Characteristics and generation of secondary jets and secondary gigantic jets

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 117,, doi:10.1029/2011ja017443, 2012 Characteristics and generation of secondary jets and secondary gigantic jets Li-Jou Lee, 1 Sung-Ming Huang, 1 Jung-Kung Chou, 1 Cheng-Ling Kuo, 1 Alfred B. Chen, 2 Han-Tzong Su, 1 Rue-Rou Hsu, 1 Harald U. Frey, 3 Yukihiro Takahashi, 4 and Lou-Chuang Lee 5 Received 8 December 2011; revised 16 April 2012; accepted 14 May 2012; published 27 June 2012. [1] Secondary transient luminous events (TLEs) recorded by the ISUAL-FORMOSAT2 mission can either be secondary jets or secondary gigantic jets (GJs), depending on their terminal altitudes. The secondary jets emerge from the cloud top beneath the preceding sprites and extend upward to the base of the sprites at 50 km. The secondary jets likely are negative electric discharges with vertically straight luminous columns, morphologically resembling the trailing jet of the type-i GJs. The number of luminous columns in a secondary jet seems to be affected by the size of the effective capacitor plate formed near the base of the preceding sprites and the charge distribution left behind by the sprite-inducing positive cloud-to-ground discharges. The secondary GJs originate from the cloud top under the shielding area of the preceding sprites, and develop upward to reach the lower ionosphere at 90 km. The observed morphology of the secondary GJs can either be the curvy shifted secondary GJs extending outside the region occupied by the preceding sprites or the straight pop-through secondary GJs developing through the center of the preceding circular sprites. A key factor in determining the terminal height of the secondary TLEs appears to be the local ionosphere boundary height that established by the preceding sprites. The abundance and the distribution of the negative charge in the thundercloud following the sprite-inducing positive cloud-to-ground discharges may play important role in the generation of the secondary TLEs. Citation: Lee, L.-J., S.-M. Huang, J.-K. Chou, C.-L. Kuo, A. B. Chen, H.-T. Su, R.-R. Hsu, H. U. Frey, Y. Takahashi, and L.-C. Lee (2012), Characteristics and generation of secondary jets and secondary gigantic jets, J. Geophys. Res., 117,, doi:10.1029/2011ja017443. 1. Introduction [2] Both sprites and jets are members of the transient luminous events (TLEs), while their generating mechanisms and luminous morphologies differ. Sprite that occurs between the thundercloud top and the lower ionosphere is mostly induced by the quasi-static electric field (QE field) establishing by the preceding positive cloud-to-ground (CG) lightning discharges [Pasko et al., 1997]. The sprite luminous altitudes span the range of 40 to 90 km and their horizontal width ranges from 5 to 30 km [Sentman et al., 1995]; while the luminous duration of sprites varies from several to 1 Department of Physics, National Cheng Kung University, Tainan, Taiwan. 2 Institute of Space, Astrophysical and Plasma Sciences, National Cheng Kung University, Tainan, Taiwan. 3 Space Sciences Laboratory, University of California, Berkeley, California, USA. 4 Department of Cosmosciences, Hokkaido University, Sapporo, Japan. 5 Institute of Space Science, National Central University, Jhongli, Taiwan. Corresponding author: R.-R. Hsu, Department of Physics, National Cheng Kung University, Tainan 70701, Taiwan. (rrhsu@phys.ncku.edu.tw) 2012. American Geophysical Union. All Rights Reserved. 0148-0227/12/2011JA017443 tens of milliseconds (ms). Jets are cone-shaped upward discharges from the thundercloud tops; they are sub-categorized into blue jets (BJ) and gigantic jets (GJ) according to their terminal altitudes. Blue jets discharge from cloud top to 40 50 km altitude, with an upward propagating velocity of 100 km/s and a luminous duration of 200 to 300 ms [Wescott et al., 1995; Chou et al., 2010]. The terminal altitudes of GJs are near the lower ionosphere boundary (90 km) and the luminous period often lasts for more than 500 ms [Pasko et al., 2002; Su et al., 2003; Cummer et al., 2009; van der Velde et al., 2010]. [3] Chou el al. [2010] studied and categorized the ISUAL GJs into three types, basing on their image and spectral properties. With the support of the associated ULF data, the type I GJs were identified to be the previously reported negative cloud-to-ionosphere discharges (-CI), which comprise of upward discharging negative streamers. Type II GJs began as BJs and then develop further to reach the lower ionosphere. The available while not conclusive ULF data indicate that the discharge polarity of the type II GJs likely is +CI a new categorization that was also independently reported in van der Velde et al. [2010]. The type III GJs always appear near the preceding lightning. The discharge polarity of the type III GJs can either be +CI or CI, 1of7

depending on the type of the charge imbalance left behind by the preceding lightning. The morphological evolution of the type I GJs is found to consist of three stages: the leading jet, the fully developed jet (FDJ) and the trailing jet (TJ) [Su et al., 2003; Kuo et al., 2009; Chou et al., 2010]. The FDJ of the type I and type II GJs both optically links the cloud top and the lower ionosphere, though the type I GJs have a much higher brightness comparing with those of the type IIs [Chou et al., 2010]. At the trailing jet stage of the type I and type II GJs, the upper edge of the luminous channel drops down to 50 km altitude and then slowly increases to 60 km altitude. [4] The generation mechanisms of jets have been studied extensively since their discoveries [Pasko, 2008, 2010, and references therein]. Most of the jets observed to date are from normal polarity thunderstorms [Krehbiel et al., 2008; Pasko, 2010]. Also blue jets are believed generally to be associated with positive leaders. Pasko and George [2002] noted that, basing on a 3D numerical work, a fast growth of a substantial amount of positive charge near the thundercloud top would be important in the generation of blue jets. Recently, Krehbiel et al. [2008] proposed that the charge imbalance in thunderclouds are the key condition for the propagation of leaders which eventually materialize into the intra/inter-cloud, cloud-to-ground, or upward discharges. In their unified model, blue jets begin as electric breakdowns between the positive upper storm charge and the screening negative charge at the cloud top; while (negative) gigantic jets start as normal intracloud discharges between the midlevel negative charge layer and positive upper charge layer that escape beyond the storm top [Krehbiel et al., 2008]. [5] Previous ground observations have reported that secondary discharges sometimes form under the preceding sprites and then propagate from the cloud top toward the lower edge of the sprites. From their appearances and their relation with the preceding sprites, these secondary discharges have carried the names of embers, trolls, palm-trees [Heavner, 2000; Lyons et al., 2000; Moudry et al., 2001; Moudry, 2003] and the sprite-initiated secondary TLEs [Marshall and Inan, 2007]. Marshall and Inan [2007] postulated that the sprite currents transport positive charge from the ionosphere, down to the bottom of sprites (40 50 km altitude), following a +CG. The +CG leaves an excess of negative charge in the thundercloud. Since the ionization is increased in the electric paths of the preceding sprites, this region then serves as an extension of the local ionosphere and effectively lowers the local ionosphere boundary. In the process, the positive charge relocating to the bottom of the sprite may enhance the electric field at the lower altitudes; thus an electric breakdown may be initiated to produce the observed secondary TLEs. In an extensive review article on blue jets and GJs, Pasko [2008] mentioned that the features of these secondary TLEs actually exhibit some close similarities with those of blue jets and gigantic jets. The lowered boundary discussed here is the same as the concept of the moving capacitor plate, which was previously proposed by Greifinger and Greifinger [1976] and also used in Pasko and George [2002]. [6] Thus far, TLE studies have focused mainly on events with relatively simple morphologies. In this work, morphologically complex events recorded by ISUAL are analyzed with the aim to elucidate the characteristics and the probable generating scenarios of the secondary TLEs that are termed as secondary jets and secondary gigantic jets. The instruments and the data are introduced in Section 2. Characteristics and the generating scenarios for the secondary jets and the secondary GJs are respectively presented Sections 3 and 4. Discussion and summary are given in Sections 5 and 6. 2. Instruments and Data [7] The data analyzed in this work are from the ISUAL experiment and the NCKU ULF sferics recording station, respectively. The ISUAL is the scientific payload onboard the FORMOSAT-2 satellite, which moves along a Sunsynchronized orbit at an 891 km altitude. ISUAL is configured with an eastward limb view, to detect TLEs and lightning in the night hemisphere [Chern et al., 2003; Chen et al., 2008]. The imagery and photometric data studied here were all recorded by the ISUAL imager through a N 2 1P (623 750 nm) filter and the ISUAL spectrophotometer (SP). The imaging area of the ISUAL iccd is 512 pixels 128 pixels, with the field-of-view (FOV) of 20 (H) 5 (V). ISUAL records six consecutive image frames for each event trigger, with the frame integration time of 29 ms and a 1 ms dead time between frames. The SP shares the same boresighted FOV with the imager. The time resolution for the SP data is 0.1 ms. The SP contains six independent channels: SP1 (150 290 nm; FUV, N 2 LBH band), SP2 (335 341.2 nm, centered at 337 nm; 2PN 2 (0,0)), SP3 (387.1 393.6 nm, centered at 391 nm; 1NN + 2 (0,0)), SP4 (608.9 753.4 nm; 1PN2), SP5 (centered at 777.4 nm; O I emission in lightning) and SP6 (244 392 nm; MUV, 2PN 2 ). [8] For an ISUAL TLE, its geographic coordinates/location can be determined from the imager frames under the assumption that the tops of the TLE or its parent thundercloud were at certain heights [Hsu et al., 2003; Chen et al., 2008]. With the inferred event coordinates and the satellite coordinates at the time of the event, the distances between satellite and the ISUAL TLE can be obtained. Hence, the lateral distance between two points on the image can be estimated from the event distance and their angular distance. In this work, the terminal altitudes of the ISUAL sprites are assumed to be at 90 km, roughly being at the lower boundary of the local ionosphere. The altitudinal heights of the events discussed in the texts and shown in the figures are all inferred basing on this assumed height. [9] The NCKU magnetic ULF station (0.3 Hz to 500 Hz) locates at the Lulin observatory, which is on the central mountain ridge of Taiwan. The system consists of a pair of EMI-BF4 magnetic induction coils to record the horizontal magnetic field emitted by the source discharges [Huang et al., 2011]. The coil orientations are parallel and perpendicular to the geomagnetic field. The recorded signals can be used to infer the discharge polarity, the vertical current moment and the time-integrated charge moment change (CMC) of TLEs and lightning [Huang et al., 2011]. The NCKU ULF station [Wang et al., 2005] has been rebuilt several times after the lightning damages and has operated continuously since June 2009. [10] From July 2004 to December 2010, more than 18,000 TLEs were recorded by the ISUAL experiment, while only 2of7

20 were secondary TLEs. Morphologically, these secondary TLEs can be categorized into two groups, according to their terminal altitudes. The first group of fifteen events is called the secondary jets and the second group of five events is the secondary gigantic jets (secondary GJs). 3. Secondary Jets [11] As shown in Figures 1a and 1b, after the occurrence of a single column sprite or a cluster of sprites, one or several secondary jets were seen to propagate from the cloud top region toward the base of the preceding sprite(s). Of the fifteen secondary jets, 12 events have morphologies similar to that shown in Figure 1a, while the rest resembles that in Figure 1b. Three characteristics can be inferred from these secondary jets: first, all of the secondary jets seem to occur near the cloud top region that is directly under the preceding sprites; second, the secondary jets all have straight luminous columns; third, the number of the secondary jets seems to correlate well with the number of the preceding sprite column(s) demonstrating the visual corresponding between the column(s) of secondary jets and the preceding sprite(s) shown in Figures 1a and 1b. [12] The first two characteristics are consistent with those reported in the previous ground observations [Heavner, 2000; Marshall and Inan, 2007]. The possible generating scenario is shown in Figure 2. After a +CG induced a sprite and left an excess of negative charge in the thundercloud (left panel), the sprite current transports positive charge from the ionosphere to the base of sprites, forming a positive capacitor plate. In the meantime, the negative charge left behind in the cloud by the sprite-producing +CG lightning forms the negative capacitor plate and becomes the charge source for the secondary TLEs (right panel). When the electric field between the capacitor plates exceeds the local breakdown field, secondary TLEs can be generated. [13] As for the third characteristic, the numbers of secondary jets may be affected by both the capacitor plate area contributing by the sprite columns and the distribution and abundance of charge left in the cloud. If the channel length, size and the sprite current [Cummer et al., 1998] for a sprite column are assumed to be equal, then, as the number of columns in the preceding spites increases, the capacitor plate area near the lower edge of sprites is larger as well. When the capacitor area is larger, the local E-field can influence a greater area. In addition, if the negative charge left in the cloud is abundant, the originating sources for the upward discharges may be many. In the case of clustering sprites, multiple secondary jets tend to occur since the capacitor plate area is comparatively large and if the residual negative charge is plentiful. While if the residual negative charge is comparatively small and the preceding single sprite left behind a small capacitor plate, a single column secondary jet will be generated. [14] If the source of the negative charge left in the cloud concentrated in a zone adjacent to the shielding area of the preceding sprite (s), the secondary jet may propagate upward along the outer electric field line toward the ionosphere boundary; Figure 2. A simple electric field model shows that if the columns of the preceding sprites are distributed in an approximately symmetrical manner, the electric field lines are straight and vertical in the region below the symmetrical center of the preceding sprites and comparatively curvy near the rim. Assuming that the negative charge sources for jets are distributed in the cloud region that are below the symmetrical center of preceding sprites, the secondary jets may propagate along the straight electric field lines, and develop straight luminous trunks. Conversely, if the negative charge sources for jets are located in the cloud near the rim of the shielding region of the preceding sprites, the secondary jets may propagate along the curved electric field lines and develop into secondary GJs with curvy trunks. 4. Secondary Gigantic Jets [15] As of December of 2010, five secondary GJs were identified in the ISUAL recorded TLEs. Depending on their location in relative to the preceding sprites, they can be shifted secondary gigantic jet (two events) or pop-through secondary gigantic jets (three events). Their characteristics and the probable generating scenarios are to be discussed in the following two sub-sections. 4.1. Shifted Secondary Gigantic Jets [16] The relative positions with the preceding sprites and the occurrence sequence for this type of secondary GJs are depicted in Figure 1c. The dashed lines in Figure 1c are fixed position markers to act as guides to the eyes. In the first frame, it can be seen that the sprites locate to the right of the dashed lines, while in the second frame, a secondary GJ occurs between the two dashed lines. This indicates that there is an obvious shift in the locations between sprites and the secondary GJ. Hence, this type of secondary GJs can also be termed as the shifted secondary GJs. In the third and fourth frames, the trailing jet of the GJ and the secondary jet, which occur directly below the preceding sprites, are evident. The fifth frame shows that a new sprite occurs to the left of the secondary GJ. The secondary GJ in the second frame and the sprite in the fifth frame morphologically are very similar, but their spectral features are distinctly different. Both SP2 and SP6 of this secondary GJ show double photometric peaks, which is a feature similar to that reported in Kuo et al. [2009] and Chou et al. [2010]; whereas a sprite is always recorded along with the 777.4 nm lightning emissions (SP5) if it occurs before the Earth s limb. Analysis of the two shifted secondary gigantic jets reveals three characteristics. First, the shifted secondary GJs originate from the cloud top region under the shielding area of the preceding sprites, which the same as that for the secondary jets. Second, sprites and jets seem to occur alternatively in space, with either a left-to-right or right-to left configurations. Third, the secondary GJs have curvy luminous bodies, in sharp contrast to the straight luminous columns of the secondary jets. [17] With the above mentioned properties, the features exhibited by the shifted secondary GJs may reflect the complexity of the charge distribution inside the thundercloud. It may be that the charge distribution in the thundercloud that produces the shifted secondary GJ is comparatively complicated, with the positive and negative charges congregate into many parcels. The +CG that induced the preceding sprites shown in Figure 1c had only removed a parcel of the positive charge, leaving parcels of negative charge inside or outside the shielding area of the preceding sprites to be the source discharge for ensuing secondary TLEs. As indicated in 3of7

Figure 1. Representative image sequences of ISUAL secondary jets and GJs (cropped): (a) secondary jets 2010/05/31 23:27:48.128 UTC, (b) secondary jet 2005/07/19 14:57:27.065 UTC, (c) shifted secondary GJ 2009/06/29 23:19:47.958 UTC, and (d) pop-through secondary GJ 2008/09/30 21:52:39.660 UTC. The integration time for each frame is 29 ms with one millisecond readout dead time between frames. Figure 3a, under the shielding region, the jets terminate at the lower local ionosphere boundary, while if they extend outward and upward, the jets can reach the normal ionosphere boundary, forming the shifted secondary GJs. After the occurrence of jets, the residual positive charge in the cloud may still be plentiful to produce another +CG that induces a new sprite nears the secondary GJ. Therefore, in this way, the sprites and the jets seem to occur alternatively in space. 4.2. Pop-Through Secondary Gigantic Jet [18] Contrasting to the shifted secondary GJ, there is another type of secondary GJs that show no discernible location shift from the preceding sprites. As exhibited in Figure 1d, following the circular sprites [Vadislavsky et al., 2009] (fourth frame), a GJ seems to develop from the cloud top region below the symmetrical center of the preceding sprites and reaches the lower ionosphere (fifth frame). In the sixth frame, the residual glow from the trailing jet is still visible. This type of secondary GJs seems to propagate right through the center region of the preceding circular sprites to reach the lower ionosphere, thus they are termed as pop-through secondary gigantic jets. [19] It is possible that, after the occurrence of the circular sprites, the ionized gas and the atmospheric conductivity increase in the volumes occupied by the sprites, while the conditions at the center of the circular sprites remain unaffected. Thus a lowered ring-like capacitor plate forms at the bottom of these sprites, while the ionosphere boundary at the Figure 2. Probable generating scenario for the secondary jets and the secondary GJs: blue cones, upward propagating negative secondary jets or GJs; red cones, downward propagating positive sprites. The dashed lines denote the electric field lines and the arrows indicate the directions of the field. (left) Sprites move the positive charge to the lower edge of sprites after a +CG, and the local ionosphere is at the lower edge of the sprites. (right) Subsequently, the negative charge left behind in the cloud may develop into the secondary jets or GJs propagating along the local electric field lines. 4 of 7

Figure 3. Proposed generating scenarios for (a) a shifted secondary GJ and (b) a pop-through secondary GJ: blue cones, upward propagating negative secondary jets or GJs; red cones, downward propagating positive sprites. The thick lines denote the probable local ionospheric boundaries. center remains undisturbed, as depicted in Figure 3b. Under this scenario and with a fast thundercloud charging time [Pasko and George, 2002], secondary GJs can be generated and propagate straight up to the ionosphere, without being cut off by the lowered ring-plate-like local ionosphere. Thus, in contrast to the shifted secondary GJ, the pop-through secondary GJs have straight luminous columns, since they develop along the straight electric field lines near the symmetrical center. [20] Arguably, the line-of-sight displacement between the preceding sprites and secondary TLEs and even the configuration of the preceding sprites cannot be conclusively determined from the images obtained in a single site. The pop-through secondary GJ may have occurred in front of or behind the sprite columns. However, judging from the images, the distribution of the sprite columns is very symmetrical and the occurrence sequences of the three pop-through GJs are all very similar. Therefore, it is likely that the arrangement of the sprite columns was circular and the secondary GJs indeed developed through the center of the preceding sprites. 5. Discussion [21] Two of the secondary TLE events analyzed in this work were observed after June 2009 and have clearly associated ULF signals. The secondary jets 2009/12/09 14:40:04.326 UTC, as shown in Figure 4, is the one of the two secondary jets with clear image, analyzable photometric data, and associated ULF signal. In Figure 4a, the first photometric peak seen in all the SP channels was from the lightning that induced the preceding sprites and likely also contained the sprite emissions. The second photometric peak is only seen in SP2 and SP6, which indicating it was from a strong blue emitting event. The SP3 signal is also expected to contain a second photometric peak, but it may have been overwhelmed by the continuing current signal of the preceding lightning. The second photometric peak was from the secondary jets shown in the cropped image of Figure 4a [Chou et al., 2010]. Figure 4b exhibits the associated ULF signal for this event. The onset time of the secondary jet is trailing that of the first peak by 7 ms. The first ULF peak was from the sprite-inducing +CG, and the smaller peak at 388 ms was from the secondary jet. Its time difference between the onsets of the sprite-inducing +CG pulse and the secondary jets is 7.7 ms, consistent with that for the photometric peaks. The small bump at 384 ms in ULF data may have come from the sprite current. The features of the ULF signal for the other secondary jet are similar to those shown in the top panel of Figure 4b while with a lower signal-tonoise ratio. [22] The horizontal magnetic ULF signal (the negativegoing pulses) indicates that the sprite-inducing lightning and the secondary jet are of the positive polarity, noting a downward pointing electric field. Therefore, the spriteinducing lightning should be a +CG or +IC and the secondary jet should be a negative cloud-to-ionosphere event to conform to the pointing direction of the deduced electric field. One contributing factor toward the detection of the ULF sferics from this secondary jet likely is due to the proximity of this event (3300 km) to the NCKU-ULF station. Also, the combined ULF emissions from the multiple upward discharge columns in this secondary jet may have elevated its detectability. The inferred charge moment change (within 2 ms of the peak) for the sprite-inducing lightning is 1300 C km, which is above the threshold value needed to induce sprites [Hu et al., 2002]. In the same period, the peak current moment is 580 ka km. The computed charge moment change (within 2 ms of the peak) associated with this secondary jets is 490 C km, while the peak current moment is 170 ka km. [23] Comparing with the ISUAL blue jet reported in Chou et al. [2010, Figure 3] (the terminal height that blue jet was <50 km), the secondary jets analyzed here have longer luminous trunks and morphologically resemble the trailing jets of the type I GJs more; Figures 1 and 4a. The recorded ULF waveforms indicate that all preceding sprite-inducing lightning are +CG events. Therefore, the electric fields above these thunderclouds likely were downward right after these +CG discharges. Since secondary jet 2009/12/09 14:40:04.326 UTC is confirmed to be a CI event, the direction of the E-field above the thundercloud probably stay the same between the occurrences of the preceding sprites and the secondary jet. Hence the discharge source of this secondary jet probably was from the mid-level negative charge reservoir, assuming it was a normally electrified cloud. In a normal polarity storm, blue jets usually originate from the upper-level positive charges, and thus have a positive discharge polarity [Krehbiel et al., 2008]. Hence, the blue jet reported in Chou et al. [2010, Figure 3] probably is a positive jet, while the secondary jets studied here likely are negative jets. The difference in the luminous form of the blue jets and the secondary jets may partially due to the nature of the discharges, while the abundance of the charge in the charge reservoir and the closeness of the capacitor plates may also contribute. [24] In this work, the secondary jets and secondary GJs were observed to closely follow the preceding sprites. Therefore, the sprite-producing lightning may have upset the charge balance in the cloud much like the IC(s) for the GJs, and then entice the jet to develop from the mid-level negative charge layer. After the sprite-producing +CG, the depletion of positive charge in the thundercloud will favor the formation of negative jets, while impede the occurrence of positive jets [Krehbiel et al., 2008, Figure 4]. 5of7

Figure 4. Secondary jets 2009/12/09 14:40:04.326 UTC: (a) imagery and photometric data and (b) the associated magnetic ULF data, the current moment, and the charge moment change. The confirmed negative polarity of the secondary jet 2009/12/09 14:40:04.326 UTC appears to be consistent with this deduction. In a few ISUAL recorded complex TLEs, sprites and jets appear to occur alternatively in space, especially in the events containing the shifted secondary GJ. It is possible that the charge distribution in the thundercloud that produce the secondary jets or GJs may be comparatively complicated, and that the positive and negative charges may have congregated into many parcels. With the occurrence of the sprite-producing +CG or IC(s), it is possible that only a portion of the charge was neutralized, while at the same time, it creates a charge imbalance in the cloud. The charge imbalance may eventually cause the jet(s) to develop from the mid-level charge layer. [25] Between July 2004 and December 2010, ISUAL recorded more than 1300 sprites while only 20 of them contain secondary jets or secondary GJs. The rarity of secondary TLEs indicates that the enhancement of electric field from the lowered ionosphere capacitor plate may not be the main contributing factor for the secondary jet/gj initiation. The lowered ionosphere plate near the bottom of the preceding sprites mainly plays a role in constraining the terminating height of the subsequent jets/gjs but is relatively less important in triggering the subsequent jets/gjs. In other words, the sprite-producing lightning probably is more important in the generation of the secondary jets/gjs, through setting up the charge imbalance in the cloud and setting up the negative cloud plate. The available radio data confirmed that two of the observed secondary TLEs are negative-gj-like events and thus they may have originated from the negative mid-level charge layer. The discharge could start with a bidirectional leader event between the upper positive and the mid negative charge reservoirs. The extensive network of positive leaders developing in the negative charge reservoir could enable the negative leaders to propagate with a high intrinsic potential, which in turn may enhance the chance to develop into a secondary jet or GJ [Krehbiel et al., 2008]. Therefore, the abundance of negative charge and the charge distribution in the thundercloud could be the more important factors for the generation secondary TLEs following the occurrence of sprites. [26] It should be noted that the occurrence scenarios of the secondary TLEs proposed here are based on the analysis of the events recorded by ISUAL up to the present moment. Other scenarios may not be excluded completely. 6. Conclusion and Summary [27] From July 2004 to December 2010, ISUAL recorded 20 secondary TLEs that occurred under or near the shielding area of the preceding sprites. The photometric data demonstrate that these secondary TLEs are predominately blue 6of7

events, which is consistent with the previous observations [Heavner, 2000; Marshall and Inan, 2007]. The photometric data also indicate that there is no coinciding lightning discharge at the time of the secondary TLEs. Meanwhile, two events were found to have clear associated ULF pulses. Therefore, these secondary TLEs are classified as upward discharging jets. [28] Depending on the terminal altitude, the secondary jets can be secondary jets with a terminal altitude of 50 km (near the lower edge of sprites) or secondary gigantic jets with a terminal altitude of 90 km (near the lower ionosphere). Furthermore, from their relative position to the preceding sprites, the secondary GJs can be shifted secondary GJs that extended outside the shielding area of the preceding sprites to reach the ionosphere or pop-through secondary gigantic jets that developed right through the center region of circular sprites. Combining the observational features of the secondary jets/gjs analyzed in this work, it is believed that the factors in influencing the generation of the secondary TLEs include the height of the local ionosphere boundary, and more importantly the abundance and the distribution of the negative charge in the cloud following the occurrence of the sprite-inducing CG lightning. It appears that the preceding sprite mainly exert its influence on the secondary jet/gj by perturbing the local ionosphere height. If the storm charge is sufficient to produce secondary breakdowns, the upward propagating jets that are cut off by the lowered ionosphere become the secondary jets, as illustrated in Figure 2; while those that reach the normal ionosphere height freely become the secondary GJs. [29] Acknowledgments. Works were supported in part by the National Space Organization (NSPO) and National Science Council in Taiwan under grants NSPO-S-100010, NSC 99-2112-M-006-006-MY3, NSC 100-2111-M-006-001, NSC 100-2119-M-006-015, NSC 100-2111- M-006-002, NSC 99-2111-M-006-001-MY3. [30] Robert Lysak thanks the reviewers for their assistance in evaluating this paper. References Chen, A. B., et al. (2008), Global distributions and occurrence rates of transient luminous events, J. Geophys. Res., 113, A08306, doi:10.1029/ 2008JA013101. Chern, J. L., R. R. Hsu, H. T. Su, S. B. Mende, H. Fukunishi, Y. Takahashi, and L. C. Lee (2003), Global survey of upper atmospheric transient luminous events on the ROCSAT 2 satellite, J. Atmos. Sol. Terr. Phys., 65, 647 659, doi:10.1016/s1364-6826(02)00317-6. Chou, J. K., et al. (2010), Gigantic jets with negative and positive polarity streamers, J. Geophys. Res., 115, A00E45, doi:10.1029/2009ja014831. Cummer, S. A., U. S. Inan, T. F. Bell, and C. P. Barrington Leigh (1998), ELF radiation produced by electrical currents in sprites, Geophys. Res. Lett., 25(8), 1281 1284, doi:10.1029/98gl50937. Cummer, S. A., J. Li, F. Han, G. Lu, N. Jaugey, W. A. Lyons, and T. E. Nelson (2009), Quantification of the troposphere to ionosphere charge transfer in a gigantic jet, Nat. Geosci., 2, 617 620, doi:10.1038/ ngeo607. Greifinger, C., and P. Greifinger (1976), Transient ULF electric and magnetic fields following a lightning discharge, J. Geophys. Res., 81(13), 2237 2247, doi:10.1029/ja081i013p02237. Heavner, M. J. (2000), Optical spectroscopic observations of sprites, blue jets, and elves: Inferred microphysical processes and their macrophysical implications, PhD thesis, Univ. of Alaska Fairbanks, Fairbanks. Hsu, R. R., H. T. Su, A. B. Chen, L. C. Lee, M. Asfur, C. Price, and Y. Yair (2003), Transient luminous events in the vicinity of Taiwan, J. Atmos. Sol. Terr. Phys., 65, 561 566, doi:10.1016/s1364-6826(02)00320-6. Hu, W., S. A. Cummer, W. A. Lyons, and T. E. Nelson (2002), Lightning charge moment changes for the initiation of sprites, Geophys. Res. Lett., 29(8), 1279, doi:10.1029/2001gl014593. Huang,S.-M.,C.-L.Hsu,A.B.Chen,J.Li,L.-J.Lee,G.-L.Yang, Y.-C. Wang, R.-R. Hsu, and H.-T. Su (2011), Effects of notch filtering on the ELF sferics and the physical parameters, Radio Sci., 46, RS5014, doi:10.1029/2010rs004519. Krehbiel, P. R., J. A. Riousset, V. P. Pasko, R. J. Thomas, W. Rison, M. A. Stanley, and H. E. Edens (2008), Upward electrical discharges from thunderstorms, Nat. Geosci., 1, 233 237, doi:10.1038/ngeo162. Kuo, C.-L., et al. (2009), Discharge processes, electric field, and electron energy in ISUAL recorded gigantic jets, J. Geophys. Res., 114, A04314, doi:10.1029/2008ja013791. Lyons, W. A., R. A. Armstrong, E. A. Bering III, and E. R. Williams (2000), The hundred year hunt for the sprite, Eos Trans. AGU, 81(33), 373, doi:10.1029/00eo00278. Marshall, R. A., and U. S. Inan (2007), Possible direct cloud-to-ionosphere current evidenced by sprite-initiated secondary TLEs, Geophys. Res. Lett., 34, L05806, doi:10.1029/2006gl028511. Moudry, D. R. (2003), The dynamics and morphology of sprites, PhD thesis, Univ. of Alaska Fairbanks, Fairbanks. Moudry, D. R., H. Stenbaek-Nielsen, D. Sentman, and E. Wescott (2001), Fingers/embers/trolls occurring in the wake of sprites, Eos Trans. AGU, 82(47), Fall Meet. Suppl., Abstract AE31A-0072. Pasko, V. P. (2008), Blue jets and gigantic jets: Transient luminous events between thunderstorm tops and the lower ionosphere, Plasma Phys. Controlled Fusion, 50, 124050, doi:10.1088/0741-3335/50/12/124050. Pasko, V. P. (2010), Recent advances in theory of transient luminous events, J. Geophys. Res., 115, A00E35, doi:10.1029/2009ja014860. Pasko, V. P., and J. J. George (2002), Three-dimensional modeling of blue jets and blue starters, J. Geophys. Res., 107(A12), 1458, doi:10.1029/ 2002JA009473. Pasko, V. P., U. S. Inan, T. F. Bell, and Y. N. Taranenko (1997), Sprites produced by quasi-electrostatic heating and ionization in the lower ionosphere, J. Geophys. Res., 102(A3), 4529 4561, doi:10.1029/96ja03528. Pasko, V. P., M. A. Stanley, J. D. Mathews, U. S. Inan, and T. G. Wood (2002), Electrical discharge from a thundercloud top to the lower ionosphere, Nature, 416, 152 154, doi:10.1038/416152a. Sentman, D. D., E. M. Wescott, D. L. Osborne, D. L. Hampton, and M. J. Heavner (1995), Preliminary results from the Sprites94 aircraft campaign: 1. Red sprites, Geophys. Res. Lett., 22, 1205 1208, doi:10.1029/ 95GL00583. Su, H. T., R. R. Hsu, A. B. Chen, Y. C. Wang, W. S. Hsiao, W. C. Lai, L. C. Lee, M. Sato, and H. Fukunishi (2003), Gigantic jets between a thundercloud and the ionosphere, Nature, 423, 974 976, doi:10.1038/ nature01759. Vadislavsky, E., Y. Yair, C. Erlick, C. Price, E. Greenberg, R. Yaniv, B. Ziv, N. Reicher, and A. Devir (2009), Indication for circular organization of column sprite elements associated with eastern Mediterranean winter thunderstorms, J. Atmos. Sol. Terr. Phys., 71(17 18), 1835 1839, doi:10.1016/j.jastp.2009.07.001. van der Velde, O. A., J. Bór, J. Li, S. A. Cummer, E. Arnone, F. Zanotti, M. Füllekrug, C. Haldoupis, S. NaitAmor, and T. Farges (2010), Multiinstrumental observations of a positive gigantic jet produced by a winter thunderstorm in Europe, J. Geophys. Res., 115, D24301, doi:10.1029/ 2010JD014442. Wang, Y.-C., K. Wang, H.-T. Su, and R.-R. Hsu (2005), Low-latitude ELFwhistlers observed in Taiwan, Geophys. Res. Lett., 32, L08102, doi:10.1029/2005gl022412. Wescott, E. M., D. Sentman, D. Osborne, D. Hampton, and M. Heavner (1995), Preliminary results from the Sprites94 aircraft campaign: 2. Blue jets, Geophys. Res. Lett., 22, 1209 1212, doi:10.1029/95gl00582. 7of7