparative analysis, the average K DP in a horizontal cross-section at different isothermal layers spanning from the 10°C layer to the cloud top could be calculated to analyze the vertical distribution of the upper charge region.
The reflectivity-based method is used to detect severe convection systems which could have the potential to produce lightning activities and thereby further be applied in lightning detection and warning by weather radars (Buechler & Goodman, 1990;Lund et al., 2009;Preston & Fuelberg, 2015;Woodard et al., 2011Woodard et al., , 2012)).The moderate reflectivity value (e.g., 30dBZ, 40dBZ) above the 0°C isothermal level is normally associated with the presence of graupel, which plays an important role in thunderstorm electrification.Buechler and Goodman (1990)    Fuelberg (2015) used the presence of graupel at different temperature levels (i.e., 0°C, 10°C and 20°C) in the mixed-phase region as a proxy for storm electrification, and showed that almost every flash initiation source was collocated with the presence of graupel at 10°C.They utilized the presence of graupel and horizontal reflectivity larger than 35 dBZ at the 10°C isothermal layer as threshold in lightning cessation algorithm.
In this study, a reflectivity criterion has been employed to identify the area with potential lightning activity.
Considering that composite reflectivity can reveal important storm structure features and intensity trends of storms compared with base reflectivity, we use composite Z H above the 10°C layer, where the reflectivity is larger than 40 dBZ at and above the 10°C isothermal layer, to define a two-dimension boundary of the analysis region in the isolated thunderstorm.Furthermore, to ensure data reliability, we screen out radar data with a correlation coefficient ρ HV smaller than 0.8 in the selected region.By using composite K DP , the electrical alignment signatures of ice particles in the upper parts of thunderstorms before the first IC lightning flash are analyzed.Based on sounding data at the Tateno site on 20 August 2019, the altitude of the 10°C, 20°C, 25°C and 30°C layers are about 7 km, 8.5 km, 9 km, and 10 km, respectively.

Initial Electrification in the Early Developing Stage of an Isolated Thunderstorm
On 20 August 2019, there was a mesoscale convective system (MCS) moved from northwest to southeast of the MP-PAWR.To investigate the electrical alignment signatures of ice crystals before the first IC lightning flash, an isolated thunderstorm moving from the west of MP-PAWR was selected for analysis.The first IC flash occurred at 15:40:03 JST, at an azimuth angle of 263.2°and a distance of 42.6 km from the radar.CAPPI of Z H and K DP at the 10°C layer are shown in Figure 6.
Figure 6 illustrates the early developing stage of the isolated thunderstorm.After 15:28:08 JST, the analyzed isolated thunderstorm echo appeared at the 10°C layer with a maximum Z H larger than 40 dBZ.Due to the Φ DP filtering process mentioned before, the top height of K DP at the same time remained below the 10°C layer.After 15:32:06, the isolated thunderstorm elevated and expanded its area at the 10°C layer, and the corresponding K DP raised above the 10°C layer as shown by a black circle in Figure 6h.As time goes on, the area of the thunderstorm continued to increase.Considering that the purpose of this paper is to analyze the negative K DP signatures related to vertical alignment ice particles before the first IC lightning flash, K DP results within the wellorganized isolated thunderstorm after 15:32:06 are used to analyze.
Figure 7 shows the evolution of the developing stage of the isolated thunderstorm.The time resolution is 30 s, and the altitude resolution is 500 m due to computational limitations.After 15:59:25, the weakened and disappeared cloud echo was caused by the movement of the thunderstorm center toward to southeast.
Figures 7a and 7b exhibit the development of the isolated thunderstorm at a range of 40 km from the radar with a 260.4°azimuth, which is roughly along the azimuth with the most negative K DP values at the layers ranging from the 10°C layer to the cloud top at 15:40:03 JST.As Figure 7a shows, it is evident that there was an updraft during the developing stage.The cloud grew upward and eventually reached a height of about 14.5 km.Especially after 15:34:35 JST (left black line in Figure 7a), the height of the echo with 40 dBZ reflectivity raised above the 0°C level and continued to move toward the 10°C layer, with continued elevated to exceed the 20°C layer when the first IC lightning flash occurred.It is obvious that negative K DP appeared on the top of the cloud after 15:34:35 JST (left black line in Figure 7b) as shown in Figure 7b.
Figures 7c and 7d show the time evolution at a range of 42.6 km from the radar with a 262.8°azimuth, which represented the closest location to the occurrence of the first IC flash.There was a strong updraft with reflectivity larger than 40 dBZ above the 0°C layer.Moreover, 30 s before the first IC lightning flash, the maximum reflectivity exceeded 50 dBZ between the 0°C layer and the 10°C layer.From 15:38:32 JST to 15:40:03 JST, approximately 1.5 min before the first IC lightning flash, the K DP values are negative between the 0°C layer and the 10°C layer, as shown by the black box in Figure 7d.Specifically, at 15:40:03 JST when the first IC flash occurred, the enhanced negative K DP values that were less than 1°/km presented in the 0°C layer.
The vertical cross-sections of the isolated thunderstorm from 15:35:05 to 15:40:03 JST at a range of 42.6 km from the radar at azimuth from 240°to 270°are shown in Figure 8.It presents a detailed development in 6 min of the isolated thunderstorm near the location where the first IC lightning discharged before it occurred.Figures 8a-8f show the evolution of Z H over a period of 6 min with a time interval of 1 min.The strong echo core grew upward and the altitude of 40 dBZ at the top of the echo core raised from about 7 to 11 km in 6 min, and the height of the cloud top raised from about 10 to 13 km in 6 min.Especially after 15:37:04 JST (Figure 8c), the region where reflectivity was larger than 40 dBZ extended above the 20°C layer and the area was getting wider.Figures 8g-8l show the evolution of K DP at 1-min intervals.The positive K DP core region where the maximum K DP was larger than 4°/km corresponded with the strong echo core below the 0°C layer.Figures 8m-8r show the ρ HV values which below the melt layer were around 0.925 while the ρ HV values above the 10°C level were around 0.975.Under the melting layer, the positive K DP region was associated with large reflectivity, and the ρ HV values were around 0.925 means that the heavy rain probably was mixed with bigger drops or hailstones in this region.The positive K DP region above the melt layer and below the negative K DP region indicates that there were probably graupel, aggregates, and ice particles mixed.The positive values of K DP may be largely related to the graupel because graupel is normally too large to be vertically aligned by strong electric fields and has the potential to mask ice crystal vertical alignment signatures in K DP associated with strong electric fields (Carey et al., 2009).The negative K DP region in the upper portions of the storm corresponded to the region where maximum reflectivity was about less than 40 dBZ and the ρ HV values were larger than 0.95.This region maybe indicates the mixtures of aggregates and ice particles.Due to the K DP of small vertically oriented ice particles dominate over larger horizontal alignment aggregates (Carey et al., 2009), the negative K DP region was most likely related to the vertical orientation of ice particles due to strong electric fields.At 15:39:03 JST, 30 s before the first IC flash, marked as a black circle in Figure 8k, there is a region with negative K DP less than 1°/km at 9 km height.It means more primarily vertically aligned ice particles in the enhanced electric fields on the cloud top.
Z DR column has been used as an indicator of strong updraft and electrification.The Z DR column is defined as a narrow vertical extension of positive Z DR values above the 0°C isotherm level associated with the updrafts in deep moist convective storms (Kumjian et al., 2012(Kumjian et al., , 2014)).It is indicative of supercooled liquid drops lifted by the updraft.The Z DR columns are likely sources of graupel embryos, which may form from the freezing of drops in these columns.The strong updrafts and the formation of graupel are favorable conditions for the electrification of thunderstorms and lightning flash initiation (Conway & Zrnić, 1993;Goodman et al., 1988;Kumjian et al., 2012Kumjian et al., , 2014;;Lund et al., 2009;Woodard et al., 2011Woodard et al., , 2012)).
In the developing stage of the isolated thunderstorm, some Z DR columns were observed.Taking one of them as an example which is marked as black ellipses in Figure 9, the reflectivity values (Z H > 35 dBZ) associated with enhanced values of Z DR (Z DR > 3 dB), extended upward above the ambient 0°C layer.The signatures of the Z DR column suggested the early electrification in the thunderstorm.

Signature of Composite K DP Before the First IC Lightning Flash
As Section 2 mentions, the composite K DP above the 10°C layer is used to analyze the characteristics of negative K DP associated with the vertical alignment signatures of ice particles before the first IC lightning flash in the upper parts of the isolated thunderstorm.Meanwhile, in the selected analysis region of the isolated thunderstorm, the mean values of composite K DP are calculated above different isothermal layers ranging from 10°C to 30°C layer, which explore the concentration of vertically aligned ice particles in different upper volumes in the thunderstorm.
Figure 10 shows the horizontal cross-section results at the 10°C layer of composite Z H , composite K DP , and the CAPPI results of K DP at the 10°C layer.The first IC flash occurred at 15:40:03 JST, indicated by the marked points in Figures 10e, 10k, and 10q, representing the azimuth locations of lightning discharges.
As Figures 10a-10f show, the Z H of the isolated thunderstorm increases above the 10°C layer, with the maximum Z H being larger than 50 dBZ after 15:34:05.Compared to the CAPPI results of K DP shown in Figure 10g-l, the results of composite K DP show a larger negative K DP area and enhanced negative K DP values at the 10°C layer as Figures 10m-10r show.It indicates that a great number of regions with higher concentrations of vertically aligned ice particles were distributed above the 10°C layer in the cloud.The larger negative K DP region with more negative K DP values, where the minimum K DP was less than 1°/km as shown in Figure 10q, appeared near the location of the first IC lightning discharge.Conversely, no such obvious negative K DP signatures are shown in Figure 10k.It implies that the use of composite K DP provides a clearer representation of the negative K DP signatures associated with the vertical alignment of ice particles in the upper parts of the cloud.Furthermore, it could indicate changes in azimuthal location and strength of the upper charge region in the thundercloud, which would facilitate observing the variations of electric field strength on vertical structures in thunderstorms.
As Figure 11a shows, the first IC lightning flash occurred at 15:40:03 JST (as marked by the right vertical black line).The mean composite Z H which is larger than 40 dBZ reached and above the 10°C layer at 15:28:37 JST, then gradually increased to about 46 dBZ.
In Figure 11b, as mentioned before, the K DP values after 15:33:06 (left black line in Figure 11b) are used for the analysis, and the composite K DP values above the 40°C layer are not analyzed in this paper because that the top height of K DP is below the 40°C layer before the first IC lightning flash.At 7 min before the first lightning flash, it is obvious that composite K DP above the 10°C layer gradually decreased from approximately 0.3 to 0.2°/km in 3.5 min (from 15:33:06 to 15:36:34 JST), with the average composite K DP above the 20°C layer decreasing, at 6 min before the first IC lightning flash, from about 0.5 to 0.65°/km in 5 min (from 15:34:05 to 15:39:03 JST).A similar signature also appeared above the 25°C layer, the average composite K DP decreased from 0 to 0.4°/km between 15:35:05 to 15:35:36 JST, 5 min before the first IC lightning flash, followed with a decrease from about 0 to 0.7°/km in 1 min (from 15:37:34 to 15:38:33 JST) which is 2.5 min before first IC lightning flash.The average composite K DP above the 30°C layer decreased at 2.5 min before the first IC lightning flash from about 0.4 to 0.8°/km in 30 s (from 15:37:34 to 15:38:03 JST) and continued to decrease from 0.65 to 1°/km in 1 min (from 15:38:33 to 15:39:33 JST).Notably, the composite K DP above the 30°C layer reached the most negative value 30 s before the first IC lightning flash.The result shows that, before the first IC lightning flash, the mean values of composite K DP , across different layers from the 10°C layer to the cloud top, gradually decreased, while the mean values of composite K DP at a higher altitude were more negative than the values at a lower altitude.As the time approached the occurrence of the first IC lightning flash, the negative value of average composite K DP appeared at higher altitudes and decreased to a more negative value compared to lower altitudes.For example, the average composite K DP above the 10°C layer decreased from 0.3 to 0.2°/km 7 min before the first IC lightning flash, while the average composite K DP above the 20°C layer decreased from 0.5 to 0.65°/km 6 min before the first IC lightning flash.Similarly, the average composite K DP above the 25°C layer and the 30°C layer decreased from 0 to 0.7°/km 5 min before the first IC lightning flash, and from 0.4 to 1°/km 2.5 min before the first IC lightning flash, respectively.These signatures probably indicate an evolution of the upper charge region associated with electrical-related ice particles in the initial electrification process during the early developing stage of a thunderstorm.The mean orientation of ice particles on the top of the cloud changed from horizontal to vertical alignment about 7 min before the first IC lightning flash.This change probably indicated a variation of the electric field, which was associated with the buildup of an upper charge region composed of charged ice particles.As the charged ice particles were carried upward by the updraft, the upper charge region raised from the 10°C layer, began approximately 7 min before the first IC lightning flash, and reached above the 30°C layer in 4.5 min.This accumulation at higher altitudes resulted in a gradual increase in the intensity of the electric field in the upper parts of the cloud that, in turn, would enhance the concentration of vertically aligned ice particles in the upper parts of the cloud.Correspondingly, the average composite K DP exhibited a decrease before the first IC lightning flash.
The evolution of the upper charge region in the early electrification stage of the isolated thunderstorm could be qualitatively analyzed by the average composite K DP at different isothermal layers.In this paper, the average composite K DP above the 25°C layer began to decrease about 2 min after the decrease in average composite K DP above the 10°C and the 20°C layers, which probably means the initial upper charge region built up under the 20°C layer (including the 20°C layer) and subsequently raised above the 25°C layer 2 min later.The average composite K DP above the layers across the 10°C and the 30°C layers increased following with the first IC lightning discharge.

Signatures of Average K DP Before the First IC Lightning Flash
Figure 12a displays a time-height analysis of the average K DP values of the isolated thunderstorm, covering altitudes ranging from ground level to 15 km with a resolution of 500 m.The analysis region was selected using composite Z H , where values exceeding 40 dBZ were observed above the 10°C layer.Notably, approximately 30 s before the first IC lightning flash, the average K DP at the 30°C layer dropped below 1°/km, indicating an intensive concentration of vertically aligned ice particles accumulating in the top parts of the cloud resulting from stronger electric fields as the lightning event approached.
Figure 12b illustrates the evolution of the average K DP at various altitudes.The occurrence of average K DP at higher altitudes was delayed due to upward air motion.Before the first lightning flash at 15:40:03, the average K DP values at and above the 20°C layers were predominantly negative, suggesting the presence of strong electric fields primarily at and above the 20°C layer.
Before the first IC lightning flash, average K DP at altitudes between the 10°C and 30°C layer exhibited varying degrees of decrease.It noticed that the average K DP values at the 10°C layer were almost positive, which is consistent with the conclusion from Mattos et al. (2016), who observed that storms with lightning showed strong positive K DP in the layer from the 0°C and the 15°C layer.It is probably because graupel has the potential to mask negative K DP signature related to vertically aligned ice particles by strong electric fields in the graupel-ice mixed region.The average K DP at the 20°C layer was kept negative before the first IC lightning flash.It probably indicates that, during the early electrification stage, the upper charge region that was composed of ice particles accumulated around the 20°C layer, or/and uncharged ice crystals vertically aligned by the electric fields between the upper charge region and the main negative charge region.At the 25°C layer, the average K DP decreased from 0.1°/km to 0.6°/km in 1 min (from 15:37:34 to 15:38:34), while at the 30°C layer, the average K DP decreased from 0.4°/km to 1.0°/km 3 min (from 15:37:34 to 15:38:03) before the first lightning flash, further decreasing to 1.2°/km at 15:39:34 JST, 30 s before the first IC lightning flash.The average K DP at the 20°C, 25°C, and 30°C layers increased with the first IC lightning discharge, indicating a weakening of the electric fields.These observations indicate that the average K DP values became more negative at higher altitudes as the first IC lightning flash approached, highlighting the enhanced negative K DP signatures of vertically oriented ice particles associated with stronger electric fields, similar to the composite K DP signatures observed before the first IC lightning flash.

Discussion and Conclusions
In this paper, an isolated thunderstorm is selected to analyze the electrical alignment signatures of ice particles in the upper parts of the thunderstorm before the first IC lightning flash.By introducing composite K DP , the negative K DP signatures associated with vertically oriented ice particles in strong electric fields are analyzed.The average composite K DP across different layers, ranging from the 10°C to the 30°C isothermal layers, exhibited a descending trend from positive to negative values or from negative to more negative values before the first IC lightning flash.Subsequently, the average composite K DP increased following the discharge of the first IC lightning.It is worth noting that, as the time approached the first IC lightning flash, the average composite K DP demonstrated a further decrease to more negative values at higher altitudes.Specifically, the average composite K DP started to decrease from a positive value to a negative value above the 10°C isothermal layer, approximately 7 min before the first IC lightning flash.The average composite K DP above the 30°C isothermal layer reached the most negative value 30 s before the first IC lightning flash.Similar signatures are obtained by analyzing the average K DP distributions in the upper parts of the thundercloud.
These results demonstrate that the mean canting angle of ice particles in the upper parts of the cloud changed from horizontal to vertical approximately 7 min before the first IC lightning flash.The concentration of vertically aligned ice particles progressively intensified with the strengthening electric field.The upper layers exhibited a higher concentration of electrically vertically oriented ice particles compared to the lower layers, probably due to the accumulation of more ice particles in the upper regions caused by updrafts.The concentration of electrically vertically oriented ice particles at the cloud top reached the maximum at about 30 s before the first IC lightning flash.This probably suggests that the electric field strength in the top parts of the cloud reached the maximum, potentially contributing to the initiation of the first IC lightning flash.With the occurrence of the first IC lightning flash, ice particles in the upper parts of the thunderstorm recovered to a horizontal or random alignment.Due to a short time from tens of milliseconds to a few seconds is taken to make the electrically aligned ice particles reorient after an abrupt change in electric fields caused by a discharge (Caylor & Chandrasekar, 1996;Hendry & McCormick, 1976;Zrnic & Ryzhkov, 1999) and the time resolution of MP-PAWR is 30 s, thus the signature that the average composite K DP increase following the first IC lightning flash indicates that the mean canting angle of ice particles restored from vertical to horizontal or more random orientation due to the weakening strength of the electric field after the lightning discharges.
Although uncharged ice particles can also be vertically aligned by the strong electric field, considering that the negative K DP value exhibits a linear relationship with the concentration of vertical alignment of ice particles in the upper parts of the cloud, the increased concentration of vertically oriented ice particles indicates the presence of stronger electric fields resulting from a higher concentration of the upper charge region.Thus, the observed electrical alignment signatures of ice particles in the upper parts of the thunderstorm provide insights into the evolution of the upper charge region.Combined with the robust reflectivity signature of graupel in the mixedphase region, the negative K DP signatures which are physically connected to the electrical-related ice particles could be helpful to offer a description of the early electrification process of thunderstorms, which probably help to improve the probability of detecting the first IC lightning flash.The changes in the average composite K DP values above different altitudes are related to variations in the height and concentration of the upper charge region, which is associated with the upper positive charge region in the typical bipolar charge structure of a thunderstorm.In this paper, 7 min before the first IC lightning flash, the upper charge region gradually formed under the 20°C layer (including the 20°C layer).With the upward motion of charged ice particles due to the updraft, the upper charge region subsequently accumulated above the 30°C layer in 4.5 min.It means the electric field built up and enhanced in the thunderstorms, resulting in an increased concentration of vertically aligned ice particles in the upper parts of the cloud.As analyzed in this paper, beneath the 20°C layer where ice particles are normally mixed with abundant graupel particles, the negative K DP signatures of vertically aligned ice particles could be masked by the K DP of graupel.At 7 min before the first IC lightning flash, the presence of a negative K DP signature at the 20°C layer suggested that the electrification initiated below the 20°C layer (including the 20°C layer) at least 7 min before the first IC lightning flash.
In summary, due to the high spatial and temporal resolution, MP-PAWR in normal operation offers the capability to detect the rapid evolution of microphysical structures in thunderstorms and observe the specific electrification processes that are challenging for traditional mechanical weather radar to detect in normal operational volume scan mode.The introduction of the composite K DP will help to observe the evolution of thunderstorm electrification processes more clearly, especially for confirming the onset of electrification combined with the presence of graupel in the mixed-phase region.The observation results in this paper illustrate the electrical alignment signatures of ice particles in the upper parts of the cloud before the first IC lightning flash, indicating an early electrification process where the electric fields build up and gradually intensify in the upper parts of the thunderstorm.However, there are still further studies that need to be done.The analysis of electrical alignment signatures of ice particles before the first IC lightning flash is based on a single isolated thunderstorm case in this study.In this case, the isolated thunderstorm was located approximately 30 km away from the radar before the first lightning flash, and the center of the thunderstorm is about 40 km away from the radar.The highest elevation angle was below 27°which allows the estimated K DP can be used without correcting the K DP bias with elevation angle.Nevertheless, the analysis of these signatures should be extended by correcting the elevation angle dependence bias of K DP and validated using three-dimensional (3-D) lightning location data in more cases.Only in this way can the phased array weather radar fully utilize its ability to observe the lightning and thunderstorms in the entire sky.

Figure 2
Figure2shows range-height-indicator (RHI) scan results of raw Φ DP , filtered Φ DP , raw K DP , and estimated K DP at an azimuth angle of 248.4°at 16:09:15 JST on 20 August 2019.The filtered Φ DP and estimated K DP are smoother and cleaner than the raw data.The discontinuity of raw Φ DP and filtered Φ DP at about 35°and 70°of elevation angle is mainly caused by different system differential phase shifts between different transmit beams of phasedarray radar in electronically scanned elevations.It is obvious that Φ DP decreased in the upper region of the cloud,

Figure 1 .
Figure 1.The radial profiles at an elevation angle of 4.99°and azimuth angle of 330°at 16:00:24 JST on 20 August 2019, Saitama University.(a) Radial profiles of raw Φ DP , filtered Φ DP and rainfall rate.The rainfall rate is estimated by the R (Z H , K DP ) relationship from Suezawa et al. (2019), (b) radial profiles of raw K DP and estimated K DP .

Figure 3 .
Figure 3. CAPPI images at an altitude of 10 km at 16:09:15 JST on 20 August 2019.(a) Corrected Z H , (b) estimated K DP .The black line is the azimuth angle of 248.4°.The blue and orange arcs represent 15 and 60 km away from MP-PAWR.The black point marks are the IC lightning discharge sources.The lightning discharge points fall out of the CAPPI results at altitude of 10 km is because the lightning discharge points were selected by the area of composite reflectivity at the 10°C layer (7 km), which the lightning discharge points distributed across all heights.

Figure 5 .
Figure 5. Radial profiles of measured Z H (dBZ) (red line) and Z H after attenuation correction (blue line) at an elevation angle of 4.99°and azimuth angle of 330°at 16:00:24 JST on 20 August 2019.

Figure 6 .
Figure 6.CAPPI results of Z H and K DP of the isolated thunderstorm at the 10°C layer on 20 August 2019, Saitama University.(a-f): Z H , (g-l): K DP from 15:28:08 to 15:47:59 JST.The time intervals are 4 min.Blue, orange, and green arcs represent 15 km, 30 km, and 60 km away from MP-PAWR.The early isolated thunderstorm echo at the 10°C layer is marked as black circles in (a, h).The analyzed isolated thunderstorm located in a region as a black box marked.

Figure 7 .
Figure 7. Time-height images of the isolated thunderstorm evolution from 15:10:15 to 16:05:22 JST.(a, b): Z H and K DP at a range of 40 km from the radar at a 260.4°a zimuth, (c, d): Z H and K DP at a range of 42.6 km from the radar at a 262.8°azimuth.The time resolution is 30 s.The isothermal levels are marked as black dot lines horizontal in all panels.The left vertical black lines in all panels represent the time of 15:34:35 JST when the moderate reflectivity raised above the 0°C layer in (a), and the right vertical black lines in all panels mark the time of 15:40:03 JST which was the first IC lightning flash occurred.The black box in (d) marks a negative K DP signature which appeared approximately 1.5 min before the first IC lightning flash between the 0°C layer and the 10°C layer.

Figure 8 .
Figure 8. Vertical cross-sections at a range of 42.6 km from the radar at azimuth from 240°to 270°on 20 August 2019, Saitama University.(a-f): Z H , (g-l): K DP , (m-r): ρ HV from 15:35:05 to15:40:03 JST.The time intervals are 1 min.The black circle in (k) marks a negative K DP region with a minimum value of less than 1°/km.

Figure 10 .
Figure 10.Horizontal cross-section images at the 10°C layer on 20 August 2019, Saitama University.(a-f): Composite Z H , (g-l): CAPPI of K DP , (m-r): Composite K DP , from 15:32:06 to 15:42:02 JST, time intervals are 2 min.Three arcs represent 15, 30, and 60 km away from MP-PAWR.The black point marks are the IC lightning discharge sources.The analyzed isolated thunderstorm located in a region as a black box marked.

Figure 11 .
Figure 11.Average composite K DP at and above different layers in the upper parts of the isolated thunderstorm.(a) Average composite Z H values above the 10°C layer (the green line) and IC lightning flash rate (the black fold line), (b) average composite K DP values above the 10°C to 30°C layer (The average composite K DP at the 10°C layer: blue line, 20°C layer: green line, 25°C layer: black line, 30°C layer: red line).

Figure 12 .
Figure 12.Time-height result of average K DP of isolated thunderstorms.(a) Time-height image of average K DP , the black fold line is IC lightning flash rate, and black horizontal lines mark the different isothermal levels; (b) average K DP at different isothermal layers from the 10°C to 30°C layer (The K DP at the 10°C layer: blue line, 20°C layer: green line, 25°C layer: black line, 30°C layer: red line), and the black horizontal dot line marks the reference line of K DP equals 0°/km.After 15:32:36 JST (as the left vertical line marks), a negative K DP region appeared at the top of the cloud and raised above the 25°C layer at 15:35:05 JST (as the middle vertical line marks).The right vertical line marks the first IC lightning flash that occurred at 15:40:03 JST.