Simultaneously measured current, luminosity, and electric ... - CiteSeerX

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, D10102, doi:10.1029/2010JD015331, 2011

Simultaneously measured current, luminosity, and electric field pulses in a rocket‐triggered lightning flash Xiushu Qie,1 Rubin Jiang,1 Caixia Wang,1 Jing Yang,1 Junfang Wang,1 and Dongxia Liu1 Received 19 November 2010; revised 21 February 2011; accepted 24 February 2011; published 20 May 2011.

[1] Current, luminosity, and electric field pulses in a rocket‐triggered negative lightning

flash have been analyzed based on the channel base current, high‐speed video images, and electric field changes at 30 m from the channel. Among the 31 distinct current pulses, there are 4 return strokes, 18 typical M components, 5 large M components with unusual large peak current in a range of kiloamperes, 3 initial continuing current (ICC) pulses, and 1 stroke–M component (RM) event which exhibits both return stroke and M component features. The geometric mean of peak current is 13.5 kA, half peak width is 28.4 ms, and risetime from 10% to 90% peak is 1.1 ms for the 4 return strokes, while the corresponding values are 243 A, 400 ms, and 319 ms, respectively, for the 18 typical M components and 5.1 kA, 76.3 ms, and 34.6 ms, respectively, for the 5 large M components. The electric field and current waveforms of ICC pulses exhibit features similar to those of the M components, indicating the similarity of their mechanisms. Detectable optical luminosity is found just prior to all the pulse events, even return strokes. The M components are superimposed on a slowly varying continuing current, while the directly measured current prior to the return stroke is not significant. The simultaneous electric field and current waveform of RM implies a superposition of dart leader and M incident wave in the channel, and the possible reason is that two branches with common lower portions coexist simultaneously in the upper part of the discharge channel. Citation: Qie, X., R. Jiang, C. Wang, J. Yang, J. Wang, and D. Liu (2011), Simultaneously measured current, luminosity, and electric field pulses in a rocket‐triggered lightning flash, J. Geophys. Res., 116, D10102, doi:10.1029/2010JD015331.

1. Introduction [2] The discharge current waveforms of lightning flashes, measured at the base of discharge channel, usually exhibit impulsive features. These pulses may be characterized by different peak current, risetime, duration and so on, but basically they are assigned into two categories, return stroke and M component. The leader–return stroke sequence and M component are two typical discharge processes with short‐ duration charge transfer to ground from cloud in both natural cloud‐to‐ground (CG) lightning and triggered lightning. Triggered lightning provides a unique opportunity to study these discharge processes and related electromagnetic field effects, especially from direct measurement of the channel base current and the simultaneous close electromagnetic field. The leader–return stroke sequences in triggered lightning are believed to be very similar to the dart leader– subsequent return strokes in natural CG lightning [e.g., Le Vine et al., 1989; Jordan et al., 1992; Fisher et al., 1993], and there have been many studies on this sequence based on the triggered lightning [e.g., Hubert et al., 1984; Uman et al.,

1 Key Laboratory of Middle Atmosphere and Global Environment Observation, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China.

Copyright 2011 by the American Geophysical Union. 0148‐0227/11/2010JD015331

1997; Rakov et al., 1998, 2005; Qie et al., 2007; Zhang et al., 2009a; Schoene et al., 2010], and will not be reviewed in detail in this paper. [3] M components are perturbations or transient enhancements in the continuing current [Fisher et al., 1993; Thottappillil et al., 1995], the associated channel luminosity [Malan and Collens, 1937; Jordan et al., 1995; Campos et al., 2007; Flache et al., 2008] and the electromagnetic field changes [e.g., Thottappillil et al., 1990; Rakov et al., 1992; Shao et al., 1995; Rakov and Uman, 2003]. Most of the recent findings on M components are from Florida triggered lightning flashes [e.g., Fisher et al., 1993; Thottappillil et al., 1995; Wang et al., 1999; Rakov et al., 1995, 2001] and upward lightning flashes initiated from tall towers [e.g., Miki et al., 2005; Flache et al., 2008; Wang et al., 2007]. A typical M component is characterized by a more or less symmetrical current pulse at the channel base with an amplitude of 100–200 A (1–2 orders of magnitude smaller than return stroke), a risetime of 300–500 ms (3 orders of magnitude longer than return stroke), and a charge transfer of 0.1 to 0.2 C [Fisher et al., 1993; Thottappillil et al., 1995; Rakov and Uman, 2003]. Some M components could have peak current in the kiloamperes range, which is comparable to smaller return strokes [e.g., Rakov et al., 1998; Flache et al., 2008]. Campos et al. [2007] found that nearly 25% of M components occur in the first 10% of the duration of the continuing current, and more than 75% occur in the first half of the continuing current. Thottappillil et al. [1995] found that the occurrence of

D10102

1 of 11

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

M components to the leader–return stroke sequences was 4 to 1 based on the triggered lightning data. [4] The return stroke is usually believed to occur only after the cessation of any preceding current through the channel base. Fisher et al. [1993] found that the return stroke current pulses, characterized by risetime of the order of a few microseconds or less and peak values in the range of 4 to 38 kA, will not to occur until after any preceding current at the bottom of the lightning channel fell below the noise level of less than 2 A, indicating the cut off of preceding channel base current is necessary to the occurrence of return stroke. This has been used as a judgment for return stroke. When there are preceding channel base current, the pulses is referred as M components even if the risetime is as short as a few microseconds [Flache et al., 2008]. [5] Based on the findings that the M component fields appear to be proportional to the time derivatives of the associated M currents, Rakov et al. [1995] proposed a “two‐ wave” mechanism of the lightning M component. According to this mechanism, an M component involves a downward progressing incident wave (the analog of a leader) and an upward progressing reflected wave (the analog of a return stroke). Both the upward and the downward processes contribute about equally to the total charge flowing from the bottom of the channel at any instant of time. Ground is sensed by the incident M wave as a short circuit, so the reflection coefficient for current at ground is close to +1, and the reflection coefficient for the associated charge density is close to −1. At each channel section, the two waves are shifted in time, the time shift being small near ground and increasing toward the cloud. Such a mode of charge transfer to ground is distinctly different from a leader–return stroke sequence in which the latter removes the charge deposited by the former. Rakov et al. [2001] further examined the M component mode of charge transfer to ground by using multiple‐station measurement of electric and magnetic field and measurement of currents at the channel base for triggered lightning channels. They found that the shapes and magnitudes of the measured close electric and magnetic fields are generally consistent with the guided wave mechanism of the lightning M component. [6] Although the “classical” M components occur during the relatively steady continuing current stage following return stroke, as first described by Malan and Collens [1937], the similar perturbations have been also found to occur during the initial continuous current (ICC) stage of rocket‐triggered lightning and upward lightning from tall towers. These events are often referred as ICC pulses or initial stage (IS) pulses [Wang et al., 1999; Miki et al., 2005; Flache et al., 2008]. A statistical comparison between the ICC pulses and the M component pulses superimposed on continuing currents following return strokes in triggered lightning studied by Wang et al. [1999] indicates that both types of pulses are due to similar physical processes. Miki et al. [2005] compared the initial stage in lightning initiated from tall objects and in rocket triggered lightning, and found that the characteristics of ICC pulses in high tower‐initiated lightning are similar within a factor of 2 but differ more significantly from their counterparts in rocket‐triggered lightning. Specifically, the ICC pulses in object‐initiated lightning exhibit larger peaks, shorter risetime, and shorter half‐peak widths than the ICC pulses in rocket‐triggered lightning.

D10102

[7] Both ICC pulses and M components were recorded in a triggered negative flash during 2009 Shandong artificially triggering lightning experiment by means of rocket‐and‐ wire technique. The flash, named 0902, contained a total of 31 distinct current pulses with a high ratio of large M components. The peak, duration, risetime, and half‐peak width, neutralized charge, and action integral determined from direct current measurements are used to characterize these pulses, and different categories were classified based on these characters. The triggered flash 0902 presented in this paper give the opportunity to make the comparison of different kinds of pulses more quantitatively, and some new insights on the M components and the background luminosity preceding return strokes are achieved.

2. Experiment and Data Description [8] The flash analyzed here was documented during the Shandong Artificially Triggering Lightning Experiment (SHATLE) in 2009. The SHATLE experiment, aimed to understand the close electromagnetic environment of lightning channel and its correlation to the discharge current, and the lightning effects on the ground objectives, has been carried out continuously since the summer of 2005 [e.g., Qie et al., 2007; Zhang et al., 2009b; Yang et al., 2008]. It was moved eastward to the dam of a water reservoir with an area of about 1 km × 1 km in 2009. Figure 1 shows the installation of the experiment during SHATLE 2009. The launching facilities were located 70 m away from the control room (Figure 1, top). The rockets were ignited by an optical fiber ignition control system which was designed to have a capability of launching up to 8 rockets during the same thunderstorm event. [9] The installation in 2009 was basically similar to the previous experiment [Qie et al., 2009], and two main sites were installed at 60 m and 450 m from the launching pad. The current of the flash was measured at the bottom of the 5 m high triggering facility by using a current monitor with a bandwidth of 0.9 Hz to 1.5 MHz and a 0.5 mW shunt with bandwidth of 0–3.2 MHz. The current signals were transmitted via an ISOBE5600 fiber‐optic link system, which has a bandwidth of DC to 20 MHz, to a DL750 digitizing oscilloscope. The ISOBE5600 system is available with individual and single channel transmitter units working with a receiver that supports up to four transmitters. Four transmitter units and one receiver were used in the summer of 2009 [Yang et al., 2010]. The signals were digitized at a 0.1 ms sampling interval and a recording length of 1 s. The current from the 0.5 mW shunt was used in the paper. The lowest measurable current of the whole system including the shunt, optical fiber and oscilloscope is 53 A. Optical images of the triggered lightning were detected with Photron FASTCAM SA1 camera, which was operated at 6000 frames per second (time resolution is 167 ms) with a corresponding spatial resolution of 640 × 1024 pixels. The focal length of the lens is 16 mm. The high‐speed camera was located 460 m away from the rocket launcher. The electric fields were measured with two flat plate antennas that had an area of 0.25 m2 each, one antenna was placed virtually flush with the ground and the other was elevated. For the antenna that was flush with the ground a theoretically determined field calibration factor is used, while the elevated antenna was

2 of 11

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

D10102

briefly discussed the luminosity and current characteristics of 26 M components contained in this flash with treating ICC pulses as M components. This paper will give an elaborate comparison of all the distinct current pulses by incorporating the data of channel base current, luminosity evolution, and electric field change at 30 m from the launcher.

3. Result and Discussion 3.1. The Pulses in the Current, Electric Field Waveform, and Luminosity Evolution [11] Figure 2 shows one of the high‐speed video frames of the triggered flash 0902 taken at a distance of 460 m from the launcher. The lower relative straight and bright section follows the trace of triggered wire vaporized by the upward leader current, and the upper tortuous channel corresponds to natural discharge channel. It is inferred from the height of the vaporized wire that the triggering height was about 340 m. [12] Figures 3a and 3b are the whole waveform of current and relative luminosity inferred from each high‐speed video frame of the discharge channel at 3 altitudes as marked in Figure 2, A and B located at the natural channel, and C located at the middle of the wire‐vaporized channel. The waveforms of relative luminosity at 3 altitudes are basically in agreement with each other, except that the luminosity at wire‐vaporized channel (point C) is larger than that at the natural channel (points A and B) because of the effect of metal ions from

Figure 1. Installation of lightning‐triggering site during SHATLE2009. (top) The building on the right side is the launching control room at 70 m from the launcher. (bottom) The main site at 450 m from the launcher where the high‐ speed camera and recording systems of electromagnetic field were installed. experimentally calibrated relative to it. The electric fields flushed with the ground at 30 m from the launcher were recorded at DL750 simultaneously with the current. Specially designed new‐type rockets were used in 2009, which is made of composite material and assembled with parachute [Qie et al., 2010]. [10] A negative lightning flash 0902 was triggered with classical triggering techniques on 5 August 2009. The flash consisted of an initial stage (upward positive leader, explosion of the triggering wire, its replacement with a plasma channel, and subsequent steady current flow) and following four leader–return stroke sequences, with each stroke attaching to the top of the 5 m lightning rod. The current at the bottom of the channel, the high‐speed optical evolution of the discharge channel with 167 ms time resolution, and the vertical electric fields at 30 m from the triggering facility (rocket launcher) were obtained simultaneously. A total of 31 distinct current pulses correlated with significant enhancement in luminosity (except the final 3 which were out of the camera memory) were documented, and Wang et al. [2010] once

Figure 2. One of the frames of triggered lightning 0902 captured by a high‐speed video camera installed at 460 m from the launcher. The relatively straight and bright section at the bottom follows the trace of triggered wire vaporized by the upward leader current, and the tortuous channel at the top corresponds to natural discharge channel.

3 of 11

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Figure 3. The evolution of channel base current and relative luminosity of discharge channel at three altitudes, A, B, and C, as marked in Figure 2. (a) Current and (b) luminosity. Curves “A” and “B” correspond to the natural channel, and curve “C” corresponds to the wire channel. wire vaporization. The luminosity pulses well correspond to the current pulses, however, the fourth return stroke was not recorded by the high‐speed camera because the camera was prematurely triggered and the record memory was limited. Eleven large pulses are found on the current waveforms with peak current reaching several kiloamperes. From the time enlarged waveforms, it can be recognized that these large pulses correspond to 4 return strokes, 1 large ICC pulse, 5 large M components with peak reaching several kiloamperes that are much larger than the typical ones, and one pulse exhibits both return stroke and M component features (here named as stroke–M component event). The 4 return strokes are marked by R1, ……, R4. The interstroke interval between the two adjacent strokes is 90.1 ms, 58.1 ms and 70.2 ms. The peak current Ip for 4 strokes is 12.1 kA, 16.3 kA, 12.1 kA and 14.0 kA. It can be seen that the second stroke has the largest peak current, and the final one is the second. Since all the 4 strokes exhibit essentially the same waveform, the current and the electric field are presented only for one stroke with enlarged time scale in Figure 4. The risetime from 10% to 90% peak t10–90% is 0.9 ms, 1.4 ms, 0.9 ms, and 1.3 ms, the risetime t30–90% from 30% to 90% peak is 0.6 ms, 1.2 ms, 0.5 ms, and 1.1 ms, and the half peak width tHPW is 20.7 ms, 30.8 ms, 29.6 ms and 34.6 ms, for the 4 strokes. [13] There are 5 large pulses that are associated with M components occurring during the continuing current stage following the return strokes, while the peak current and luminosity are comparable to the return strokes. The M components are marked by M1, M2……, M5 in Figure 3. All

D10102

these 5 large M components exhibit essentially the same behavior of single peak, although the risetime exhibits a wide distribution from 12.4 ms to 71.7 ms. Figure 5 shows two examples of M components, M2 about 20.2 ms before R1, and M3 about 9.2 ms after R1. The peak current for M2 and M3 is 4.3 kA and 5.6 kA, the risetime from 10% to 90% peak is 56 ms and 12 ms, and the half peak width is 104 ms and 58 ms, respectively. It is evident from Figures 4 and 5 that the current and electric field of M components have a very different character from return stroke. The return stroke current pulse exhibits a relatively fast rise to peak, while both the electric field and current of M event exhibit a V‐shaped characterization, similar to the previous results [e.g., Fisher et al., 1993; Rakov et al., 1995; Thottappillil et al., 1995]. However, the amplitudes of peak current for these 5 large M components are 1 order of magnitude larger than the typical ones. Thottappillil et al. [1995] presented a geometric mean (GM) of just 117 A for 124 M components in 14 triggered lightning flashes. [14] It is interesting to note that RM exhibits different behavior. The RM current waveform exhibits two peaks with the secondary peak being larger and about 27.2 ms later than the first. Figure 6 shows the simultaneously recorded channel base current and electric field changes of RM at 30 m from the launcher. The risetime from 10 to 90% of the first peak is similar to that of return stroke, but after the first peak, the waveform is more like M component behavior. The peak current of this event is 3.4 kA, 2.7 kA for the first peak, and 1.4 kA for the second peak (the referring current level is 2.0 kA when the second peak starts). The risetime from 10% to 90% of the first peak is 5.0 ms and from 10% to 90% of the total peak is 27.8 ms. A similar event was reported by Rakov et al. [1996]. Because it is unusual and exhibits both return stroke and M component features, we treat it separately as stroke–M component event, and marked as RM in Figure 3.

Figure 4. The time‐expanded current waveform and electric field at 30 m from the launcher for the first return stroke R1 in flash 0902. The black dashed lines indicate the times of base current starting, the peak electric field, and the peak current. Note that the time of base current starting and the time of peak electric field are basically the same (the shift time is less than about 0.3 ms), so only two dashed lines can be seen.

4 of 11

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Figure 5. Same as Figure 4 but for two large M components. (a) M2 and (b) M3. The possible reason for this feature will be discussed in section 3.2. [15] In addition to these 11 large pulses, other 20 pulses with peak current of several tens to several hundreds of Amperes are found based on the current and the relative luminosity records. Among them, 18 are typical M components occurred after return strokes, while other 2 are typical ICC pulses in the initial stage. The time intervals between each adjacent typical M pulses range from 0.7 ms to 56.4 ms. The waveshapes of these M components are quite similar to the 5 large M components with simple and single peak. The one distinct difference is their intensity, characterized by low current, low electric field and relatively weak luminosity. Because they are similar to the large M components, the waveform will not be presented here. The comparison of these parameters with large M components will be discussed in section 3.2. [16] By analyzing the simultaneous observation of the channel base current and electric field at 30 m from the launcher, it is found that the features of the 5 large M components are also basically in good agreement with the observation given by Rakov et al. [1995] and support their “two wave” theory. The electric field produced by M com-

D10102

ponent at 30 m from the lightning channel is initially negative going (removal of negative charge overhead produces a negative field change, for easy comparison with the current waveshape). The field change begins when there is no obvious current at the bottom of the channel, consistent with the view of the M component as a leader like process transporting negative charge from a cloud charge source toward the ground. The M current on the ground emerges from the background negative continuing current when the corresponding electric field at 30 m from the launcher is still negative going, implying that initially the negative charge is removed by the current from the lower sections of the channel, while at the same time a larger negative charge is transported downward along the higher channel sections causing the continuing negative field change. Another remarkable feature is that the M current peak lags behind the electric field peak by 17–48 ms and the waveshapes of current and electric field pulses are dissimilar. The trailing edge is somewhat slower than the leading edge for the current pulse, whereas the field pulse is more or less symmetrical. All these observation facts support the “two wave” theory of Rakov et al. [1995]. [17] Wang et al. [2007] speculated that an M component is a composition of many waves occurring at different times and different locations, and those waves may interfere with each other complicatedly. This “many waves” theory is based on the impulsive nature in the correlated electric field changes of luminous variation event. Based on our simultaneous records of current and electric field change, we did not find any remarkable impulsive variations. During the M component, it could be possible that many electrical breakdowns occur at different time and locations along the discharge channel and the breakdown‐induced waves propagate along the same channel and superimpose on the two main waves. However, these induced waves could be weak and not distinguishable on our electric field records. 3.2. Comparison Between Different Categories of Pulses 3.2.1. Waveform Characterization [18] Figure 7 shows histogram distributions of current parameters of return strokes and M component pulses,

5 of 11

Figure 6. Same as Figure 4 but for the RM event.

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

D10102

Figure 7. Histogram distributions of current waveform parameters: (a) peak current, (b) risetime, (c) half peak width, (d) duration, (e) charge transfer, and (f) action integral. including their peak current, risetime, half peak width, duration, charge transfer and action integral. The ICC pulses in the initial stage are not included here and will be discussed later. The large M components are presented separately for the comparison of typical M components with previous results, also because they exhibit some special waveform behavior as shown below. The parameter definition is the same as the previous definition by Thottappillil et al. [1995]. The Ip is the difference between the peak of maximum value of the current pulse and the background current level. Risetime t10–90% is the time interval on the wavefront between the 10% and 90% values of the peak value. The half‐peak

width thpw is the time interval between 50% values of the peak on the wavefront and on the falling portion of the event waveform. The duration tw is the time interval from the beginning of the wavefront to the somewhat subjectively selected point at which the trailing edge of the current pulse becomes indistinguishable from the overall continuous current waveform. Furthermore, charge transfer Q is the time integral ofR the current above the background level. The action integral ( i2 dt) time is within the duration tw. Differences between the parameters of typical M components and return strokes are quite distinct, and the large M component is in between. For further comparison, Table 1 shows the GM

6 of 11

D10102

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Table 1. Comparison of Parameters for the Four Categories of Current Pulses Events

Sample

Type of Value

IP (kA)

thpw (ms)

tw (ms)

t10–90% (ms)

Q (C)

Return stroke

4

Stroke M event

1

Large M component

5

Typical M events

18

GM min max first peak (return stroke) second peak (M component) GM min max GM min max

13.5 12.1 16.3 2.7 1.4 5.1 3.8 7.0 0.24 0.10 0.72

28.4 20.7 34.6 99 85 76.3 58 104 400 247 787

0.37 0.33 0.42 0.31a – 0.61 0.43 0.84 1.95 0.60 6.35

1.1 0.9 1.4 5.0 27.8 34.6 12 72 319 137 709

0.76 0.53 0.99 0.32a – 0.55 0.46 0.68 0.11 0.03 0.31

R

i2 dt (A2 s−1) 6100 2443 6878 630a – 1620 978 2641 135 54 400

a

The value of the whole stroke‐M event.

values and ranges of each parameter for the 4 categories of pulses based on the current measurement. [19] As expected, the return stroke is the strongest event among the 4 categories of pulses, and the mean peak current was 13.5 kA. The 10% to 90% risetime and the half peak width are the shortest, with the values of 1.1 ms and 28.4 ms, respectively. The return stroke parameters are in good agreement with previous results [e.g., Fisher et al., 1993; Schoene et al., 2010; Zhao et al., 2009]. The typical M components are the weakest with a peak current GM value of 243 A. The 10% to 90% risetime and the half peak width of the typical M component are the longest, with the values of 319 ms and 400 ms. It is interesting to note that the 5 large M components have significant larger peak current and smaller half‐peak width, duration and risetime compare with the typical M components. The difference could be an order of magnitude. For the charge transfer and action integral, the large M component and return stroke are comparable, although the return stroke is relatively larger than the large M component. The peak current of these large M components are even comparable to some of the weak subsequent return strokes in the previous results [Zhao et al., 2009; Yang et al., 2010]. [20] If we compare the parameters of return stroke, large M components and typical M components, we can find a tendency that the higher the peak current, the shorter the risetime and the width. The action integral are more likely proportional to the peak current not the duration. When all the pulses were analyzed for correlation between the parameters, no significant correlation is found between either the magnitude of current or charge (the two are moderately correlated) and any other parameters listed in Table 1, even only the M components were considered. [21] The results for typical M events in our case are basically consistent with the results of Thottappillil et al. [1995] for M components that occur during the continuing current

stage following return stroke and by Wang et al. [1999] and Miki et al. [2005] from the ICC pulses in Florida. Table 2 shows a comparison of statistic parameters between different authors in triggered lightning. The peak current, from 10 to 90% peak risetime, half‐peak width and charge transfer of M components are 117 A, 422 ms, 0.8 ms and 129 mC based on Thottappillil et al. [1995] and 144 A, 528 ms, 1.0 ms and 143 mC based on Wang et al. [1999], respectively. Although the results of typical M components are basically in good agreement with Wang et al. [1999] and Thottappillil et al. [1995], the M peak current will be more than 3 times larger than theirs if the larger M components are incorporated together. [22] Five large M components, the peak current of which is an order of magnitude greater than the typical ones, were recorded in one flash. The occurrence ratio of large M component to the typical one was 1:3.6. The M components with such large peak current are rare reported, though Rakov et al. [1998], Rakov [2001], and Miki et al. [2002] showed a few cases. Miki et al. [2002] once reported 2 large M components whose current peaks were between 2.3 and 3.2 kA. Flache et al. [2008] reported M peak current in upward natural lightning from tall towers ranged from 3 to 11 kA with a GM of 6.6 kA for 4 “fast” M component pulses. The range of variation of peak current for 7 “fast” IS pulses was from 3.6 kA to 13 kA with a GM value of 6.5 kA. In fact, the flash 0902 was not the only case with such kind of large M components in SHATLE data set, but usually just one or two occurred after the return stroke in a flash. 3.2.2. The Comparison Between ICC Pulses in the Initial Stage and M Components Following Return Strokes [23] The triggered flash 0902 contained both ICC pulses and M pulses that occurred during the continuing current following return stroke, providing an opportunity of their quantitative comparison. Three ICC pulses are found in the

Table 2. The Comparison of M Component Parameters in Triggered Lightning 0902 and Those Found in Other Literature (M and ICC Pulses)

This work This work This work Thottappillil et al. [1995] Wang et al. [1999] Miki et al. [2005]

Type of Pulse

IM (A)

thpw (ms)

tw (ms)

typical M large M all M components M components (Florida, Alabama) ICC pulses (Florida, Alabama) ICC pulses (ICLRT)

243 (n = 18) 5140 (n = 5) 472 (n = 23) 117 (n = 124) 144 (n = 230) 113 (n = 296)

0.40 (n = 18) 0.08 (n = 5) 0.28 (n = 23) 0.82 (n = 113) 1.0 (n = 175) 943 (n = 247)

1.95 (n = 18) 0.61 (n = 5) 1.52 (n = 23) 2.1 (n = 114) 2.5 (n = 179) 2.59 (n = 254)

7 of 11

Q (mC) 105 553 151 129 143 –

(n (n (n (n (n

= = = = =

18) 5) 23) 104) 175)

t10–90% (ms) 319 (n = 18) 34.6 (n = 5) 197 (n = 23) 422 (n = 124) 528 (n = 200) 464 (n = 267)

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Figure 8. Same as Figure 4 but for the large ICC pulse (ICCP). initial stage of flash 0902, including one large and two typical ICC pulses. Figure 8 shows the simultaneous records of current and electric field for the large ICC pulse (marked as ICCP in Figure 3). The peak current of this large ICC pulse is 6.7 kA. The 10–90% peak risetime and the half peak width are 48 ms and 91 ms, respectively. It can be seen that the current and the simultaneous electric field waveform of this large ICC pulse is quite similar to the waveforms of large M components. The electric field starts and reaches its peak earlier than the channel base current, and when the current starts to change, the electric field is still negative going. The similarity of the large ICC pulse and large M components indicates that they are due to similar physical processes. However, it should be noted that ICC pulses happen during the ICC stage of triggered lightning, and M components during the usual continuing current that happens after the return stroke. The ICC pulses are absent in the normal natural downward lightning. Also, the origin of ICC pulses and the traditional M components inside the cloud could be different because they happen at two different stages of a flash. [24] Jiang et al. [2011] conducted a simulation on the large ICC pulse and M components based on the “two wave”

D10102

theory proposed by Rakov et al. [1995], the simulated result is basically in good agreement with the observed electric field, although an unexpected overshooting occurs in the recovery edge. This further confirms the similarity in the mechanism of ICC pulse and M component. Except the large ICC pulse, there are also 2 typical ICC pulses in the IS stage, the parameters of them are consistent with the statistic result of typical M components. This is basically in agreement with the results by Wang et al. [1999]. 3.2.3. The Luminosity and Current Immediately Prior to the Pulse Occurrence [25] Most of the current pulses are simultaneously recorded by the high‐speed video camera, except the final three current pulses including the fourth stroke and the final 2 typical M component because of the memory limitation of high‐speed camera. When we checked the measured current and optical records, we found that all of the optical pulses were superimposed on slowly varying luminosity intensity, even for return strokes. Figure 9 shows the video images immediately preceding the 10 large events, including 1 large ICC pulse (ICCP), 1 RM event, first 3 return strokes, and 5 large M components in time sequence. These pictures are composed from original frame images. The pictures show clearly the apparent luminosity of natural discharge channel immediately preceding each of the 10 large events, even preceding the return strokes although their luminosity is weaker than that before the M components. [26] The channel luminosity could imply a possibility of current flowing in the channel preceding these large pulses. After checking the current waveform, it is found that the current immediately preceding these large pulses were submerged into background noise of 53A. To identify if there is current flowing through the channel preceding the return strokes, a frequency domain low‐pass filter was used for the original current data. After the high‐frequency noise and DC offset were filtered from the current data, the background continuing current just prior to the typical M components, large M components and large ICC pulse is recognized. Table 3 shows the luminosity and the corresponding currents preceding each category of current pulses inferred from luminosity and after the filtration of high‐frequency noise and DC offset. It is found that nearly all of the typical M event pulses are superimposed on a slowly varying continuing

Figure 9. The video images of discharge channel, including the bottom wire‐related channel and higher natural discharge channel, immediately prior to 1 large ICC pulse (ICCP), 1 RM event, 5 large M components, and 3 return strokes in time sequence. 8 of 11

D10102

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Table 3. Luminosity and Corresponding Currents (After High‐Frequency and DC Offset Filtered) Immediately Preceding Each Category of Current Pulsesa Events

Sample (Lcc/I)

Return stroke

3/4

Stroke M event

1/1

Large M component

5/5

Typical M component

16/18

Type of Value

IP (kA)

Background Lcc

Background Icc After Filter (A)

GM min max first peak second peak GM min max GM min max

13.5 12.1 16.3 2.7 3.4 5.1 3.8 7.0 0.22 0.06 0.72

91 53 169 293 – 552 323 858 995 156 7183

2.0 1.7 2.2 5.4 – 43 19 87 100 8 469

a Lcc and Icc are the background luminosity value and continuing current value immediately preceding the M current pulse and return stroke. The GM value of Lcc is the average relative luminosities obtained from four points between A and B as marked in Figure 2.

current with the value ranged from about several tens to several hundreds Amperes, similar to the previous results by Fisher et al. [1993] and Thottappillil et al. [1995]. For the large M components, the background continuing current ranged from 19 A to 87 A with a GM value of 44 A. A very small current with a GM value of 2.0 A is found preceding the return strokes, and 5.4 A preceding RM. These small values are consistent with 2 A noise level in the work by Fisher et al. [1993] and indicate that there is no significant current in the channel or the current is at a very low level just prior to the return strokes. The currents dropping below 1 A and 0.1 A prior to return strokes were reported by Berger [1967] and McCann [1944], respectively. On the other hand, Hubert et al. [1984] reported that they observed optically, although rarely, leader–return stroke sequences during the continuing current stage in triggered flashes in New Mexico. [27] Diendorfer et al. [2003] once analyzed the brightness and the current of the initial continuous current in upward initiated flashes to the Gaisberg tower. They found an excellent linear correlation (r2 = 0.96) between brightness and current in the range of 10 to 250 A. Flache et al. [2008] also used channel luminosity as a proxy for current in determining which channel or branch carried the current at a certain time when they studied the initial stage pulses in upward lightning. From Table 3, we can see that Lcc and Icc are basically of the same tendency. Though it may not be quite rigorous, if we make a linear hypothesis to the relationship of luminosity and current data for the current smaller than 500 A, the background Lcc indicates that the current prior to the return strokes is about 10 A. The inferred current result provides another possibility of the channel current, for the background Lcc is the original value of the channel relative luminosity. Of course, we tend to deem that the directly measured current is more credible (though being filtered). 3.3. Discussion [28] From above analysis it seems that there are some unusual features of this triggered lightning flash. First, the RM exhibits both return stroke and M component features. Second, there is no obvious luminosity cutoff time prior to the return strokes, and the third, high ratio of the large M components. [29] The shift time of two peaks of RM is 27.8 ms, and the rising edge of the second peak starts 7.5 ms after the first peak (or the return stroke peak), much less than an ordinary time

of a return stroke current trailing from peak to a relatively steady continuing current. Based only on the current waveform, this event can be judged as a return stroke immediately followed by an M component. But some other features can be identified when checking the electric field change. As can be confirmed from Figures 4 and 5, the rising edge of the electric field is caused by the dart leader for the leader‐stroke sequence or the M component incident wave for the M event. If the RM event is an ordinary return stroke and followed by an M component, then the electric field should logically present the dart leader feature. But the risetime of electric field change of this event (as shown in Figure 6) is much longer than that of the downward dart leader (as shown in Figure 4). This indicates that the dart leader and the M component incident wave may be superimposed on each other in the channel. As also show in Figures 4 and 5, the ratio of the peak magnitude of the electric field pulse to the current pulse for return stroke and large M component are different, about 6–7 for return stroke and 2–3 for large M component, note that the unit is “kV/m” for the electric field is and “kA” for the current. As shown in Figure 6, this ratio for the return stroke part of RM is 10.7 (and for the M component part is 8.5), approximately the sum of both values for return stroke and M component. [30] How is the RM event formed? and why does the discharge channel show continuing luminous prior to the return strokes? After we checked the high‐speed images carefully we found that an independent faint branch appears on some of the video images, which is a hint of branches in the upper part of the channel. It is possible that two branches with common lower portions exist simultaneously in the upper part of the discharge channel. As a speculation, Figure 10 shows a schematic diagram for this possible discharge channel. The branched point could be out of the view of the video camera. Under this speculation, the unusual result can be explained. The continuing current flows in one branch while a leader– return stroke sequence propagates along another branch, producing a superposition of the continuing luminosity and leader–return stroke images in the time‐resolved optical records and channel base current records. [31] For the RM event, quite possibly the return stroke and M component are initiated from two different upper branches. Almost simultaneous occurrence of this two events leads to the superposition of the dart leader and M component incident wave in the common lower channel. Actually,

9 of 11

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Figure 10. A schematic diagram for the possible discharge channel of RM event. The branched point could be out of the view of the video camera. A, B, and C correspond to the three points in Figure 2. similar lightning flash of different upper braches was reported by Flache et al. [2008]. Based on the current measurement and high‐speed video images, they found that the IS current pulses with shorter risetime each developed in a newly illuminated branch, and IS pulses with longer risetime occurred in already luminous (current‐carrying) channels when they studied the initial stage pulses in upward lightning. Fisher et al. [1993] also suggested that the continuing current preceding the leader–return stroke sequences in New Mexico triggered flashes could be associated with the upper branches of the discharge channel. 3.4. Conclusions [32] Thirty one pronounced pulses superimposed on current waveforms during one negative triggered lightning flash have been studied, including return stroke pulses, ICC pulses in the initial stage and M components in continuing current stage following return stroke. Based on the waveshapes of current, luminosity from high‐speed video images, and electric field at 30 m from the channel, the 31 distinct current pulses are discriminated as 4 return strokes, 18 typical M components, 5 large M components with a single peak current in range of kiloamperes, 1 RM event which exhibits both stroke and M component behavior, and 3 ICC pulses. The GM value of 4 return stroke peak current is 13.5 kA, the half peak width is 28.4 ms, and the risetime from 10% to 90% peak is 1.1 ms, while the corresponding values for 18 typical M components are 243 A, 400 ms and 319 ms, and the values are 5.1 kA, 76.3 ms and 34.6 ms, respectively, for the 5 large M components. The only stroke‐M event exhibits two peaks with the first being the return stroke mode and secondary

D10102

being the M mode. The return stroke parameters and typical M parameters are all basically in good agreement with the previous results. [33] There are 5 larger‐than‐usual M components observed in the flash, the peak current of which is in a range of several kiloampere and comparable to the smaller return strokes. The occurrence ratio of large M component to the typical one is 1:3.6 in flash 0902. This is significant because the large M components transport charge similar to the return stroke, and has severe thermal effect to the nearby ground objectives. The large M components have been rarely reported quantitatively in the previous statistic studies. The result in this paper should be good supplementary data for the M event. The ICC pulses exhibit similar features to the M components. The waveforms of current and simultaneous electric field of the large ICC pulse are basically the same as the large M components. This indicates that they have similar physical processes. [34] All of the pulse events of flash 0902 are superimposed on slowly varying luminosity intensity, and there is no obvious luminosity cutoff time prior to the return strokes. The luminosity of the channel implies that there may be a current flowing in the channel preceding all the pulses, even for return strokes. The current result after filtering the high‐ frequency noise and DC offset indicates that there may be current flowing in the channel prior to the return strokes but not significant, or the current is at a very low level. The unusual simultaneous electric field and current waveform of RM implies a superposition of dart leader and M incident wave in the channel, suggesting the possibility of coexistence of two branches in the upper part of the discharge channel. [35] Acknowledgments. The research was supported by the Knowledge Innovation Program of the Chinese Academy of Sciences (grant KZCX2‐YW‐206), the National Natural Science Foundation of China (grants 40774083 and 40930949), and the One Hundred Person Project of the Chinese Academy of Sciences. The authors would like to thank all the staff involved in the Shandong Artificially Triggering Lightning Experiment. They also thank the two anonymous reviewers for their valuable suggestions which improved the quality of this paper.

References Berger, K. (1967), Novel observations on lightning discharges: Results of research on Mount San Salvation, J. Franklin Inst., 283, 478–525, doi:10.1016/0016-0032(67)90598-4. Campos, L. Z. S., M. M. F. Saba, O. Pinto Jr., and M. G. Ballarotti (2007), Waveshapes of continuing currents and properties of M‐components in natural negative cloud‐to‐ground lightning from high‐speed video observations, Atmos. Res., 84, 302–310, doi:10.1016/j.atmosres.2006.09.002. Diendorfer, G., M. Viehberger, M. Mair, and W. Schulz (2003), An attempt to determine currents in lightning channels branches from optical data of a high speed video system, paper presented at International Conference on Lightning and Static Electricity, Natl. Weather Assoc., Blackpool, U. K. Fisher, R. J., G. H. Schnetzer, R. Thottappillil, V. A. Rakov, M. A. Uman, and J. D. Goldberg (1993), Parameters of triggered‐lightning flashes in Florida and Alabama, J. Geophys. Res., 98, 22,887–22,902, doi:10.1029/93JD02293. Flache, D., V. A. Rakov, F. Heidler, W. Zischank, and R. Thottappillil (2008), Initial‐stage pulses in upward lightning: Leader/return stroke versus M‐component mode of charge transfer to ground, Geophys. Res. Lett., 35, L13812, doi:10.1029/2008GL034148. Hubert, P., P. Laroche, A. Eybert‐Berard, and L. Barret (1984), Triggered lightning in New Mexico, J. Geophys. Res., 89, 2511–2521, doi:10.1029/ JD089iD02p02511. Jiang, R., X. Qie, C. Wang, J. Yang, Q. Zhang, J. Wang, and D. Liu (2011), Lightning M‐components with peak current in the range of kilo amperes and their mechanism [in Chinese], Acta Phys. Sin., in press.

10 of 11

D10102

QIE ET AL.: ARTIFICIALLY TRIGGERED LIGHTNING

Jordan, D. M., V. P. Idone, V. A. Rakov, M. A. Uman, W. H. Beasley, and H. Jurenka (1992), Observed dart leader speed in natural and triggered lightning, J. Geophys. Res., 97, 9951–9957. Jordan, D. M., V. P. Idone, R. E. Orville, V. A. Rakov, and M. A. Uman (1995), Luminosity characteristics of lightning M components, J. Geophys. Res., 100, 25,695–25,700, doi:10.1029/95JD01362. Le Vine, D. M., J. C. Willett, and J. C. Bailey (1989), Comparison of fast electric field changes from subsequent return strokes of natural and triggered lightning, J. Geophys. Res., 94, 13,259–13,265, doi:10.1029/ JD094iD11p13259. Malan, D. J., and H. Collens (1937), Progressive lightning. III. The fine structure of return lightning strokes, Proc. R. Soc. London, Ser. A, 162, 175–203, doi:10.1098/rspa.1937.0175. McCann, G. D. (1944), The measurement of lightning currents in direct strokes, Trans. Am. Inst. Electr. Eng., 63, 1157–1164, doi:10.1109/ T-AIEE.1944.5058859. Miki, M., V. A. Rakov, K. J. Rambo, G. H. Schnetzer, and M. A. Uman (2002), Electric fields near triggered lightning channels measured with Pockels sensors, J. Geophys. Res., 107(D16), 4277, doi:10.1029/ 2001JD001087. Miki, M., V. A. Rakov, T. Shindo, G. Diendorfer, M. Mair, F. Heidler, W. Zischank, M. A. Uman, R. Thottappillil, and D. Wang (2005), Initial stage in lightning initiated from tall objects and in rocket‐triggered lightning, J. Geophys. Res., 110, D02109, doi:10.1029/2003JD004474. Qie, X., Q. Zhang, Y. Zhou, G. Feng, T. Zhang, J. Yang, X. Kong, Q. Xiao, and S. Wu (2007), Artificially triggered lightning and its characteristic discharge parameters in two severe thunderstorms, Sci. China, Ser. D Earth Sci., 50(8), 1241–1250, doi:10.1007/s11430-007-0064-2. Qie, X., et al. (2009), Characteristics of triggered lightning during Shandong artificial triggering lightning experiment (SHATLE), Atmos. Res., 91, 310–315, doi:10.1016/j.atmosres.2008.08.007. Qie, X., J. Yang, R. Jiang, J. Wang, D. Liu, C. Wang, and Y. Xuan (2010), A new‐model rocket for artificially triggering lightning and its first triggering lightning experiment [in Chinese], Chin. J. Atmos. Sci., 34(5), 937–946. Rakov, V. A. (2001), Surges superimposed on continuing currents in lightning discharges, paper presented at International Conference on Lightning and Static Electricity, Soc. of Automotive Eng., Seattle, Wash. Rakov, V. A., and M. A. Uman (2003), Lightning: Physics and Effects, 687 pp., Cambridge Univ. Press, New York. Rakov, V. A., R. Thottappillil, and M. A. Uman (1992), Electric field pulses in K and M changes of lightning ground flashes, J. Geophys. Res., 97, 9935–9950. Rakov, V. A., R. Thottappillil, M. A. Uman, and P. P. Barker (1995), Mechanism of the lighting M component, J. Geophys. Res., 100, 25,701–25,710, doi:10.1029/95JD01924. Rakov, V. A., M. A. Uman, M. I. Fernandez, R. Thottappillil, A. Eybert‐ Berard, J. P. Berlandis, F. Rachidi, M. Rubinstein, S. Guerrieri, and C. A. Nucci (1996), Observed electromagnetic environment close to the lightning channel, paper presented at 23rd International Conference on Lightning Protection, Comm. of Int. Conf. on Lightning Prot., Florence, Italy. Rakov, V. A., et al. (1998), New insights into lightning processes gained from triggered‐lightning experiments in Florida and Alabama, J. Geophys. Res., 103(D12), 14,117–14,130, doi:10.1029/97JD02149. Rakov, V. A., D. E. Crawford, K. J. Rambo, G. H. Schnetzer, M. A. Uman, and R. Thottappillil (2001), M‐component mode of charge transfer to ground in lightning discharges, J. Geophys. Res., 106, 22,817–22,831, doi:10.1029/2000JD000243.

D10102

Rakov, V. A., M. A. Uman, and K. J. Rambo (2005), A review of ten years of triggered‐lightning experiments at Camp Blanding, Florida, Atmos. Res., 76, 503–517, doi:10.1016/j.atmosres.2004.11.028. Schoene, J., M. A. Uman, and V. A. Rakov (2010), Return stroke peak current versus charge transfer in rocket triggered lightning, J. Geophys. Res., 115, D12107, doi:10.1029/2009JD013066. Shao, X. M., P. R. Krehbiel, R. J. Thomas, and W. Rison (1995), Radio interferometric observations of cloud‐to‐ground lightning phenomena in Florida, J. Geophys. Res., 100, 2749–2783, doi:10.1029/94JD01943. Thottappillil, R., V. A. Rakov, and M. A. Uman (1990), K and M changes in close lightning ground flashes in Florida, J. Geophys. Res., 95, 18,631–18,640, doi:10.1029/JD095iD11p18631. Thottappillil, R., J. D. Goldberg, V. A. Rakov, M. A. Uman, R. J. Fisher, and G. H. Schnetzer (1995), Properties of M components from currents measured at triggered lightning channel base, J. Geophys. Res., 100, 25,711–25,720, doi:10.1029/95JD02734. Uman, M. A., V. A. Rakov, K. J. Rambo, T. W. Vaught, M. I. Fernandez, D. J. Cordier, R. M. Chandler, R. Bernstein, and C. Golden (1997), Triggered‐lightning experiments at Camp Blanding, Florida (1993– 1995), Trans. IEE Jpn., 117‐B, 446–452. Wang, D., V. A. Rakov, M. A. Uman, M. I. Fernandez, K. J. Rambo, G. H. Schnetzer, and R. J. Fisher (1999), Characteristics of the initial stage of negative rocket‐triggered lightning, J. Geophys. Res., 104, 4213–4222, doi:10.1029/1998JD200087. Wang, D., N. Takagi, and T. Watanabe (2007), Observed characteristics of luminous variation events during the initial state of upward positive leaders, J. Atmos. Electr., 7(1), 61–68. Wang, C., X. Qie, J. Yang, R. Jiang, J. Wang, D. Liu, M. Yang, L. Pian, and H. Yang (2010), The luminosity and current characteristics of M‐component from triggered‐lightning in Shandong, paper presented at Asia‐Pacific Symposium on Electromagnetic Compatibility, Tsinghua Univ., Beijing. Yang, J., X. Qie, G. Zhang, and H. Wang (2008), Magnetic field measuring system and current retrieval in artificially triggering lightning experiment, Radio Sci., 43, RS2011, doi:10.1029/2007RS003753. Yang, J., X. Qie, G. Zhang, Q. Zhang, G. Feng, Y. Zhao, and R. Jiang (2010), Characteristics of channel base currents and close magnetic fields in triggered flashes in SHATLE, J. Geophys. Res., 115, D23102, doi:10.1029/2010JD014420. Zhang, Q., X. Qie, Y. Zhao, X. Kong, and J. Yang (2009a), Characteristics and simulation of lightning current waveforms during one artificially triggered lightning, Atmos. Res., 91(2–4), 387–392, doi:10.1016/j. atmosres.2008.04.015. Zhang, Q., X. Qie, Z. Wang, T. Zhang, and J. Yang (2009b), Simultaneous observation on electric field changes at 60 m and 550 m from altitude‐ triggered lightning flashes, Radio Sci., 44, RS1011, doi:10.1029/ 2008RS003866. Zhao, Y., X. Qie, X. Kong, G. Zhang, T. Zhang, J. Yang, G. Feng, Q. Zhang, and D. Wang (2009), Analysis on the parameters of the current waveforms of triggered lightning [in Chinese], Acta Phys. Sin., 58(9), 6616–6626. R. Jiang, D. Liu, X. Qie, C. Wang, J. Wang, and J. Yang, Key Laboratory of Middle Atmosphere and Global Environment Observation, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China. ([email protected])

11 of 11