VHF (Very High Frequency) lightning interferometers can locate and observe lightning discharges with a high time resolution. Especially the appearance of continuous interferometers makes the 2-D location of interferometers further improve in time resolution and completeness. However, there is uncertainty in the conclusion obtained by simply analyzing the 2-D locating information. Without the support of other 3-D total lightning locating networks, the 2-station interferometer becomes an option to obtain 3-D information. This paper introduces a 3-D lightning location method of a 2-station broadband interferometer, which uses the theodolite wind measurement method for reference, and gives the simulation results of the location accuracy. Finally, using the multi-baseline continuous 2-D locating method and the 3-D locating method, the locating results of one intra-cloud flash and the statistical results of the initiation heights of 61 cloud-to-ground flashes and 80 intra-cloud flashes are given. The results show that the two-station interferometer has high observation accuracy on both sides of the connection between the two sites. The locating accuracy will deteriorate as the distance between the radiation source and the two stations increases or the height decreases. The actual locating results are similar to those of the existing VHF TDOA (Time Difference of Arrival) lightning locating network.

A low-frequency lightning location system was established and operated. The time-reversal approach in the frequency domain (TRFD) is applied to the system for three-dimensional (3-D) lightning source location finding. The TRFD method uses back propagation of the signals detected at each site and determines the source location automatically using the focusing point of the back-propagated waves. Three cases are reported to showcase the capability of the method. The first case is the 3-D localization of an intra-cloud flash over the network. Both TRFD and time-of-arrival (TOA) method provide LF lightning sources that are in consistent with very high frequency (VHF) sources. The lightning sources located by TRFD method reveal more details and are more continuous compared to those obtained by TOA method used in this paper. The second case concerns the localization of an artificially triggered lightning flash with a horizontal error of 30 m by TRFD. The third case is a multi-stroke cloud-to-ground flash that shows location differences between four return strokes (RSs) being less than 60 m. Furthermore, a special case was used to demonstrate the ability of the proposed method in locating RS grounding points by combining two successive RSs when the TOA is not available.

High‐speed video data were used to analyze the initiation and propagation of 36 needles and their associated 306 flickering events observed in a single‐stroke positive cloud‐to‐ground (+CG) flash. The needles occurred during the return‐stroke later stage and the continuing current, within approximate 10 ms after the onset of the +CG return stroke. They initiated near the lateral surface of the predominantly horizontal channel and extended almost perpendicular to that channel. Flickering events are recoil type streamers (or leaders) that retrace the channels created by needles. Flickering events can be repetitive and are classified into four categories based on different scenarios of their occurrence. Needles are caused by the radial motion of negative charge from the hot core of the positive‐leader channel into the positive corona sheath surrounding the core, when the core is rapidly recharged (its radial electric field reversed) by the return‐stroke process and during the following continuing current.

To investigate the characteristics of regions exhibiting multiple lightning initiations within a finite volume and a short time, the lightning location data obtained from the convective regions of 14 mesoscale convective systems were analyzed in combination with data from radar. In total, 415 out of 5996 radar grids (1 km × 1 km × 0.5 km) were found to initiate more than one flash within 6 min. Only 49 grids showed an initiation density of more than two flashes within 6 min. The grids with high flash initiation densities were found to have a similar distribution to those with one lightning initiation within 6 min, in terms of altitude and reflectivity relative to altitude. They also showed similar trends in their frequency evolution. The grids with higher initiation densities seemed to be more concentrated in the altitude range of 9–13 km. However, only one was found to form at a lower altitude near the melting level when lightning initiation clearly declined. Moreover, the spatial relationship of this lower higher-initiation density grid to the reflectivity core was different to that in the main altitude range. In this paper, the possible dynamic and electrical mechanisms of the formation of this lower higher-initiation density grid are discussed.

Using a broadband (from 1 to 250 MHz) interferometer with high temporal‐spatial resolution, we have observed the lightning initiation process of a winter cloud‐to‐ground lightning in Japan with great details. We found that the lightning initiation involved with multiple fast breakdowns behaving like a series of back (downward) and forth (upward) consecutive reflections in a constrained space near the main negative charge region of the lightning. We also found that some initial negative fast breakdowns could propagate with an unusually spread manner. We suggest that fast breakdowns with either reflecting features or spread manners could efficiently utilize the electrostatic energy in a local region with strong electric field for driving subsequent streamers, and therefore may widely exist in lightning initiation processes.