Discussion of paper by M. A. Uman, “The diameter of lightning”

This is the authors' abstract. We don't add key points for this paper.

In the preceding article, Uman [1964] describes an ingenious method of observing the diameter of the lightning flash to a grounded metal conductor mounted on an elevated structure. His method gives significant results, but he is disturbed because there are discrepancies between his observations and those of Hill [1963] on damage to metal conductors and those of his colleagues Evans and Walker [1963], who used photography. These results are not necessarily contradictory, as they represent observations of different aspects of a most complex phenomenon. The lightning flash, which may consist of a sequence of several strokes to a well-grounded metal conductor which projects some distance above the ground, consists of the following chain of events. Initially a stepped leader [Schonland, 1956] stroke originating from an internal predischarge within the cloud [Loeb, 1953, 1958; Clarence and Malan, 1957] forges earthward at some 107 cm/sec. It may have a rather large diameter but its more conducting core is not too extensive in area of cross section. As the leader approaches the earthed conductor its high negative tip potential creates such a high field at the metal conductor that a positive streamer process moves upward toward the stepped leader from one to several meters. Under special conditions this streamer process may extend to kilometers [McEachron, 1939; Berger, 1961, 1962]. In doing so the positive streamer process may have several highly conducting branches [Nasser, 1959a, b, 1962; Nasser and Loeb, 1963; Kritzinger, 1962]. The intense field distortion at the junction of positive streamer branches and negative streamers then sends ionizing space waves of potential [Loeb, 1957 ; Westberg, 1959; Loeb and Cravath, 1935], called return strokes [Schonland, 1956], up the axis of the stepped leader channel and down the positive streamer channel at speeds of 5 × 108 to 1010 cm/sec. These render the channels conducting. Usually the main discharge will favor the most conducting of the branched streamer channels, but one or two others could be involved as well.