Properties of positive and negative leaders developing in air gaps ranging from 4 to 10 m that were subjected to 100/7,500‐μs voltage impulses were examined using a two‐frame, high‐speed video camera with image enhancement. Abrupt extension (stepping) that culminated in a bright and structured corona streamer burst was observed for both negative (expected for the “classical” stepping process) and positive (expected for the so‐called restrike process) leaders. Selected high‐quality images of five negative and four positive leaders with pronounced corona streamer bursts are presented here. The morphology of corona streamer bursts was essentially independent of polarity. Streamer bursts exhibiting nearly spherical symmetry were observed. For the four positive leaders, the newly added channel sections (steps) were almost straight and had lengths ranging from about 50 to over 120 cm. For the five negative leaders, most of the steps were curved and their 2‐D lengths were some tens of centimeters. It is generally thought that positive leaders in both long sparks and lightning extend continuously or exhibit optically unresolvable steps whose length is comparable to the leader tip size (1 cm or less) and that for sparks only when the absolute humidity is relatively high (>10 g/m³ or so) or voltage rise time is relatively long (around 1 ms or more) can larger steps occur. In this study, both modes of propagation for different branches of the same positive leader were observed.

Based on experimental results of recent years, this article presents a qualitative description of a possible mechanism (termed the Mechanism) covering the main stages of lightning initiation, starting before and including the initiating event, followed by the initial electric field change (IEC), followed by the first few initial breakdown pulses (IBPs). The Mechanism assumes initiation occurs in a region of ~1 km³ with average electric field E > 0.3 MV/(m·atm), which contains, because of turbulence, numerous small “E_th volumes” of ~10⁻⁴–10⁻³ m³ with E ≥ 3 MV/(m·atm). The Mechanism allows for lightning initiation by either of two observed types of events: a high‐power, very high frequency (VHF) event such as a Narrow Bipolar Event or a weak VHF event. According to the Mechanism, both types of initiating events are caused by a group of relativistic runaway electron avalanche particles (where the initial electrons are secondary particles of an extensive air shower) passing through many E_th volumes, thereby causing the nearly simultaneous launching of many positive streamer flashes. Due to ionization‐heating instability, unusual plasma formations (UPFs) appear along the streamers' trajectories. These UPFs combine into three‐dimensional (3‐D) networks of hot plasma channels during the IEC, resulting in its observed weak current flow. The subsequent development and combination of two (or more) of these 3‐D networks of hot plasma channels then causes the first IBP. Each subsequent IBP is caused when another 3‐D network of hot plasma channels combines with the chain of networks caused by earlier IBPs.

An experimental method to measure the electric fields existing in a plasma wake produced by a ∼0.24 GW, ∼0.5 ns, 9.5 GHz microwave pulse traversing a plasma-filled waveguide is presented. The intensity of the second harmonic of a 30 fs 800 nm laser generated inside a gas-filled dielectric tube placed inside the waveguide is used to characterize the wakefield parameters. Three distinct decaying oscillations of the plasma wakefield, with peak amplitude of ∼20 kV/cm, were observed. The experimental results were confirmed by 3D large-scale plasma particle-in-cell simulations.

This study investigates the ionization pressure threshold of a gas (air, helium, argon, and SF6 across a wide pressure range) filled dielectric tube when a ∼300 MW, ∼0.7 ns, 9.6 GHz high-power microwave (HPM) pulse propagates through it. The thresholds are determined as the pressure for which the energy of the transmitted HPM pulse decreases to ∼30%, which is close to the same HPM pulse's transmission coefficient when a metal rod fills the tube. These thresholds are found to be 0.4 × 105 Pa,105 Pa, 1.8 × 105 Pa, and 0.2 × 105 Pa, for air, argon, helium, and SF6, respectively. The measured intensity of the plasma light emission starts to decrease at a pressure which coincides with the pressure threshold determined by HPM pulse propagation. Additionally, at gas pressures 105 Pa transforms into streamer like plasma. Simplified numerical simulations of a microwave discharge in air at 1 × 105 Pa and 4 × 105 Pa are consistent with the experimental plasma light observations.