01:47:52 UT Sprite-producer

The second sprite event of the June 22nd storm was associated with a 83.7 kA +CG at 01:47:52.7112 UT. The video sequence of the sprite cluster is shown in Figure 3.8. The sprite reached maximum brightness in its first frame and then faded into the next 3 frames until it was at the limit of detectability.

A large number of IC and CG flashes occurred in the intervening time between the first and second sprite-producing flashes. However, none of these flashes developed very far into the stratiform region.

The 01:47:52 UT flash began as an ordinary -CG discharge. A -31.2 kA -CG occurred at 01:47:51.4763 UT, about 0.3 s after the start of the flash, according to Figure 3.9. Only a single data point was located by LDAR in the subsequent 0.3 s, though the interferometer log-RF record indicates that there was activity throughout this interval. Figure 3.10a shows that the VHF sources within interval t1 were relatively localized.

The flash developed horizontally outward into the stratiform region during interval t2 (Figure 3.10b). The lateral extent grew to $\sim$64 km in $\sim$1 s during interval t2, corresponding to a progression velocity of 6.4$\times$10$^4$ m/s. This is somewhat less than typically observed for negative leaders, but is not unrealistically low.

The NLDN indicated that a 9.7 kA +CG occurred at 01:47:52.3209 UT about 5 km west of the earlier -CG. The ``weak'' current (relative to ordinary +CGs) indicates that this was probably an IC event, as discussed earlier in Section 3.4. Unfortunately, the interferometer buffer filled up due to an excessive number of RF-triggers during the flash. This resulted in the loss of all fast antenna, log-RF, and phase data for the remaining flash activity, starting at $\simeq$0.24 s prior to the weak +CG. The slow antenna data, which is stored in a separate buffer, was also affected by the lock-up and was lost for the entire flash.

Figure 3.8: The video sequence of the 01:47:52 UT sprite cluster. The sprites reached maximum brightness within the first frame. They then faded over the next 3 frames to the limit of detectability. Scattered light from the parent discharge can be seen in the lower portion of the final 5 frames. Note that the first frame was relatively dark, which suggests that there was relatively little breakdown activity in the parent discharge prior to the +CG. The start time of each video field, accurate to $\pm$5 ms, is shown at the bottom of each image.

Figure 3.10: The spatial development of the 01:47:52 UT sprite-producing discharge as a function of time. The flash produced a -CG ($\triangle$) in the first time interval (t1) and then developed horizontally outward through the anvil in the second (t2). A weak $\simeq$10 kA ``+CG'' (intracloud?) was indicated by the NLDN in interval t2. The sprite-producing +CG's location (+), LDAR-indicated sprite-producing discharge center (*), and sprite locations (x) are shown within time interval t3. As before, many of the sprites were near the periphery of the discharge. The NLDN detected a -CG beneath the discharge in the final interval, t4.

A 83.7 kA +CG occurred at the start of interval t3 and was concurrent with the appearance of sprites on video (Figure 3.8). The sprite positions were determined in the same way as before; the maximum sprite altitude was assumed to be 87$\pm$6 km. Figure 3.10c indicates that many of the sprites occurred near the periphery of the discharge, as in Figure 3.5d.

However, two of the sprites occurred to the north of the apparent northern periphery of the discharge. The discharge appeared to extend further northward during interval t3 only after the sprites appeared. Figure 3.11 shows that these sprites were tilted somewhat from vertical. The direction of the tilt (away from the center of the overall activity) is qualitatively consistent with what would be expected due to a somewhat nonvertical electric field below the ionosphere at a significant distance from the parent CG charge center. Sprites at significant tilt angles have been observed previously by Winckler et al. (1996).

Figure 3.11: The azimuth and elevation orientation of the camera is shown for the 01:47:52 UT sprite event. The same camera orientation was used for the other sprite events up to $\sim$2 UT. Most of the sprites were vertically oriented, with the notable exception of the slanted pair of sprites at $\sim$310$^{\circ}$ azimuth. The slanted sprites occurred north of the apparent northernmost extent of the flash at the start of interval t3 in Figure 3.10c.

During the final time interval, numerous VHF sources were observed to descend to 4-5 km and to propagate at that altitude. As was noted for the previous flash, these sources were likely associated with negative leaders (spider lightning) propagating through the positive charge layer near the 0$^{\circ}$C isotherm. These sources produced a significant secondary maximum below 5 km in the height distribution of VHF sources (Figure 3.12). The primary maximum in the VHF source distribution just below 8 km altitude should correspond to the average height of charge removal for this sprite-producing discharge. The total duration of the flash was about 2.7 s, nearly twice that of the previous sprite-producing discharge.

Figure 3.12: The density of VHF sources in 500 m height bins for the 01:47:52 UT flash. Two peaks in density are evident at just below 8 km altitude and just below 5 km altitude, near the 0$^{\circ}$ C isotherm. The former likely corresponds to the altitude of charge removal for the sprite-producing +CG while the latter was due primarily to development in the late phase of the flash after the sprites had decayed.

An interesting feature of time interval t4 is that the NLDN detected a -CG stroke emanating from underneath the horizontally expansive discharge. The -CG may have been produced by a negative leader spawned from the parent discharge. However, another possible origin will be discussed in Section 3.8, in which several -CG strokes were observed beneath another sprite-producing discharge.