Search

We developed a framework merging unsupervised and supervised machine learning to classify lightning radio signals, and applied it to the possible detection of terrestrial gamma-ray flashes (TGFs). Recent studies have established a tight connection between energetic in-cloud pulses (EIPs, >150 kA) and a subset of TGFs, enabling continuous and large-scale ground-based TGF detection. However, even with a high peak current threshold, it is time-consuming to manually search for EIPs in a background of many non-EIP events, and it becomes even more difficult when a lower peak-current threshold is used. Machine learning classifiers are an effective tool. Beginning with unsupervised learning, spectral clustering is performed on the low-dimensional features extracted by an autoencoder from raw radio waveforms, showing that +EIPs naturally constitute a distinct class of waveform and 67% of the total population. The clustering results are used to form a labeled dataset (~10,000 events) to further train supervised convolutional neural network (CNN) that targets for +EIPs. Our CNN models identify on average 95.2% of true +EIPs with accuracy up to 98.7%, representing a powerful tool for +EIP classification. The pretrained CNN classifier is further applied to identify lower peak current EIPs (LEIPs, >50 kA) from a larger dataset (~30,000 events). Among 10 LEIPs coincident with Fermi TGF observations, 2 previously reported TGFs and 2 unreported but suspected TGFs are found, while the majority are not associated with detectable TGFs. In addition, unsupervised clustering is found to reflect characteristics of the ionosphere reflection height and its effect on radio wave propagation.

Gravitational wave high-energy Electromagnetic Counterpart All-sky Monitor (GECAM) is a space-borne instrument dedicated to monitoring high-energy transients, thereinto Terrestrial Gamma-ray Flashes (TGFs) and Terrestrial Electron Beams (TEBs). We propose a TGF/TEB search algorithm, with which 147 bright TGFs and 4 TEBs are identified during an effective observation time of $\sim$ 9 months. We show that, with gamma-ray and charged particle detectors, GECAM can effectively identify and distinguish TGFs and TEBs, and measure their temporal and spectral properties in detail. Moreover, we find an interesting TEB consisting of two pulses with a separation of $\sim$ 150 ms, which is expected to originate from a lightning process near the geomagnetic footprint. We also find that the GECAM TGF’s lightning-association ratio is $\sim$ 80\% in the east Asia region using the GLD360 lightning network, which is significantly higher than previous observations.

Elves are rapidly expanding rings of optical emissions in the lower ionosphere. Narrow bipolar events (NBEs) are signatures in radio signals from intra-cloud discharges. They are thought to be fast streamer breakdown that may trigger the onset of lightning and blue jets. However, there is a lack of experimental evidence on whether the streamer discharges of NBEs carry sufficient currents to generate elves in the lower ionosphere. Here, we report the first simultaneous observation of NBEs and elves that confirm this hypothesis. The NBEs are observed simultaneously from the ground by an array of wave receivers located in China and from space by spectral measurements by the Atmosphere-Space Interactions Monitor (ASIM) on the International Space Station (ISS). We observe thirteen negative and six positive NBEs produced in four thunderclouds penetrating into the stratosphere. Five NBEs are accompanied by elve emissions observed in the near-ultraviolet of the Lyman–Birge–Hopfield (LBH) band. They were at ~18 km altitude, and their peak currents, estimated by a ground-based lightning detection network, were larger than 135 kA. The observations show that the impulse currents of the streamers are of sufficient magnitude to power elves, thereby adding to the new pathways that thunderstorms affect the lower ionosphere.

Elves are rapidly expanding rings of optical emissions in the lower ionosphere, excited by the transient electromagnetic pulses from lightning in the clouds below. Narrow bipolar events (NBEs) are signatures in radio signals from intra-cloud discharges. They are thought to be fast streamer breakdown that may trigger the onset of lightning and blue jets. Recent observations from space suggest they may carry sufficient current to excite elves. Here, we report the first simultaneous observation of NBEs and elves that confirm this hypothesis. Three negative NBEs produced in a tropical storm are observed from the ground by a radio wave receiver located in China and their optical signatures from space by the Atmosphere-Space Interactions Monitor (ASIM) on the International Space Station (ISS). The NBEs are associated with flashes of emissions in N22P at 337 nm, with faint or no activity in the dominant lightning leader line OI at 777.4 nm, confirming their streamer nature. The elves are seen as delayed emissions in the near-ultraviolet of the Lyman–Birge–Hopfield (LBH) band. Our analysis suggests that negative NBE discharges can generate weak, but observable elves when the associated impulse current is above ~ 140 kA. The appearance of elves would provide a new method to estimate the currents of corona discharges at cloud tops, crucial for the characterization of chemical perturbations to greenhouse gas agents at the tropical tropopause.

Most lightning location networks are based on real-time analytical solutions of certain simplified models, while the reality is much more complicated. In this paper, we introduce a graphics processing unit (GPU)-based parallel computing algorithm that can extensively benefit lightning geolocation networks. For a network running this GPU-based algorithm, one can build up a geolocation database based on numerical solutions of certain complete models in advance, and lightning geolocations can then be easily determined with a grid-searching technique in real time. One such grid-searching technique, is the grid traverse algorithm (GTA) for the traditional time of arrival technique. By running GPU-based GTA in a six-station two-dimensional (2-D) and a five-station 3-D networks, we show that extremely high network performance can be achieved, with a processing speed of about 2700 times faster than CPU-based GTA. The location accuracy of GPU-GTA is examined with Monte Carlo simulations, showing that GPU-GTA can locate a lightning source in real time with high accuracy. We also find that when the grid step is comparable with the inherent time uncertainty of a network, the location accuracy cannot be improved further with a finer grid step.

We examine the broadband ( +500 kA, and the associated charge transfer is also considerable (typically >+20 C km). The observed complexity of TGF-associated lightning emissions can also be interpreted in the context of the initial development of normal IC lightning flashes, where the upward negative leader drives the millisecond-scale current and may also provide the seed electrons for avalanche multiplication in the upper part of active thunderstorms. Our analyses show that the thunderstorms in the land area of South China produce TGFs that can be readily observed by Fermi/GBM, and the future ground-based coordinated observations in this area would be rather promising to gain more insights into the physical mechanism of TGFs.

This data folder contains the data included in the paper titled "A Distinct Class of High Peak Current Lightning Pulses over Mountainous Terrain in Thunderstorms" submitted to "Geophysical Research Letters". To use the data, load them in Matlab, and the meaning of the parameters can be found from their variable names. &nbsp

Narrow bipolar events (NBEs) are occasionally reported to occur within a few ms after Terrestrial Gamma-ray Flashes (TGFs), while the formation mechanism remains mysterious partially due to the lack of sufficient observations. Here, nine more TGFs of this scenario are reported with concurrent LF sferics and lightning location data. The gamma-ray production in these TGFs preceded the occurrence of NBEs by a minimum of 60 µs up to 13.5 ms, and no other fast leader discharge was found within 20 ms before the TGF. The TGF-preceded positive NBEs occurred at altitudes of 8.6–11 km in thunderstorms, likely in the high electric field (E-field) region of lightning initiation. The analyses show that the NBE-preceding TGFs bear harder energy spectrum with larger proportion of high-energy photons than EIP-related TGFs produced in association with lightning leader. Our findings support the relativistic feedback mechanism of gamma-ray generation in the large-scale thunderstorm E-field. © 2021. American Geophysical Union., This work was supported by the National Key R&D Program of China (2017YFC1501501), National Natural Science Foundation of China (U1938115, 41805004, and 41875006), and the Chinese Meridian Project. The Insight-HXMT mission is a project funded by China National Space Administration (CNSA) and Chinese Academy of Sciences (CAS). The Fengyun-2G satellite images are provided by National Satellite Meteorological Center (NSMC) of China (www.nsmc.org.cn)., With funding from the Spanish government through the Severo Ochoa Centre of Excellence accreditation SEV-2017-0709.

Thunderstorms in the troposphere produce lightning flashes and cause charge transfer of different strength at varying spatial and temporal scales, leading to various forms of transient electromagnetic effects in the vast space above thunderstorms. In particular, normal intra-cloud (IC) lightning can generate ionizing hard X-rays and gamma rays, forming Terrestrial Gamma-ray Flashes (TGFs). We briefly summarize the progress in TGF studies that has been achieved in the past decade based on multiple space-borne platforms: (1) TGFs are usually associated with the upward negative leader during the initial stage of IC flashes and are often accompanied by relatively strong IC discharge with high peak current and large charge transfer, which is called the energetic IC pulse (EIP); (2) based on the characteristics of TGF-related radio-frequency signals, we can develop a remote sensing approach with ground-based measurements of lightning signals, thereby greatly enriching the investigation dataset of TGFs and parent thunderstorms; (3) till date, no unified mechanism for TGF production has been developed due to a lack of effective observation with respect to the source region. Thermal runaway breakdown and relativistic runaway electron breakdown are the two mainstream theories to explain TGF production. Compared with transient luminous events (TLEs; e.g., red sprites, gigantic jets, and blue jets) as the lightning-induced dielectric breakdown in the mesosphere, studies on TGFs, in terms of both observations and theoretical interpretation, lag behind the research in Europe and the USA. However, along with China's latest progress in space detection technology (particularly the implementation of the Insight Hard X-ray Modulation Telescope (Insight-HXMT) and the Gravitational-wave high-energy Electromagnetic Counterpart All-sky Monitor (GECAM)), researchers in China desire to make steady progress in the field of TGF studies through continuous efforts in developing ground-based lightning detection techniques. © 2020.

On August 19, 2012, the Imager of Sprites and Upper Atmospheric Lightning (ISUAL) onboard the FORMOSAT-2 satellite captured a sequence of seven blue discharges within one minute that emanated from a parent thunderstorm over Lake Taihu in East China. The analysis of lightning activity produced in the thunderstorm indicates that at least six of these events occurred in association with negative NBEs that were concurrent with the blue discharge by less than 1 ms, and negative CGs occurred within 6 s before each blue discharge, which is in agreement with the modeling presented by Krehbiel et al. [2008]. Therefore, the frequent occurrence of negative CGs could provide the favorable condition for the production of blue discharges, and negative NBEs are probably the initial event of blue discharges. The detection of negative NBEs might provide a convenient approach to detect the occurrence of blue discharges as lightning bolt shooting upward from the top of energetic thunderstorms.

Differing from the traditional Time Difference of Arrival (TDOA) lightning location technique, Grid Traverse Algorithm (GTA) is to build up a series of grid points based on numerical discretion of lightning locations in a network region concerned and then to find the lightning location solution by traversing the grid points. Although GTA has advantages over traditional TOA method in many aspects, it is computationally ineffectiveness. To accelerate the GTA processing speed, we have recently proposed a Graphics Processing Unit (GPU) based GTA parallel computing technique, which is proven to be much faster than the existing CPU based GTA technique. In this paper, we give firstly a briefing of the GPU-GTA technique and then the results of performance test of the technique. The performance test is done by applying the GPU-GTA to a specific lightning location network (LLN) with the Monte Carlo simulation approach. The test results show that GPU-GTA can take about 4.3 seconds, 3.6 seconds and 3.2 seconds to locate 1000 points in 3D at the accuracy of 25 m, 50 m and 100 m respectively. The high performance of the GPU-GTA makes it feasible for a real-time LLN with high accuracy.

Terrestrial gamma ray flashes (TGFs) are very short bursts of gamma radiation associated to thunderstorm activity and are the manifestation of the highest-energy natural particle acceleration phenomena occurring on Earth. Photon energies up to several tens of megaelectronvolts are expected, but the actual upper limit and high-energy spectral shape are still open questions. Results published in 2011 by the AGILE team proposed a high-energy component in TGF spectra extended up to $\approx$100 MeV, which is difficult to reconcile with the predictions from the Relativistic Runaway Electron Avalanche (RREA) mechanism at the basis of many TGF production models. Here we present a new set of TGFs detected by the AGILE satellite and associated to lightning measurements capable to solve this controversy. Detailed end-to-end Monte Carlo simulations and an improved understanding of the instrument performance under high-flux conditions show that it is possible to explain the observed high-energy counts by a standard RREA spectrum at the source, provided that the TGF is sufficiently bright and short. We investigate the possibility that single high-energy counts may be the signature of a fine-pulsed time structure of TGFs on time scales $\approx$4 {\mu}s, but we find no clear evidence for this. The presented data set and modeling results allow also for explaining the observed TGF distribution in the (Fluence x duration) parameter space and suggest that the AGILE TGF detection rate can almost be doubled. Terrestrial gamma ray flashes (TGFs) are very short bursts of gamma radiation associated to thunderstorm activity and are the manifestation of the highest-energy natural particle acceleration phenomena occurring on Earth. (...continues)

Lightning mapping with radio frequency location systems plays a more and more important role in understanding the mechanism of lightning discharges during thunderstorms. Here we present a comprehensive analysis of the electromagnetic field measurements and 3-dimensional (3D) mapping of two coincident lightning flashes: an upward positive flash (+CG) with long preceding upward negative leader (UNL), which was triggered by the in-cloud (IC) activity of a nearby downward negative flash (−CG). The flashes were mapped with a low frequency (LF) mapping array deployed near Duke University and during a late-spring thunderstorm passing a plain area. The 3D LF maps show that the IC activity related to the −CG moved negative charge away from the location overhead the UNL. The UNL developed firstly upward for about 2 km and then extended horizontally for about 8 km at that height with a mean speed of 6.2 × 105 m/s. The channel line charge density of the UNL was estimated to range from −542 μC/m to −9.21 mC/m with a mean value of −3.46 mC/m. The total positive charge transferred from the cloud to ground by the +CG flash was estimated to be about 5.8C. From the 3D maps, we inferred that the IC activity relating to the −CG flash which moved negative charge away from overhead region of the UNL could be responsible for the UNL initiation. And the existence of extensive lower positive charges further favored the long horizontal propagation of the UNL inside the cloud.

By comparing the source position of pulses during initial leaders and the paths of subsequent dart leaders from typical bilevel intracloud (IC) flashes, we demonstrate a new method to image the detailed three-dimensional (3-D) structure and stepping dynamics of IC initial leaders. Using this approach, we show that nearly half of the initial breakdown pulses are not involved in the main channel extension but instead originate in nonpropagating branches. The primary upward but significantly tilted initial leader channels propagated with measured mean/median 3-D step length of 295/265 m and mean/median step interval of 1.05/0.6 ms. The structure and dynamics of IC initial upward negative leaders are distinct from those of highly branched cloud-to-ground downward negative leaders. We suggest that the tilted initial leader channel and electric field could affect the observation on leader-associated high-energy radiation phenomena, such as the position-dependent observation of terrestrial gamma ray flashes from space-based detectors.

On the morning of 5 June 2013, two cameras of the SpriteCam network concurrently captured a red sprite with diffuse halo over a mesoscale convective system (MCS) passing the panhandle area of Oklahoma. This sprite was produced by a negative cloud-to-ground (CG) stroke with peak current of −103 kA in a manner different from previous observations in several aspects. First of all, the causative stroke of sprite is located by the National Lightning Detection Network (NLDN) in the trailing stratiform of MCS, instead of the deep convection typically for negative sprites. Second, the sprite-producing stroke was likely the first stroke of a multistroke negative CG flash (with ≥6 CG strokes) whose evolution was mainly confined in the lower part of thunderstorm; although the parent flash of sprite might contain relatively long in-cloud evolution prior to the first stroke, there is no evidence that the negative leader had propagated into the upper positive region of thundercloud as typically observed for the sprite-producing/class negative CG strokes. Third, as shown by the simulation with a two-dimensional full-wave electrodynamic model, although the impulse charge moment change (−190 C km) produced by the main stroke was not sufficient to induce conventional breakdown in the mesosphere, a second impulse charge transfer occurred with ~2 ms delay to cause a substantial charge transfer (−290 C km) so that the overall charge moment change (−480 C km) exceeded the threshold for sprite production; this is a scenario different from the typical case discussed by Li et al. (2012). As for the source of the second current pulse that played a critical role to produce the sprite, it could be an M component whose charge source was at least 9 km horizontally displaced from the main stroke or a negative CG stroke (with weak peak current for the return stroke) that was not detected by the NLDN.

Radio emissions continue to provide insight into the production of terrestrial gamma ray flashes (TGFs) by thunderstorms, including the critical question of the conditions under which they are generated. We have identified several TGF-associated lightning radio emissions in which the altitudes of in-cloud lightning leader pulses that precede and follow the TGF can be measured. We combine these with high absolute timing accuracy TGF observations from the Fermi satellite to determine the development of the lightning channel before, during, and after the TGF production. All of these TGFs were produced several milliseconds after the leader had initiated and when the leaders reached 1–2 km in length. After the TGFs, the leaders all continued to ascend for several more kilometers with no dramatic change in their characteristics, although they all exhibited high average velocities of 0.8–1.0 × 106 m/s. Implications in the context of TGF models are discussed. These results paint the first clear picture of the lightning processes that occur before, during, and after TGF production.

We investigated National Lightning Detection Network reports and lightning radio waveforms in a 44 day observation period to analyze the in-cloud (IC) events producing currents above 200 kA. The results show that there are two distinct classes of IC lightning events with very high peak currents: the well-known narrow bipolar events, and a previously unreported type that we call energetic in-cloud pulses (EIPs). Their temporal and spatial context shows that EIPs are generated from existing negative polarity leaders that are propagating usually upward but sometimes downward. The nearly identical characteristics of EIPs and some previously reported terrestrial gamma ray flashes (TGFs) indicate a likely connection between the two, which further suggests the possibility of downward directed TGFs. These very high peak current IC events also suggest the association of EIPs with ionospheric perturbations and optical emissions known as elves.

Many details of how thunderstorms generate terrestrial gamma ray flashes (TGFs) and other forms of high-energy radiation remain uncertain, including the basic question of where they are produced. We exploit the association of distinct low-frequency radio emissions with generation of terrestrial gamma ray flashes (TGFs) to directly measure for the first time the TGF source altitude. Analysis of two events reveals source altitudes of 11.8 ± 0.4 km and 11.9 ± 0.9 km. This places the source region in the interior of the thunderstorm between the two main charge layers and implies an intrinsic TGF brightness of approximately 10 18 runaway electrons. The electric current in this nontraditional lightning process is found to be strong enough to drive nonlinear effects in the ionosphere, and in one case is comparable to the highest peak current lightning processes on the planet.

We examined the broadband (< 1 Hz to 30 kHz) lightning sferics associated with 395 sprites observed near North America by the Imager of Sprites and Upper Atmospheric Lightning (ISUAL) onboard the FORMOSAT-2 satellite in a 12-year period from 2004 - 2015. Our analysis indicates that the ISUAL dataset contains a significant fraction (69, or ~18%) of negative sprites, which were predominantly (> 80%) observed over oceanic and coastal thunderstorms mostly in tropical areas. The mean and median of impulse charge moment change (iCMC) associated with positive (negative) sprites are +346 C km (-709 C km) and +280 C km (-649 C km), respectively. The morphology and parent lightning properties (e.g., typically with high peak currents > -80 kA and large iCMCs > -300 C km) of negative sprites observed by the ISUAL are generally consistent with that documented in ground-based observations, but the ISUAL dataset does imply that sprites are sometimes produced by negative strokes with sub-critical iCMCs (less than -300 C km). Consequently, the future survey of global occurrence of sprites is desired to be based on complementary ground and space-borne observations.

We report on the development of an easily deployable LF near-field interferometric-time of arrival (TOA) 3-D Lightning Mapping Array applied to imaging of entire lightning flashes. An interferometric cross-correlation technique is applied in our system to compute windowed two-sensor time differences with submicrosecond time resolution before TOA is used for source location. Compared to previously reported LF lightning location systems, our system captures many more LF sources. This is due mainly to the improved mapping of continuous lightning processes by using this type of hybrid interferometry/TOA processing method. We show with five station measurements that the array detects and maps different lightning processes, such as stepped and dart leaders, during both in-cloud and cloud-to-ground flashes. Lightning images mapped by our LF system are remarkably similar to those created by VHF mapping systems, which may suggest some special links between LF and VHF emission during lightning processes.