JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 97, NO. D9, PAGES 9935-9950, JUNE 20, 1992 ELECTRIC FIELD PULSES IN K AND M CHANGES OF LIGHTNING GROUND FLASHES Vladimir A. Rakov, • Rajeev Thottappillil, and Martin A. Uman Department of Electrical Engineering, Universit? of Florida, Gainesville Abstract. From electric field records of likely to have pulses than M changes during 27 ground flashes near Tampa, Florida, and 19 continuing-current field changes of longer ground flashes at the NASA Kennedy Space duration. Center (KSC) the occurrence and waveshape of microsecond-scale electric field pulses Introduction and Literature Review associated with both millisecond-scale steplike K changes and millisecond-scale K changes. The term "K change"is used to hook-shaped M changes are examined to test and denote the relatively small electric field disprove the following two hypotheses: (1) changes that occur in the intervals between that K changes contain a microsecond-scale and after the strokes of a groundflash and in pulse component which can be described by the the final (J type) portion of a cloud flash characteristic pulse waveform proposed by [e.g., Kitagawa and Brook, 1960; Uman, 1987]. Arnold and Pierce (1964) and (2) that there is At ranges up to some tens of kilometers, essentially no difference between K and M instrumentally undistorted (system decay times processes, as argued by Kitagawa et al. of the order of some tens of milliseconds or (1962). Microsecond-scale electric field longer) K changes in both ground and cloud variations exceeding by at least 50% the flashes appear as steplike or ramplike field system noise level were observed in 23% of 135 changes,depending on the time scale on which K changes from Tampaand in 25% of 128 K the changes are observed [Brook and Kitagawa, changes from KSC,while such field variations 1960; Kitagawa and Brook, 1960; Ogawaand were found in 44%of 88 M changesfrom Tampa Brook, 1964; Krehbiel et al., 1979, Thomson, and in 77% of 30 M changes from KSC. In the 1980; Bils et al., 1988; Thottappillil et al., majority of the K changes having microsecond- 1990]. Krehbiel et al. [1979] report that scale pulse activity, that activity did not ground-flashKchanges recorded simultaneously occur at the beginning of the K step, while in at eight stations had, at each station, the most cases the pulses associated with M same polarity as the slower interstroke field changes occurred at the initial portion of the change (J change) on which the K changes were M hook. These results can be interpreted to superimposed,in spite of the fact that the J imply that K changes and M changes are field changes could be of different polarity associatedwith dissimilar physical processes, at different stations. A similar tendencyhas in refutation of hypothesis (2) above. The been found in single-station studies [e.g., microsecond-scale pulse activity during K Kitagawa, 1957; Kitagawa and Brook, 1960; this changes and M changes was highly variable and study], although opposite polarity K-type sometimes irregular in waveshape. Not all the field changeswere also observed. The most pulses had the same polarity as the K step or recent review of the various characteristics the initial portion of the M hook on which of K changes is given by Thottappillil et al. they were superimposed. The relation of the [1990] who found, in Florida ground flashes, microsecond-scale variations to the overall K a geometric mean K-change duration of 0.7 ms changes in ground flashes as regards the and a geometric meantime interval between K frequency of occurrence, the position of changesof about 13 ms. pulses within the slower field change, and the The physical process giving rise to what shape of the pulses is similar to that later [Kitagawa, 1957] would be termed a K reported by Bils et al. (1988) for K changes change was apparently first inferred by Malan in cloud flashes. The observed microsecond- and Schonland [1951, pp. 161-162] from their scale field variations associated with K electric field and photographic observations changes are not consistent with the of ground flashes in South Africa. They characteristic electric field pulse waveform interpreted steplike field changes (see their attributed by Arnold and Pierce (1964) and Figure 9d) occurring between strokes and after someother investigators to K changes, in the last stroke, which were not accompanied by refutation of hypothesis (1) above. No any detectable luminosity in the channel relation was observedbetween the magnitudeof between cloud base and ground, as due to a a K change and the presence or absence of leader not being able to reach ground and corresponding microsecond-scale field instead producing "a small readjustment of variations. M changes during continuing- charge within the cloud." Subsequently, the current field changes of relatively short K process in both groundand cloud flashes has duration (less than 20 ms or so) are more been hypothesized [e.g., Kitagawa, 1957; Ogawa and Brook, 1964] to be a negative "recoil streamer" that occurs when a positive J type •Permanently at High Voltage Research streamer propagating within the cloud Institute Tomsk 634050, Russia. encounters a region of concentrated negative charge. Clegg and Thomson [1979] found that gaps of typically 6 to 9 ms in the HF (10 MHz) radiation following return strokes were Copyright 1992 by theAmerican Geophysical invariably terminated by a K change. They Union. suggested that these quiet periods were associated with nonradiating positive J Paper number 92JD00797. streamers propagating from the top of the 0148-0227/92/92JD-00797505.00 preceding return-stroke channel and producing 9935