Mazur and Ruhnke (2011) and Yoshida et al. (2012), the connection of a RL with a grounded conducting lightning channel can result in an M-component or "M-event." For the bidirectional leader, Montanyà et al. (2015) proposed that the subsequent discharge process was very similar to a return stroke (RS) in a negative cloud-toground flash. However, due to limited temporal resolutions (90-1,000 μs) in those previous studies, the connection of a RL with a conducting channel and its subsequent discharge processes have never been properly resolved before, thus their detailed features remain unknown.Using a high-speed video camera NAC and a high-speed optical imaging system Lightning Attachment Process Observation System (LAPOS3) , we observed a special discharge progression which extended its channel not through an ordinary leader, but through a series of recoil leaders and the subsequent discharges caused by the connection of the recoil leaders to other conducting leader channels. Similar discharge progression was recently reported by Wu et al. (2021). However, due to the resolution limitation (20 or 50 μs), detailed features of such a special discharge progression were not well-resolved. In this study, the sub-microsecond time resolution of LAPOS3 allows us to provide direct evidences for the detailed features of the subsequent discharges. This paper reports discharge features of recoil leaders and subsequent discharge processes.
InstrumentationWe have been making observation of lightning to a windmill and its lightning protection tower at Uchinada, Japan for more than 10 years. The detailed observation items have been listed up by . For this study, an optical observation site was set about 1.45 km from the windmill and its lightning protection tower as illustrated by Takamatsu et al. (2015). The upward lightning studied in this paper was triggered from the