INTRODUCTION

The development of high depth cumulonimbus clouds can lead to the occurrence of such dangerous phenomena as heavy rainfall, squalls, tornadoes, hail, and thunderstorms. The formation of these phenomena is closely related to both the microphysical, dynamic and electrical characteristics of clouds [1]. Studies aimed at studying the relationship between the parameters of electrical activity and the characteristics of thunderstorm clouds are of great practical importance. In [1–4], a strong correlation was shown between the lightning frequency, the volume of the supercooled part of the cloud, and the intensity of precipitation.

One of the most dangerous and destructive phenomena associated with convective clouds is a tornado. The wind speed in it can reach 50–100 m/s with a strong upward component and cause significant destruction. The power and lifespan of waterspouts are less than those of continental ones, and this greatly complicates their detection and study.

Currently, magnetohydrodynamic and electrohydrodynamic theories of the appearance of tornadoes are being actively discussed [5–8]. One of the important signs that confirms the validity of the role of electric forces in the formation and maintenance of rotational movements in a tornado is the increased electrical activity of the cloud observed in some cases where a tornado exists.

This article analyzes the case of the formation and development of a thunderstorm over the Gulf of Finland from 14:00 to 16:00 UTC on August 12, 2018. During its evolution, a waterspout formed above the water, and heavy rainfall was observed. The characteristics of the Cb and the atmosphere obtained using the C-band Doppler radar, the SEVERI radiometer installed on the Meteosat satellite, the Blitzortung lightning detection system, and aerological observations are considered.

The purpose of this work is to reveal the features of the microphysical structure of a cumulonimbus cloud at the beginning of thunderstorm activity and during its maximum, during the existence of a waterspout, as well as to identify regression relationships between the radar and electrical characteristics of the cloud.

Weather Conditions and Structure of the Atmosphere

At the time under consideration, the state of the atmosphere was characterized by the passage of the occlusion front; Cb was in close proximity to the occlusion point and the center of the cyclone. This location corresponds to one of the most common situations of waterspout formation over the Baltic Sea [9]. To analyze the structure of the atmosphere and assess the convective instability, the data of rawinsonde sounding at Voeikovo station, was modified in the surface layer to take into account the underlying surface—the waters of the Gulf of Finland.

The level of free convection and lifted condensation level (LCL) coincided in height (59 m), the pressure at this level was 989 hPa, i.e., the cloud base could be located very low (a similar feature of the structure of the atmosphere was noted by us when considering Cb with a tornado over Lake Ladoga [1]). The equilibrium level (EL) was at 8831 m; thus, the convection depth (the thickness of the convectively unstable layer) was 8772 m (EL−LCL). The total energy of instability CAPE (Convective Available Potentional Energy) is 1542 J/kg, and this value indicates the likelihood of developing hazardous phenomena.

To study the conditions for the formation of a waterspout, one can rely on research data [9], where the Baltic Sea region was also considered. In our case, there is a greater instability than the average values in [9]: the level of condensation is much lower, the convection depth is greater, and there is a significant difference in water surface temperature and the temperature at the level of 850 hPa, exceeding 13°C.

Characteristics of a Cb during Development

Let us consider the change in the characteristics of the cloud during its existence. For the starting point “0 minutes” we will take 14:00 UTC, the time when the radar echo first appeared. The investigated cloud existed until the 120th minute, after which it merged with several convective cells. At the 20th minute, the first lightning discharges were recorded, and at the 60th minute, their frequency was maximum. The tornado was observed at the 80th minute. Immediately before observing the tornado at the 70th minute, the cloud under consideration began to merge with another Cb, which ended by the 80th minute.

In Fig. 1a, the time variation of the cloud top height under study is shown Htop according to the measurements of the C-band Doppler radar and the SEVERI radiometer, the frequency of lightning is also given according to the data of the Blitzortung lightning detection system. Note that the results of measuring the lightning frequency using this system correspond to the lightning frequency recorded by the Alves lightning detection system of Rosgydromet [10].

Fig. 1.
figure 1

Temporal variations (a) in the altitude of cloud top Htop according to radar (radar) and radiometer (Meteosat) data and the frequency of lightning discharges F; (b) in the radar lightning hazard criterion Y, volume of the supercooled part of a cloud V35, and the precipitation flux G.

Full size image

Before the start of thunderstorm activity (0th and 10th minutes) Htop did not exceed 7.5 km (Fig. 1a), maximum reflectivity Zmax was less than 50 dBZ. At the 10th minute, heavy rainfall begins to fall out with maximum intensity ∼30 mm/h. At the moment the first discharges appear (20th minute) Htop reaches 10 km, Zmax > 60 dBZ, lightning frequency F = 0.6 min−1 (Fig. 1a). From this moment on, Cb reaches a maximum in its development and passes into mature stage, being in it for the rest of the observed lifetime. In this time interval Zmax = 53–63 dBZ; according to radar Htop was within the range from 9.9 to 10.8 km.

Ci were present in the study area, radiation temperature equaled −46°C at the cloud top. From the 50th minute, Cb became sufficiently developed, and its top exceeded the Ci top height. According to radiometric measurements (taking into account rawinsonde data), the height of cloud top was in the range of 10.6–10.8 km, and according to radar measurements, it was 9.7–9.9 km. This difference can be explained by the fact that the cloud was at a distance ~90 km from radar. At such a distance, the errors of the radar method are quite large (the standard deviation of the measured cloud height for the radar is 0.75 km). The error in measuring the Cb top height from the data of measurements of the radiation temperature by the SEVERI device is determined by the instrumental error of the device itself (0.13 K), the methodical error associated with the fact that the blackbody radiation is formed by a certain cloud layer, as well as the error in the temperature measurements by the radiosonde. It was shown that the layer path that forms the blackbody radiation for dense Cb does not exceed 100 m [11]. The temperature measurement error by the radiosonde is close to 0.5°C. Thus, the total error in measuring the high depth Cb top height can be estimated at ±100 m. Taking into account the parallax phenomenon, as well as the presence of Ci, it can be assumed that this error is about two times greater. Similar results were obtained in the study [12].

The lightning frequency at the 60th minute was maximum (F = 8.3 min−1), and in the 70th minute, it sharply decreases to 1.6 min−1. Further, after the clouds merge, an increase in F up to 3.7 min−1 was observed. A similar effect was recorded earlier [13]. The appearance of a tornado (80th minute) barely changed the frequency of lightning, in contrast to our earlier work [1]. At the end of the observations, discharges were recorded with a frequency of 2.3 min−1.

Thus, the increased electrical activity of the cloud was observed 20 min before the formation of the tornado. The frequency of lightning at the time of observation of the tornado was 55% less than at the peak of thunderstorm activity, which contradicts the electrical theory of the formation of a dangerous vortex [5–8]. It is possible that a more significant contribution to the formation of the tornado was made by the process of merging of convective cells [14, 15].

Microphysical Structure

A generalized simplified diagram of the microphysical structure of a cloud at the main moments of thunderstorm activity is shown in Fig. 2. The diagram was drawn up based on the analysis of vertical sections of the polarization radar characteristics of the cloud using works [1, 16].

Fig. 2.
figure 2

A simplified diagram of the microphysical structure of a cloud at the main moments of thunderstorm activity. Arrows at the bottom of the cloud indicate the direction of vertical flows; the gray arrows at the top of the diagram are the cloud displacement vectors

Full size image

At the 10th minute, rather intense precipitation was already falling out of the cloud, and large nonspherical drops of rain and drizzle were recorded in its lower part (drops of precipitation with a diameter of less than 0.5 mm). Cloud drops, drizzle, small and large crystals were recorded above the zero isotherm. These fractions were recorded up to the isotherm −22°C. Above was only the crystalline fraction. It is known that the interaction of small crystals with graupel in the presence of a liquid-droplet fraction generates electrification of the cloud, which was observed in the case under consideration. Areas of updrafts and downdrafts are registered in the cloud, the latter being associated with precipitation.

The transition of the cloud to the thunderstorm stage occurred between the 10th and 20th minutes. Significant development of Cb was noted, the top height increased by more than 2 km, reaching 10 km. The top growth rate was on average 3.4 m/s. The microphysical structure of the cloud at the 20th minute is basically preserved, in the upper part of the cloud only small crystals are recorded. Large crystals appear below 8 km. In the lower part of the cloud, there are watered crystals, and the process of sedimentation is enhanced. Thus, the conditions that lead to the electrification of the cloud remain. It can be assumed that the rapprochement of precipitation particles with cloudy particles and between themselves triggers the electrification caused by corona discharges [17].

The microstructure of the cloud at the moment of the maximum lightning frequency at the 60th minute in general resembles the microstructure recorded at the moment of the first discharges. The differences are that, before the peak of thunderstorm activity, watered crystals are fixed in the layer from the cloud base to a height slightly exceeding the level of the zero isotherm, apparently in the zone of updrafts, and at the moment of the maximum discharge frequency, these hydrometeors were detected only in the immediate vicinity of the cloud base.

A distinctive feature of the cloud microstructure at the time of observation of the tornado (80th minute) is that large flattened drops lifted by a strong updrafts associated with the tornado are observed up to a height of 4 km, while large crystals are recorded up to 9 km. The volume of the supercooled part of the cloud with large crystalline particles is less than at the 60th minute (see Fig. 1b). This could be the reason for a decrease in the frequency of lightning at the time of observation of the tornado.

Relation of Lightning Frequency to Radar Characteristics

In Fig. 1b, the time course of radar characteristics is presented: the volume of the cloud is above the zero isotherm with a reflectivity of more than 35 dBZ (V35), precipitation flux (G) (the volume of water falling out of the convective cell per unit time (m3/s)), the lightning hazard radar criterion Y calculated by the method described in [3]. Temporal changes in V35 and G occur quite synchronously with the lightning frequency. The maximum correlation coefficient is observed between F and V35 (r = 0.95). A close relationship is present between the lightning frequency and the volumes of the supercooled part with reflectivity of 30 and 35 dBZ (r = 0.96 and 0.95). Similar results were obtained earlier when analyzing the development of Cb both in the Northwest and for the North Caucasian region [3, 10]; however, for a cloud with a tornado over Lake Ladoga [1], these correlation coefficients were only 0.66 and 0.62, respectively.

From radar data characterizing precipitation, a closer relationship is observed in the precipitation flux with the discharge frequency (r = 0.90). A high correlation coefficient between F and Y, which we noted earlier [3], has not been established in the present experiment.

For an ultra-short-term forecast of the formation of a waterspout, the characteristics of a cloud at the stage of development preceding its appearance are of particular importance. The analysis of the regression relationships of the parameters studied above was carried out separately for the development period and for the entire lifetime of the cloud. It was found that the coefficients of the regression equations change little. This may indicate that the vortex of the observed scale did not lead to a significant change in the characteristics of the relationships between the lightning frequency and the parameters of the cloud.

CONCLUSIONS

A thunderstorm with a waterspout observed over the Gulf of Finland was investigated. The top height of Cb reached a height of 10.8 km, and the frequency of lightning discharges was 8.3 discharges per minute, which is a significant value for the North–West of the Russian Federation.

Analysis of the microstructure of the cloud, obtained from the data of Doppler and polarization radar measurements, showed that during the transition of the cloud to the thunderstorm stage, it contained cloud drops, drizzle drops, small and large crystals. Immediately before the appearance of the first discharges, watered hailstones appear in the cloud. At the moment of the maximum lightning frequency, watered crystals were found only directly at the cloud base, up to this moment they were recorded up to the zero isotherm and somewhat higher. At the time of the tornado, large crystals reach a maximum height of 9 km.

The maximum frequency of lightning was observed 20 minutes before the formation of the tornado, at the time of observation of the tornado it was 55% less than at the peak of the thunderstorm activity, which does not confirm the electrical theory of the formation of a dangerous vortex.

The previously established close correlation between the frequency of lightning and the volume of the supercooled part of the cloud, as well as the precipitation flux, was confirmed.