Preliminary Breakdown Pulses in Lightning

• PBPs are short, high-amplitude bipolar pulses marking the initial phase of lightning breakdown, crucial for understanding leader development.
• They are classified into Classical and Narrow types based on duration and morphology, linking pulse characteristics to specific initiation stages.
• Electrodynamic models correlate PBPs with rapid current surges and corona dynamics, offering quantitative diagnostics for lightning channel evolution.

Preliminary Breakdown Pulses (PBPs) are short, intense electromagnetic or electric-field pulses emitted during the initial breakdown stage of natural lightning, before the first return stroke. In downward negative cloud-to-ground flashes they are the impulsive electric-field changes that mark the initial stages of leader development and are traditionally associated with the early Breakdown stage in the BIL sequence proposed by Clarence and Malan (1957). In parts of the lightning literature, especially in work emphasizing initiation physics, the same phenomena are termed initial breakdown pulses (IBPs). Across these usages, PBPs denote discrete signatures of early charge acceleration, leader stepping, and related fast electrodynamic reconfiguration of the channel (Carlson et al., 2016).

1. Terminology, scope, and classification

In the broader lightning literature, PBPs are the initial, high-amplitude, bipolar electric-field pulses that occur soon after the very first electromagnetic signature of a flash. In the terminology of one qualitative initiation model, these pulses are called IBPs, and “classic IBPs” are defined as the larger subset of IBPs with durations $\ge 10~\mu\text{s}$ and amplitudes $\ge 25\%$ of the largest IBP in a given flash (Kostinskiy et al., 2019).

A widely used observational distinction separates “Classical” PBPs from “Narrow” PBPs. “Classical” PBPs are typically bipolar electric-field signatures, with typical durations on the order of $20$–$40~\mu\text{s}$ in many cloud-to-ground studies and, more broadly, $5$–$100+~\mu\text{s}$ with mean $\sim 20~\mu\text{s}$ in literature cited in recent comparative work. They are predominantly observed during the Breakdown stage. “Narrow” PBPs can be unipolar or bipolar, are significantly shorter, and in the literature are classified as $<2~\mu\text{s}$, sometimes $<4~\mu\text{s}$; they are commonly observed during the Leader stage and can also appear superimposed on Classical PBPs during Breakdown (Oregel-Chaumont et al., 13 Aug 2025).

Class	Morphology and duration	Stage association
Classical PBP	Typically bipolar; on the order of $20$–$\ge 25\%$0 in many CG studies; more broadly $\ge 25\%$1–$\ge 25\%$2	Predominantly Breakdown stage
Narrow PBP	Unipolar or bipolar; $\ge 25\%$3 in one literature criterion, sometimes $\ge 25\%$4	Commonly Leader stage; can be superimposed on Classical PBPs

A recent upward-lightning comparison maps Category A upward pulses to Classical PBPs and Category B upward pulses to Narrow PBPs. In that mapping, Category A pulses are bipolar electric-field pulses with microsecond-scale risetimes and half-widths and are correlated with tower-base current pulses, whereas Category B pulses have sub-microsecond risetime and FWHM, occur later in the flash, and typically show no correlating tower-base current (Oregel-Chaumont et al., 13 Aug 2025).

2. Observed waveform structure and measurement signatures

PBPs recorded by the Huntsville Alabama Marx Meter Array typically exhibit a characteristic pulse shape: a short, intense initial excursion in the measured vertical electric field change $\ge 25\%$5, generally of one polarity, immediately followed by a weaker, longer excursion of opposite polarity. Nearby sensors often register a stepwise DC change superposed on the fast pulse, consistent with net charge transport near the sensor. Pulse trains are common: in the event of 26 October 2010 at 19:04:59 UT, a trio of pulses appeared within roughly $\ge 25\%$6, from $\ge 25\%$7 to $\ge 25\%$8, with inter-pulse spacings on the order of a few tens of microseconds. Geometry strongly affects the recorded waveform: sensors at different azimuths and ranges relative to the channel tip can record different DC changes and sometimes opposite fast-pulse polarity, consistent with dipole-like geometry and leader extension direction. The fast radiation component remains visible at large ranges with $\ge 25\%$9 scaling, whereas the quasi-static DC change decays more rapidly with distance and depends on the motion of charge relative to the sensor (Carlson et al., 2016).

In the HAMMA observations summarized in the same work, PBPs commonly show the negative-first, positive-later pattern in $20$0 at moderate and large ranges, while at close range such as $20$1 stepwise DC changes can be substantial. In the analyzed event, detector 5 at $20$2 saw positive DC steps, while detectors at $20$3 and $20$4 saw net negative DC changes. A trio at $20$5–$20$6 in detector 5 exhibited differing DC offsets and fast-pulse amplitudes, consistent with successive steps directed at differing azimuths (Carlson et al., 2016).

The Säntis tower study provides complementary microsecond-scale statistics through simultaneous channel-base current and electric-field measurements of upward pulses explicitly compared with PBPs. For Category A pulses, measured over $20$7 pulses, the electric-field parameters normalized to $20$8 were $20$9, $40~\mu\text{s}$0, first half-cycle peak $40~\mu\text{s}$1, and maximum slope $40~\mu\text{s}$2. The associated current parameters were $40~\mu\text{s}$3, $40~\mu\text{s}$4, peak current $40~\mu\text{s}$5, maximum slope $40~\mu\text{s}$6, and total pulse duration $40~\mu\text{s}$7. For Category B pulses, measured over $40~\mu\text{s}$8 pulses, the electric-field values were $40~\mu\text{s}$9, $5$0, $5$1, and $5$2 (Oregel-Chaumont et al., 13 Aug 2025).

Polarity statistics in that upward analogue are also notable: about $5$3 of bipolar Category A electric-field pulses start with a positive half-cycle and about $5$4 are inverted, a distribution stated to be consistent with prior PBP observations in downward flashes. The same study reports a bimodal temporal distribution, with an early phase containing both Category A and B pulses and a later phase dominated by Category B pulses, separated by a quiet Intermediate phase. This was noted as similar to bimodal stepping distributions reported in downward flashes (Oregel-Chaumont et al., 13 Aug 2025).

3. Electrodynamic description as a signature of stepped leader extension

A detailed time-domain electrodynamic model treats the lightning channel as a thin, conducting wire embedded in three-dimensional space, carrying line current $5$5 and line charge density $5$6 along a known path parameterized by arc length $5$7. The electric field at an observation point is computed from the time-domain electric field integral equation, retaining retarded-time effects and explicitly separating radiation, induction, and electrostatic contributions. In thin-wire form, the channel field is written as

$5$8

with charge continuity enforced by

$5$9

The singular self-terms are regularized by replacing $100+~\mu\text{s}$0 with $100+~\mu\text{s}$1, where the effective channel radius is taken as $100+~\mu\text{s}$2. The channel is segmented into straight current segments of length $100+~\mu\text{s}$3, time is discretized with $100+~\mu\text{s}$4, and the resulting sparse linear system is solved with sparse matrix methods using CXSparse and parallelized ScaLAPACK/MPI (Carlson et al., 2016).

Within this framework, PBPs arise from the electrodynamics of stepwise leader-channel extension inside the cloud. A new segment of the conducting channel rapidly heats, its conductivity rises, current surges onto the new segment, and charge migrates outward from the channel to a corona sheath. The turn-on of current produces a strong, fast radiation pulse, and the slower filling of the corona sheath produces a slower turn-off as the current relaxes. Ohm’s law is imposed locally along the wire, $100+~\mu\text{s}$5, with resistance per unit length $100+~\mu\text{s}$6 (Carlson et al., 2016).

Resistance evolution is modeled through Joule heating on sub-$100+~\mu\text{s}$7 timescales, with cooling neglected:

$100+~\mu\text{s}$8

where $100+~\mu\text{s}$9. An approximate conductivity model motivated by the Saha equation and nonideal plasma transport gives

$\sim 20~\mu\text{s}$0

with $\sim 20~\mu\text{s}$1, so that $\sim 20~\mu\text{s}$2. The proportionality is anchored by $\sim 20~\mu\text{s}$3. In the representative simulations, the pre-existing main channel is set to $\sim 20~\mu\text{s}$4, while the new step begins at about $\sim 20~\mu\text{s}$5, corresponding to $\sim 20~\mu\text{s}$6, so that it can self-heat into conduction (Carlson et al., 2016).

Charge transfer to a corona sheath is represented by a second set of sheath charges co-located with each channel charge segment but using an effective radius $\sim 20~\mu\text{s}$7 and no longitudinal current. Outward charge migration is modeled as an exponential with characteristic timescale $\sim 20~\mu\text{s}$8:

$\sim 20~\mu\text{s}$9

The simulated geometry uses a straight, unbranched $<2~\mu\text{s}$0 leader with its bottom at $<2~\mu\text{s}$1 altitude, allowed to reach quasi-static equilibrium under a uniform applied field $<2~\mu\text{s}$2 along the channel. A new $<2~\mu\text{s}$3 step is appended at the tip, and its resistance is enforced to be uniform along its length at each time step to emulate space-stem-like behavior without assuming detailed pre-ionization structure (Carlson et al., 2016).

This model reproduces the broad features of PBPs measured by HAMMA: a strong, fast initial excursion driven by the rapid increase of current onto the newly conducting step; a subsequent weaker, longer opposite-polarity excursion as current decays; and a DC offset whose sign and magnitude depend on the net direction of charge motion relative to the sensor. The far-field radiation is dominated by the EFIE radiation term proportional to $<2~\mu\text{s}$4, giving

$<2~\mu\text{s}$5

so that $<2~\mu\text{s}$6 (Carlson et al., 2016).

4. Competing and complementary physical interpretations

The stepped-wire electrodynamic model explains PBPs as the electromagnetic signatures of impulsive leader extension driven by channel heating and current surge, with corona-sheath formation governing the subsequent relaxation. In that interpretation, the fast initial excursion is the impulsive turn-on of current as the step heats, the slower opposite-polarity excursion is the turn-off as the corona sheath fills and reduces the driving field, and the polarity and DC offset depend on how net charge transport is oriented relative to the observer (Carlson et al., 2016).

A distinct qualitative initiation mechanism proposes a more distributed origin for PBPs. In that framework, lightning initiation occurs within an “EE-volume” of order $<2~\mu\text{s}$7–$<2~\mu\text{s}$8 where the average density-scaled electric field is $<2~\mu\text{s}$9–$<4~\mu\text{s}$0 and where turbulence creates many small “Eth-volumes” with local field $<4~\mu\text{s}$1. Extensive air showers provide secondary particles that, in a sufficiently strong ambient field, undergo relativistic runaway electron avalanching and pass through many Eth-volumes in $<4~\mu\text{s}$2–$<4~\mu\text{s}$3, thereby triggering many positive streamer flashes nearly simultaneously. Those streamers then develop unusual plasma formations (UPFs) through ionization-heating instability, and the subsequent merging of three-dimensional UPF networks is proposed to produce the first and later IBPs/PBPs (Kostinskiy et al., 2019).

In that mechanism, the streamer-to-UPF transition is associated with contraction of cold plasma channels from about $<4~\mu\text{s}$4 to about $<4~\mu\text{s}$5–$<4~\mu\text{s}$6, with temperatures rising to $<4~\mu\text{s}$7–$<4~\mu\text{s}$8. The instability develops in about $<4~\mu\text{s}$9–$20$0 at $20$1 and scales approximately as $20$2 with $20$3 at altitude. UPFs are short hot channels, initially about $20$4–$20$5 long, and they can merge into chains and networks when the local inter-UPF field exceeds the positive streamer threshold $20$6–$20$7. Minimal per-channel current for UPF survival is stated as $20$8, while currents through UPF chains are frequently $20$9–$\ge 25\%$00 (Kostinskiy et al., 2019).

The same proposal ties classic IBPs to discrete mergers of large UPF networks. The first classic pulse occurs when two large networks at opposite ends launch bidirectional leaders that contact and merge, producing a breakthrough phase followed by a miniature return stroke. Subsequent classic pulses occur when another three-dimensional UPF network connects to the already-formed chain. This account is intended to explain large bipolar fast-antenna pulses, strong VHF emission, bright optical bursts, and inter-pulse spacing commonly in the hundreds of microseconds (Kostinskiy et al., 2019).

A further refinement comes from upward-lightning observations at Säntis. There, Category A pulses are interpreted as stepping of an upward negative leader and are used as analogues of Classical PBPs. Category B pulses lack channel-base current correlation, are faster, and in one flash were temporally coincident with a downward-propagating recoil leader retracing a pre-existing channel. This suggests that at least some Narrow PBP-like fast pulses can be produced by recoil leader activity. The same study explicitly argues that Narrow PBPs in downward flashes could be similarly produced by recoil leader activity, but that statement is presented as an interpretation rather than a direct downward-flash measurement (Oregel-Chaumont et al., 13 Aug 2025).

5. Quantitative relations, sensitivities, and diagnostic content

PBPs encode several physically distinct components of the discharge. In the thin-wire model, the radiation amplitude scales with current slew rate according to $\ge 25\%$01, while the induction and electrostatic terms determine the DC and near-field components. This separates fast-radiation observables from quasi-static charge-transport observables and provides a direct link between waveform morphology and channel microphysics (Carlson et al., 2016).

Parameter studies in the same simulation identify four key controls on single-step waveform features. Increasing the channel heat capacity per unit length $\ge 25\%$02 lengthens the negative excursion, reduces peak amplitude, and increases asymmetry. Increasing the corona-sheath timescale $\ge 25\%$03 reduces peak amplitude and increases asymmetry without strongly affecting the initial rise duration. Increasing the step length yields longer pulses of similar peak amplitude and increased symmetry. Increasing the applied field $\ge 25\%$04 increases amplitude, reduces duration, and reduces asymmetry. The representative heating and relaxation timescales are likewise separated: in the model the step turns on over tens of microseconds, with the onset of high current near about $\ge 25\%$05 in a representative $\ge 25\%$06 step, while $\ge 25\%$07 controls how quickly the current rolls off and the degree of asymmetry (Carlson et al., 2016).

The Säntis measurements add a direct field-current scaling. Using bipolar Category A pulses and correcting for the known topographic enhancement factor $\ge 25\%$08 between Mt. Säntis and Herisau, the reported ratio is

$\ge 25\%$09

while the Kaspar et al. (2017) model prediction quoted in that study is

$\ge 25\%$10

The peak electric field versus peak current correlation for Category A pulses is strong, with $\ge 25\%$11, $\ge 25\%$12, $\ge 25\%$13, and best-fit $\ge 25\%$14. For Category A morphology alone, first versus second half-cycle widths show $\ge 25\%$15, $\ge 25\%$16, $\ge 25\%$17, and first versus second half-cycle amplitudes show $\ge 25\%$18, $\ge 25\%$19, $\ge 25\%$20. Category B pulses exhibit a very strong linear relation between amplitude and slope, with $\ge 25\%$21, $\ge 25\%$22, $\ge 25\%$23, implying a minimum characteristic risetime $\ge 25\%$24 and reported $\ge 25\%$25 (Oregel-Chaumont et al., 13 Aug 2025).

This suggests that PBP rise time and peak amplitude can serve as diagnostics of effective heating rate and current turn-on, whereas asymmetry and the delayed opposite-polarity excursion are especially sensitive to corona dynamics. The same inference is stated explicitly in the time-domain simulation study: $\ge 25\%$26 and rise time diagnose the leader’s effective heating rate through parameters such as $\ge 25\%$27 and $\ge 25\%$28, while asymmetry and the positive excursion diagnose corona dynamics through $\ge 25\%$29 (Carlson et al., 2016).

6. Limitations, unresolved questions, and broader usage of the term

Current PBP models remain deliberately simplified. In the time-domain electrodynamic treatment, the channel geometry is fixed, the thin-wire approximation is used, the ground is treated as perfectly conducting, the step resistance is artificially enforced uniform over its length during evolution, cooling is neglected on sub-$\ge 25\%$30 timescales, and the corona sheath is represented as a co-located capacitance-like structure with radius $\ge 25\%$31 and timescale $\ge 25\%$32 but no longitudinal current or explicit streamer dynamics. These approximations affect pulse-shape accuracy, especially the timing and height of the positive excursion and some DC offsets, and limit the ability to reproduce branching and complex multi-step overlap (Carlson et al., 2016).

The discrepancies between model and data are specific. In observations, the positive excursion often peaks later and higher than in the simulation, implying that in nature the current remains near its peak longer and then decays faster than the present corona model produces. Some DC changes also disagree in sign or magnitude with simple single-step simulations, indicating concurrent charge transfers elsewhere on a branched channel or different step directions. Proposed refinements include more realistic stepping physics, non-uniform initial resistance or temperature along the step, improved corona dynamics, explicit streamer velocity fields, longitudinal sheath currents where appropriate, better temperature-dependent conductivity microphysics, radiative and conductive cooling on longer timescales, and complex channel branching and tortuosity (Carlson et al., 2016).

The upward-lightning comparison has its own limitations. The dataset consisted of $\ge 25\%$33 upward positive flashes, with $\ge 25\%$34 Category A pulses and $\ge 25\%$35 Category B pulses. Only one flash provided usable high-speed-camera data for direct step-length estimation and recoil-leader identification, and no speeds were extracted. The downward-flash comparison relies on literature rather than simultaneous downward current measurements. Accordingly, the proposed linkage between Narrow PBPs and recoil leaders is suggestive but not universalized in the reported measurements (Oregel-Chaumont et al., 13 Aug 2025).

The qualitative IBP mechanism also identifies open questions. It relies on abundant small-scale Eth-volumes generated by turbulence, and the number density and lifetimes of those Eth-volumes remain to be quantitatively verified. The proposal likewise depends on uncertain line charge densities for UPFs and early hot channels, as well as incompletely constrained conductivity and heating-cooling rates in evolving hot plasma formations (Kostinskiy et al., 2019).

Outside lightning physics, the term “preliminary breakdown pulses” has also been used in a different sense for capacitor-assisted dielectric breakdown of a nanoscopic NbOx layer. In that context, PBPs denote transient, short-lived electrical breakdown events that occur in the dielectric nanolayer before the final, stable, low-resistance conducting channel is established. They were inferred from abrupt voltage collapse, capacitor discharge, and cathode-film morphological damage, with a characteristic discharge time $\ge 25\%$36 for $\ge 25\%$37 and $\ge 25\%$38, and an initial breakdown timescale of order $\ge 25\%$39. This usage is physically distinct from lightning PBPs and underscores that the acronym is context-dependent (Krevsun et al., 2022).