A Study of Radar Echoes and their Relation to Lightning. Discharges of Thunderclouds in the Hokuriku District

Transcription

1 April 1993 K. Michimoto 195 A Study of Radar Echoes and their Relation to Lightning Discharges of Thunderclouds in the Hokuriku District Part II: Observation and Analysis of "Single-Flash" Thunderclouds in Midwinter By Koichiro Michimoto Department of Geoscience, National Defense Academy, Hashirimizu, Yokosuka, Kanagawa,239, Japan (Manuscript received 21 November 1990, in revised from 2 February 1993) Abstract In the present work, winter thunderclouds and active convective clouds were observed by means of radar with CAPPI (Constant Altitude Plan Position Indicator). The lightning activity is monitored by a local network of sferics direction finders around Komatsu Airport. In midwinter, from January to early March, active convective clouds, 30 dbz echo tops of which reason for keeping such clouds in a weak or non-lightning situation. Concerning the classification of the lightning activity, the author has proposed the following criteria: Along with the results of thundercloud investigation stated in Part I of the present article (Michimoto, 1991), it is concluded that the necessary conditions for convective clouds to generate lightning discharge are as follows: (3) The clouds have to involve rapid development of graupel particle precipitation; specifically, they have to involve formation and rapidly vertical movement of 40-to-50 dbz echo cells. 1. Introduction In late autumn and winter, very active convective clouds frequently form along the coastline of the Sea of Japan, being associated with strong advection of Siberian air masses over the Sea of Japan. Their electrical activity is generally very low and clouds exhibiting only one or two lightning flashes throughout their whole duration are the most common, and are popularly called "Single-Flash Thunderclouds" or "Ippatsu-Rai" in Japanese. In the present paper, the thunderclouds of this type are consequently referred to as "IPR", and the similar convective clouds that exhibit no lightning activity, as "NLA". Takahashi (1984) carried out a numerical simulation of thunderclouds and pointed out that, as the main charge separation process, graupel particles were negatively charged in the cloud between the els. Further, the charge separation process around portant role in the electrical activity of the clouds. Tomine et al. (1986) and Michimoto (1988a, 1988b and 1989a) carried out numerous winter thunderclouds observations in 1982/1983, 1983/1984, 1984/1985 and 1987/1988 in the area surrounding

2 196 Journal of the Meteorological Society of Japan Vol. 71, No. 2 Fig. 1. The relationship of lightning activity ter convective clouds. The observation was made in the vicinity of Komatsu Airport. Temperature at the top of echoes is indicated on the abscissa and the altitude When the temperature at the top of echoes line), convective clouds exhibit no lightning activity (black circles). (B) When the and-dotted line), convective clouds exhibit relatively intense lightning activity (cross lower than 1.8km, convective clouds exhibit no lightning activity (open circles) or only very weak lightning activity (black triangles). (after Michimoto (1989b) ) Komatsu Airport and revealed many features of winter thunderclouds and convective clouds. The main points of the results are illustrated in Fig. 1 as follows: Convective clouds, the 20 dbz echo tops of which exhibit lightning activity (area A in Fig. 1). When 3. Case study of observational results and in altitude, convective clouds, the 20 dbz echo tops summary hibit relatively high electrical activity (area B in Fig. ture level is lower than 1.8km in altitude, convective clouds exhibit either very low or no lightning activity, even if their 20 dbz echo tops develop higher Kitagawa (1989) proposed a formular concerning vertical air motion in convective clouds. He classified vertical motions: at an altitude higher than 3km, the updraft velocity takes an arbitrary value depending on the dynamic condition of the cloud, but at an altitude lower than 3km, the velocity is restricted to a lesser value determined by its altitude from the ground (e. g., less than 3m/s at an altitude of 1.8km). This can physically explain the results temperature level is higher than 1.8km in altitude, the updraft velocity at that level becomes large enough to separate the charge that causes lightning in altitude the updraft velocity at that level is not sufficient to cause lightning activity. Thus, the alti- factor for judging the electrical activity of convective clouds. Subsequently, Michimoto (1991) investigated many summer and early winter thunderclouds, paying attention to the behavior of 30 dbz radar echoes in addition to that of 20 dbz echoes. He emphasized that clouds exhibit lightning activity without temperature level. In Part II, the author aims to reveal the meteorological property of convective cloud systems in mid- ture level to exhibit very weak or no electrical activity. He also aims to identify meteorological factors that are contributing causes of lightning discharges in midwinter cloud systems. 2. Instrumentation and observation The author used a local network consisting of a 5.7-cm radar and MHz sferics direction finders installed in the vicinity of Komatsu Airport. The detailes of the systems are described in Part I of the paper. The author carried out the observation from January to March in both 1988 and 1990 and obtained examples of 10 cloud systems which were either IPR or NLA clouds. The meteorological data at a given altitude and time were interpolated from the rawinsonde data obtained routinely at 0900 and 2100 JST at Wajima Aerological Station (operated by the Japan Meteorological Agency) 120km north-northeast of Komatsu Airport. In this section, six typical cloud systems (four IPR and two NLA) are introduced and illustrated in Fig. 2 to 16. These figures are composed of the CAPPI image delineated by radar reflectivity contours at 5 dbz intervals from the outermost 20 dbz contour at

3 April 1993 K. Michimoto 197 Fig. 2. Temporal variation of altitude of echoes (reflectivity from 20 to 40 dbz) on 9 January Fig. 4. Temporal variation of altitude of echoes (reflectivity from 20 to 40 dbz) on 13 January Fig. 3. Radar reflectivity pattern shown by the contours of 25, 30, 35 and 40 dbz at selected altitudes between 1804 and 1812 JST on 9 January The echo area higher than 40 dbz is blackened. An arrow indicates the 40 dbz echo (1km altitude) 1km in altitude (omitted in Cases 1 to 3 for want of space), the temporal variation of the altitude of each radar reflectivity, and the horizontal radar reflectivity pattern at 5 dbz intervals from the surface to an altitude of 3 or 4km (omitted instead of the use of two CAPPIs in Case 6). Case 1: 9 January 1988 A cloud system which caused a lightning flash at 1810 JST was located over the Sea of Japan. Figure 2 indicates an instantaneous appearance of the ture levels (0.5 and 1.7km in altitude) at 1810 JST, and persistence of the 20 and 25 dbz echoes at an almost constant altitude. In this figure, both the 30 and 35 dbz echoes exhibited relatively rapid ascent km in altitude). All of the echoes remained stationary for a while after the occurrence of a lightning flash at 1810 JST. Comparing Fig. 2 and Fig. 3, it can be seen that a 40 dbz echo core, which is indicated by an arrow in at the time of the flash. Before the flash, the 30 dbz echo moved upward at 1804 JST and downward at the 40 dbz echo suddenly appeared at temperature Case 2: 13 January 1988 A ligntning flash occurred in a cloud system over the Sea of Japan at 0042 JST. As shown in Fig. 4, the 30 dbz echo ascended rapidly very close to the while the 20 dbz echo remained stationary at the Fig. 4 that the temporal curves of the 25-to-35 dbz echoes exhibited a repetitive ascent and descent motion around each constant level before the occurrence of the flash at 0042 JST. As shown in Fig. 5, formations of the strongest echo cells (cell renewals), which are indicated by arrows, were observed three times: at the altitude of 3 km at 0038 JST, between 2 and 4km at 0042 JST, and below 1km at 0044 JST. It should be noted that the lightning flash at 0042 JST occurred simultaneously with the sencond renewal of the cell (between Case 3: 7 March 1988 A cloud system first developed over the Sea of Japan, moved overland and caused a lightning flash at 0221 JST. Figure 6 shows that the 35, 40 and 45 (about 3km in altitude), while the 25 and 30 dbz echoes ascended slowly and then remained at an almost constant altitude. This figure also shows the echo center position when the system was crossing the coastline, which is indicated by a broken line. Figure 7 shows that the 35-to-45 dbz echoes, which are indicated by arrows, formed rapidly the flash at 0221 JST (a cell renewal occurred).

4 198 Journal of the Meteorological Society of Japan Vol. 71, No. 2 Fig. 5. Radar reflectivity pattern shown by the contours of 20, 25, 30, 35 and 40 dbz at selected altitudes between 0034 and 0044 JST on 13 January The echo area higher than 40 dbz is blackened. Arrows indicate renewal cells which appeared before and after a lightning flash. Fig. 7. Radar reflectivity pattern shown by the contours of 35, 40 and 45 dbz at the selected altitudes between 0211 and 0223 JST on 7 March (In this case, the echo area higher than 40 dbz is not blackened.) Two arrows indicate the 40 and 45 dbz echoes which rapidly formed when the cloud system moved over the land and a lightning flash occurred simultaneously. Fig. 6. Temporal variation of altitude of echoes (reflectivity from 20 to 45 dbz) on 7 March Case 4: 22 January 1990 Figure 8 indicates the 2km altitude CAPPI at 2229 JST. A cloud system containing one 40 dbz echo cell was located mostly over the sea near the coast between Komatsu and Wajima, and a lightning flash occurred at 2229 JST near the coastline. The location of the lightning flash, which was detected by the direction finders, is indicated by a cross mark. Figure 9 shows that the lower edge of the 40 dbz echo reached the sea surface at 2224 JST, about 5 minutes before the occurrence of the flash, while The location of the flash which was de- tected by the VHF sferics direction finders (3km in altitude) at 2130 JST and was descending very slowly around the time of the flash. Figure 10 shows that, before the occurrence of the flash, the strong echo of 40 dbz persisted for 10 minutes at Fig. 8. CAPPI on 22 January 1990 at 2km, 2229 JST from the 5.7-cm wavelength weather radar. The arrow indicates an echo cell of the IPR cloud. The 20 dbz contour is indicated by a bold line and the echo area higher than 35 dbz is blackened. is indicated by a cross mark. for about 4 minutes immediately before the flash oc- currence. The downward motion of the strong echo portion of this strong echo descended 1km vertically core, indicated by arrows, suggests that the whole

5 April 1993 K. Michimoto 199 Fig. 9. Temporal variation of altitude of echo (reflectivity from 20 to 40 dbz) on 22 January Fig. 11. Same as Fig. 8, but on 23 January 1990 at 1km, 1800 JST. The arrow indicates an echo cell of the NLA cloud. The 20 dbz contour is indicated by a bold line and the echo area higher than 35 dbz is blackened. Fig. 10. Radar reflectivity pattern shown by the contours of 30, 35 and 40 dbz at selected altitudes between 2200 and 2235 JST including the time of flash on 22 January The echo area higher than 40 dbz of the cloud shown in Fig. 8 is blackened. Arrows indicate the descent motion of the 40 dbz echo from 2km to the ground surface. graupel region descended at a velocity of about 5 m/s. Case 5: 23 January 1990 Figure 11 shows the 1km altitude CAPPI display at 1800 JST, indicating a snowcloud over the Sea of Japan. The 20 dbz echo area of the cloud, which is indicated by an arrow, had an equivalent diameter larger than 50km, while the 30 dbz echo area had one of only 20km. Figure 12 indicates that the lower edge of the 40 dbz echo reached the sea surface at 1800 JST, by which time the 30 dbz echo had already crossed the Fig. 12. Same as Fig. 9, but on 23 January CAPPI displays at 0029 and 0046 JST, respectively. The 20 dbz echo area of this cloud system, indicated detected during the whole observation period, this by an arrow, had an equivalent diameter greater was a NLA cloud system. The rapid descent of 30- than 50km, while the 30 dbz echo area had an to-40 dbz echo cells was observed around 1755 JST equivalent diameter of only 20km over the Sea of as indicated by arrows in Fig. 13. Japan. The lack of sferics during the whole obser In this case, although the growth of a new cell was observed at an altitude of 3 to 4km at about 1755 JST (see Fig. 13) in a manner similar to that in Case 3 (see Fig. 7) and the lower edge of the 40 dbz echo reached the sea surface at about 1800 JST (see Fig. 12) in manner similar to that in Cases 3 and 4 (see Figs. 6 and 9), no flash occurred, and the cloud system maintained the NLA status. It should be level was significantly low, namely, 1.3km. Case 6: 24 January 1990 Figures 14 and 15 indicate the 1km altitude

6 200 Journal of the Meteorological Society of Japan Vol. 71, No. 2 Fig. 13. Same as Fig. 10, but at selected altitudes between 1740 and 1810 JST on 23 January Arrows indicate the descent motion of the 40 dbz from 3km to the ground surface. vation period indicated that this was an NLA cloud system. It can be seen in Fig. 16 that, among the several cells associated with this NLA cloud system, the 30 (about 3km in altitude). In the first half, the 35 dbz echo cell grew rapidly and its lower edge reached the sea surface at 0029 JST. And in the latter half, the lower edge of the 40 dbz echo cell reached the sea surface at 0044 and 0055 JST. The behavior of echo cells was similar to that in Case 5 (see Fig. 12). The significantly low, namely, 1.3km. Fig. 14. Same as Fig. 8, but on 24 January 1990 at 1km, 0029 JST. The arrow indicates an echo cell of the NLA cloud. The 20 dbz contour is indicated by a bold line and the echo area higher than 35 dbz is blackened. Concerning the above six cloud systems, the main quantitative results are listed in Table 1. As seen in the table, all cloud systems were most active of midwinter season, because their cloud tops developed very high in the troposphere, the 20 and 30 dbz ature levels, respectively. However, of the six cloud systems, two cases exhibited no lightning activity, and the other four, only one lightning flash for their duration, i. e., they were typical so-called "Single- Flash Clouds" ("Ippatsu-Rai" in Japanese). Considering the fact that, in the summer season, cloud 1990 at 1km, 0046 JST. The arrow indi- Fig. 15. Same as Fig. 8, but on 24 January systems that develop to the same temperature altitude always exhibit much stronger ligtning activity, 20 dbz contour is indicated by a bold line cates an echo cell of the NLA cloud. The the above results seem very noteworthy. Comparing the non-flash clouds group with that of single- blackened. and the echo area higher than 35 dbz is flash clouds, no marked activity difference was found between the two groups, and the only difference be discussed in the following sections. Namely, it was 1.3 to 1.2km for non-flash clouds and 4. Characteristic features of active concevtive 1.6 to 1.3km for single-flash clouds. A meteorological explanation as to why lightning activity clouds in winter from active convective clouds is so rare in midwinter will The observational results of the present work, to-

7 April 1993 K. Michimoto 201 Fig. 16. Same as Fig. 9, but on 24 January gether with those in Part I of the same titled paper, revealed the following features of winter active convective clouds: 4-1. When sufficient unstable stratification is formed in the troposphere, e. g., when cold air masses advect over a relatively warm sea surface, active convective clouds are formed and develop high in the troposphere. Though their cloud tops develop to the temstrong lightning activity. nor penetrate the tropopause. As for the cell structure, clouds develop as a single cell or plural isolated cells, and except for synoptic-scale cold fronts, cloud systems seldom develop into organized multicell clouds with strong convection. Though Sakakibara et al. (1988) observed multi-cell typed cloud systems in the same Hokuriku district, those systems were very weak convective clouds without lightning flashes. This feature is a point of contrast with the summer active convective clouds that frequently develop as multi-cell or supercell storms and fully penetrate the tropopause (Browning, 1982). Nevertheless, temporal renewals of individual echo cells are frequently observed in winter convective clouds. MacGorman et al. (1989) pointed out that renewal of echo cells reinforces interaction of precipitation particles and strengthens the precitpitaion rate The echo cells of 35 and 40 dbz initially ap- In fact, Table 1 shows that the ascending velocities of 30-to-40 dbz echo cells are much higher in glaciation of supercooled droplets occurs most vigorously (Ryan et al., 1976). The clouds necessarily involve definite graupel precipitation regions and exhibit strong echo cells of at least 40 dbz. However, in winter, the volume of the graupel precipitation regions is rather small and their duration in the clouds is held to less than 15 minutes When active convective clouds are formed in separation of charge is most effectively going on (Takahashi,1984), is positioned at an altitude ranging from 3 to 1km. The proximity of the sea or ground surface to this level confines the maximum updraft velocity at the level to the range from 6 to 1.5 m/s (Kitagawa, 1992). This limitation of up- urally influences the rate of charge separation in the clouds. Looking at the features described in Items 4-1 and 4-2, the clouds fully satisfy the necessary conditions for generation of lightning discharges stated in Part I of the present article. This raises the question: Why it is that, in midwinter, such active clouds exhibit only weak lightning activity or fail to cause lightning discharges? The features described in Item 4-3 provide the answer. The author (1989b) has already proposed the perature level is higher than 1.8km, clouds exhibit rather strong lightning activity, and when it is lower than 1.8km, there is only very low or no lightning activity. Referring to the present observational results and other midwinter observations (Michimoto, 1993a), these criteria may be supplemented as follows: level is higher than 1.8km, clouds exhibit rather level is between 1.8 to 1.4km, clouds exhibit either weak or no lightning activity. level is lower than 1.4km, clouds never exhibit a lightning discharge. According to Kitagawa's formula (1992) referred to in Item 4-3, the maximum updraft velocities at m/s when altitudes of that level are between 1.8 and 1.4km. Consequently, it reasonably follows that Lure level is confined to less than 3.0m/s, the rate of charge separation is sufficient to cause lightning temperature level is confined to less than 2.0m/s, the rate of charge separation is not sufficient to cause a lightning discharge. Cases 1, 2 and 3 than in Cases 4, 5 and 6. In the level is 1.5km or higher, while in the latter cases, it is 1.3km or less. Averaging the data in the 10 observed cases, including the six cases described in Section 3, the author calculated the maximum ascending velocity of each reflectivity echo and listed the results in Table 2. Although movements of echo cells cannot be equated with air motion, the maximum ascending velocities of echo cells may reflect the updraft velocity to a certain extent. Table 2 shows that the maximum ascending velocities of echo cells in IPR clouds are three to five times as high as those in NLA clouds. Reynolds and Brook (1956) pointed

8 202 Journal of the Meteorological Society of Japan Vol. 71, No. 2 Table 1. List of IPR and NLA clouds in 1988 and 1990 midwinter. Table 2. The average radar echo ascending velocity before IPR and during NLA. out that the formation of a radar precipitation echo did not lead to lightning discharges unless the precipitation echo grew rapidly in the vertical direction. It is apparent that when the ascending velocities of echo cells are lower than certain values, the clouds are not capable of producing a lightning flash. In Table 1, the occurrence of a flash in Case 4 seems to be exceptional, because the altitude of range, and the criterion has to be strictly applied ascending echo motion was observed. It should be noted that the echo center of the cloud system was approaching the coastline to cross it almost simultaneously with the occurrence of a flash (specifically, one minute after the flash occurrence, see Fig. 9). As stated in Section 3, prior to the flash occurrence, the 40 dbz echo cell descended 1km at a velocity of 4m/s. It is highly probable that the topographical effect caused this rapid descent of the graupel region, and that this motion triggered the lightning flash. 5. Conditions for clouds to generate lightning discharges In Part I of the present article, the author pointed out that the necessary conditions for the generation of lightning discharges by convective clouds are that temperature level and that the clouds involve the rapid development of 40 dbz or stronger echo cells. However, these conditions are valid for seasons other than midwinter. To obtain general conditions which are applicable for all seasons of the year, one has to take account of the features of winter convective clouds revealed in the preceding section. The criterion proposed from the present work level must be at least, 1.4km for the generation of a natural lightning discharge. Therefore, this should be added to the conditions for the generation of lightning discharges. level frequently comes down to the aforementioned in discussing the lightning activity of clouds. In Section 3, it is observed that very active convective clouds, the 30 dbz echo tops of which reach Flash lightning activity or no lightning activity in midwinter. This is a direct consequence of the fact that the al- 1.8km (and sometimes lower than 1.4km) and that the updraft velocity at that level is less than 3.0m/s (and sometimes less than 2.0m/s). 6. Conclusions In Part I of the present work, the author investigated summer and early winter thunderstorms using CAPPI radars and the local network of sferics direction finder settled around Komatsu Airport. In

9 April 1993 K. Michimoto 203 Part II, he extended the same field work to midwinter thunderstorms. Summarizing the whole results, he obtained two categories of conclusions. The first one concerns the necessary conditions for the generation of lightning discharges by convective clouds applicable for all seasons of the year and over all latitudes. The major necessary conditions consist of the following three points: 1. For the generation of lightning discharges, the cloud top has to develop high enough in the troposphere; specifically, the 30 dbz echo of the clouds level. 2. For the generation of lightning discharges, there References perature level from the ground where precipitation particles are formed and electrified; specifically, it should be at least 1.4km. 3. For the generation of lightning discharges, the May J. Atmos. Sci., 46, is higher than 1.8km, active convective clouds generate rather strong lightning activity. When the al- 1.8km, active convective clouds exhibit either weak or no lightning activity, even if their 30 dbz echo than 1.4km, convective clouds exhibit no lightning discharge. The conclusions of the second category are highly suitable for midwinter thunderstorms along the coast of the Sea of Japan. In fact, in midwinter almost all lightning activities that occur are weak ones known as "Single-Flash Thunderstorms" ( "Ippatsu- Rai" in Japanese). The necessary condition 2 refers solely to the generation of lightning discharges, and the critical altitude becomes slightly lower when some dischargetriggering effect is at work. On 22 January 1990 winter around Komatsu Airbase, Japan. J. Atmos. Electr., 13, level was 1.3km and yet a single flash was observed. In this case, the radar echo cell was just crossing the coastline and the discharge was inferred to have been triggered by some topographic effect. On 29 January 1991, an aircraft took off from Komatsu Airport and was struck by lightning that was obviously triggered by its invasion into the active convective clouds (Michimoto, 1993b). Acknowledgements The author wishes to express his sincere thanks to the staff members of the Department of Geoscience, the National Defense Academy (NDA), for their encouragement and kind advice throughout this research. Dr. Kitagawa also provided helpful comments. Thanks are also due to the Komatsu Air Weather Service Squadron, JASDF, for providing the thundercloud detection system data. Data processing was carried out with NDA HITAC M-680H. Comments provided by anonymous reviewers were very useful for revising this manuscript. Browning, K. A., 1982: General circulation of middle latitude thunderstorms. Thunderstorms edited by E. Kessler, Vol. II, NOAA, Department of Commerce, clouds have to involve rapid development of graupel U. S. A., Washington D. C., precipitation; specifically, they have to involve formation and vertical movement of 40-to-50 dbz echo ter thunderclouds along the Japan Sea coast. Pa- Kitagawa, N., 1989: Meteorological features of win- cells. pers from a joint technical meeting on electrical discharges, high voltage, IEE Japan, ED , HV- The second category of conclusions concerns the (in Japanese). aerological condition that effectively controls the Kitagawa, N., 1992: Charge distribution of winter thunderclouds. Res. Lett. Atmos. Electr., 12, grade of lightning activity. The altitude of the MacGorman, D. R., D. W. Burgess, V. Mazur, W. D. sary condition 2 also determines the grade of lightning activity as follows: ning rates relative to tornadic storm evolution on Rust, W. L. Taylor and B. C. Johnson, 1989: Light- 22 Michimoto, K., 1988a: A study on thunderstorms in winter in the area surrounding Komatsu by radar. Tenki, 35, (in Japanese). Michimoto, K., 1988b: A method of prediction of thunderstorms in winter by radar echo. Tenki, 35, (in Japanese). Michimoto, K., 1989a: A study on thunderstorms in summer in the area surrounding Komatsu by radar. Tenki, 36, (in Japanese). Michimoto, K., 1989b: A study on thunderstorms in winter in the area surrounding Komatsu. Tenki, 36, (in Japanese). Michimoto, K., 1991: A study of radar echoes and their relation to lightning discharge of thunderclouds in the Hokuriki district, Part I: Observation and analysis of thunderclouds in summer and winter. J. Meteor. Soc. Japan, 69, Michimoto, K., 1993a: A study of the variation of surface electric fields below the winter thunderclouds in the Hokuriku district. J. Atmos. Electr., 13, Michimoto, K., 1993b: On lightning strikes to aircraft in Reynold, S. E. and M. Brook, 1956: Correlation of the initial electric field and the radar echo in thunderstoms. J. Met., 13, Ryan, B. F., E. R. Wishart and D. E. Shaw, 1976: The growth rates and densities of ice crystls between

10 204 Journal of the Meteorological Society of Japan Vol. 71, No. 2 Sakakibara, H., M. Ishihara and Z. Yanagisawa, 1988: Squall line like convective snowbands over the Sea of Japan. J. Meteor. Soc. Japan, 66, Takahashi, T., 1984: Thunderstorm electrification-a numerical study. J. Atmos. Sci., 41, Tomine, K., K. Michimoto and S. Abe,1986: Studies on thunderstorm in winter in the area surrounding Komatsu by radar. Tenki, 33, (in Japanese).