AE31A-0151 0800h
Transient luminous jets recorded in the Taiwan 2004 TLE campaign
The gigantic jets were first reported by Su et al. [Su et al., Nature, 423, 974-976, 2003]. After nearly two years in hunting, another gigantic jet was recorded from Lulin observatory (23.47N, 120.87E, 2862m) on June 18, 2004. The newly observed event occurred above a frontal system which was stationed at Anhui province of China about 700km away. This fact indicated that gigantic jets could also occur over land. Moreover, during the 2004 Taiwan TLE campaign, several jet-like TLEs were recorded at Kenting, Taiwan (21.92N,120.85E, 116m) on the night of August 3, 2004. These events occurred over a thunderstorm located at the coast of Guandong province, China about 500km away. The luminous duration of these events lasted for ~200ms, which is shorter than the GJs and the large blue jet observed by Pasko et al [Nature, 416 , 152-154, 2002]. The upper altitude of the jet-like TLEs is ~70km, which is lower than the GJs but is the same as the large blue-jet or the trailing jets of GJs. The luminous structure of these events looks like a fountain in the dark, which is very similar to trailing jets of GJs but is different from that of the large blue jet. In the 2004 TLE campaign, we deployed a high speed ICCD with 4ms time resolution and three NTSC ICCD cameras that were equipped with different filters. In this talk, the temporal evolution of the jets, the associated electromagnetic emissions, and their spectral properties will be presented. Comparisons with the existing theoretical models will also be presented. *Work supported in part by grants from NSPO (93-NSPO(B)-ISUAL-FA09-01) and NSC (NSC93-2112-M-006-007, NSC93-2111-M-006-001) in Taiwan.
AE31A-0152 0800h
Characteristics of Sprites in Relation to Characteristics of Lightning Determined by the OK Lightning Mapping Array and the NLDN
During TELEX II in May and June of 2004, we launched balloon-borne systems to observe X-rays and electric-field changes (ΔE) in storms in central Oklahoma and deployed ground-based low-light cameras in Clayton, NM to observe TLE that occurred over the storms. Our goal was to obtain contemporaneous observations of E, ΔE, X-rays, and TLE for storms within the purview of the Oklahoma Lightning-Mapping Array (OKLMA). Our observations were motivated by the work of Roussel-Dupre and Gurevich (1996) on runaway breakdown and upward propagating discharges, which suggested that runaway breakdown is the driving mechanism for TLE and that the discharges would be sources of observable RF and γ or x-ray emissions. Indeed, they suggested that confirmation could be attained by measuring optical, γ /X-ray, and RF emissions (ΔE) simultaneously. On three occasions we launched instruments to record ΔE and X-rays in storms and we obtained observations of Sprites on 5 days during the program. Unfortunately, both Clayton, NM and Fort Collins were overcast on days when we obtained ΔE/X-ray soundings. However, we did obtain contemporaneous observations of Sprites, discharges mapped by the OKLMA, and NLDN CG flash locations. We compare the characteristics of lightning discharges associated with observed Sprites and those not associated with observed Sprites, in the context of runaway hypotheses.
AE31A-0153 0800h
Statistical Characterizations of TLEs and Their Parent Lightning Discharges
The characteristics of sprites, halos and elves (collectively termed Transient Luminous Events [TLEs]) and their parent lightning discharges have been documented for individual storms (Lyons 1996; Lyons et al. 2000; 2004). Several summer observation campaigns conducted at the Yucca Ridge Field Station (YRFS), notably the Severe Thunderstorm Electrification and Precipitation Study (STEPS) of 2000 (Lang et al. 2004), have allowed the determination and analysis of the characteristics of a large number (>1000) of TLEs and their parent CG strokes. The optical characteristics of the TLEs are now being linked to various relevant parameters including NLDN polarity and peak currents, charge moment changes (CMCs) determined from ELV/VLF signatures. (Cummer and Lyons 2004), LMA-determined altitudes and areas of charge removal, and storm morphological characteristics (cloud heights, areal coverage, reflectivity, age, etc.) The assembly of this large database is an ongoing process that will take another year for completion. However, some preliminary results will be presented. The geographical distribution of TLEs optically detected from YRFS will be mapped. The distribution of sprite parent +CGs shows a broad range of peak currents from 9 kA to 237 kA (mean of 58 kA) while TLEs with halos and/or elves ranged from 11 kA to 266 kA (mean of 112 kA). This is consistent with previous studies indicating peak currents for halos/elves are roughly twice that for sprites, and both are significantly larger than the average +CG value in the storm. However, given its variability, peak current is not a good predictor of TLE potential. Various approaches to computing CG charge moment changes are showing a clear separation in distributions between TLE and non-TLE producing +CGs. Statistics on time of occurrence, sprite morphological relationships to CMCs, and sprite onset time delays will also be presented. Given the temporal imprecision of video, the TLE onset delay after the return stroke can only be estimated. Approximately one third of TLEs onset within the first 17 ms, though almost 20% delay for 60 ms or longer. No High Plains sprites associated with negative CGs have been confirmed, though there are some negative CG halo candidates. No blue jets have been monitored from the ground at YRFS in a decade. Is this due to (1) difficulties in monitoring a blue signature through >100 km of atmosphere, and/or or (2) the rarity of this phenomenon over the U.S. High Plains? The unexpectedly high frequency of large peak current (>200 kA) negative (but not positive) CGs detected by the NLDN over tropical oceanic regions, an environment in which giant blue jets have been observed, has lead to questions as to whether such large peak currents are real or an artifact (Lyons et al. 1998). A climatology of large peak current CGs for the New Zealand Lightning Detection Network found some evidence of enhanced large peak current -CGs over surrounding oceanic regions. The pattern, however, is complicated by apparent enhancements in high peak current CGs of both polarities in mountainous regions, an affect not documented by the U.S. NLDN.
AE31A-0154 0800h
Numerical Study on Branching Properties of Sprite Streamers
The complex morphology of sprites, dominated by highly branched and predominantly vertical filamentary structures, is well documented in the existing literature [e.g., {\it Gerken and Inan, JASTP, 65,} 567, 2003; {\it Moudry et al., JASTP, 65,} 509, 2003], and has recently been interpreted in terms of thin channels of ionization called streamers [{\it Liu and Pasko, JGR, 109,} A04301, 2004, and references therein]. The commonly observed, but not yet fully understood, tree-like structures in sprites are likely related to the branching properties of individual streamer channels. The branching phenomenon has received increasing attention in recent experimental and modeling studies of streamers at ground pressure [e.g., {\it Arrayas et al., PRL, 88,} 174502(R), 2002; {\it Yi and Williams, JPD (Appl. Phys.), 35,} 205, 2002; {\it Hallac et al., JPD, 36,} 2498, 2003]. However, there is still a disagreement on whether the branching of model streamers is a pseudo-branching (i.e. a consequence of numerical instability) or just correctly reflects the real splitting physics [{\it Pancheshnyi and Starikovskii, JPD, 36,} 2683, 2003]. Furthermore, the modeling results on branching morphology reported by different research groups remain highly controversial [{\it Kulikovsky, JPD, 33,} 1514, 2000; {\it Arrayas et al.,} 2002; {\it Rocco et al., PRE, 66,} 035102(R), 2002; {\it Hallac et al.,} 2003; {\it Liu and Pasko,} 2004; and references therein]. The purpose of this study is to clarify the physical nature of the streamer branching by means of numerical simulations using several numerical schemes. For this study we have developed a minimal model of streamers in a non-attaching nitrogen-like gas. The model allows studies of streamers propagating between planar electrodes in a strong applied electric field, and is similar to the model used by {\it Arrayas et al.} [2002]. Although this model only accounts for the effects of the ionization of neutral molecules and the transport of electrons, it has an advantage by minimizing the total number of involved physical processes and by allowing us to investigate more clearly the variations in branching morphology obtained by using different numerical schemes. In this talk we report modeling results of branching as a function of magnitudes of applied electric field and medium pre-ionization using the following numerical schemes to calculate the electron flux: (1) the 1st order upwind, (2) the 2nd order, (3) the 3rd order upwind-biased, (4) the improved Scharfetter-Gummel (ISG), and (5) the flux-corrected transport (FCT). In all studied cases, we observe the same pre-branching features of streamers (i.e., extremely high peak field and electron density in the head with a weak curvature) as reported previously by {\it Arrayas et al.} [2002], {\it Rocco et al.} [2002] and {\it Liu and Pasko} [2004]. The results obtained by different numerical schemes show slight variations in onset time and morphology of branching; however, it is shown in all considered cases that the branching is very sensitive to ambient electron density and that the expanding streamers can reach a branching stage for a wide range of applied electric field. These results agree with the earlier suggestion by {\it Liu and Pasko} [2004]. The results on two distinct types of branching morphology observed in our numerical experiments and their relationship to physical and numerical parameters of the model will be presented.
AE31A-0155 0800h
Three Dimensional Fully Electromagnetic Simulations of Sprites Including the Effects of Runaway Breakdown
In an effort to simulate the detailed spatial structure of sprites and other high-altitude discharges, a three dimensional, fully electromagnetic, hydrodynamic code has been written. The hydrodynamic equations for four species (relativistic electrons, secondary electrons, negative ions, and positive ions) are solved in time along with Maxwell's equations. Both an externally applied electric field resulting from thunderstorm processes (e.g., a temporally varying field produced by thunderstorm electrification and/or by a parent lightning discharge) and a magnetic field (i.e., the Earth's magnetic field) are included. Seed electrons produced by cosmic rays accelerate and avalanche in the external field and develop into a runaway discharge and/or a conventional streamer. In three dimensions it is possible to simulate the development of compact streamers and to investigate the effect of spatially compact and statistically random (on the time scales and spatial scales of interest) cosmic ray showers. The implications for our understanding of the small-scale structures observed in sprite-like discharges are discussed. The corresponding optical, radio, and X-ray emissions for two simulations are presented for comparison with observations.
AE31A-0156 0800h
Parallelization of the PLUME and UNIMAX Computer Codes
PLUME computes the evolution of an electron distribution function through the relativistic Fokker-Planck equation and continuity equations. Features include transport, drag, scattering, production of secondary electrons by primary ionization, recombination reactions, and an applied external electric field. Geophysical applications include sprites. The UNIMAX code is used for lightning modeling; it computes breakdown and avalanche processes for cosmic-ray initiated electron showers in an electric field. Explicit finite difference numerical algorithms are used for both codes. Options for numerical treatment of fluxes range from simple donor differencing to higher order Godunov. The PLUME code employs two space dimensions plus time plus two phase space dimensions; consequently, computer simulations are very computationally intensive. In addition, for UNIMAX to accurately follow the development of lightning, extremely fine gridding is required. To extend the range of applicability of these codes, we are parallelizing them. Parallelization is achieved through the MPI library. Performance of the parallelized codes vs. problem size and number of CPUs is discussed.
AE31A-0157 0800h
Simulation of Streamer Initiation using a Particle in Cell Code with Monte Carlo Collisions : Application to Sprite ignition.
A particle code designed to simulate electrical breakdown of air is presented. By the use of 2D axisymetrical coordinates, the code can simulate 3D development of an electron avalanche, its transition into a streamer and the beginning of the streamer propagation. The method employed is a classical Particle in Cell method plus Monte Carlo treatment of collisions. The particles are followed inside a Cartesian mesh, collision processes are performed randomly accordingly cross sections and the self-consistent electric field is calculated. The use of a particle method gives us a kinetic description of the electron population; in particular we can obtain the electronic distribution function in the head of the streamer where the electric field varies strongly. The mobility, mean energy, and every considered collisional events rates are calculated without the assumption of locally constant electric field commonly used in fluid simulations. Some results of streamer development under atmospheric condition of sprite ignition are presented. The use of such simulation makes possible the comparison with laboratory discharge experiments or sprite observation campaigns by recovering informations on optical emissions for example.
AE31A-0158 0800h
Monte Carlo Model for Analysis of Runaway Electrons in Streamer Tips in Sprites
The observed filamentary structures in sprites [Gerken and Inan, JASTP, 65, 567, 2003] have recently been interpreted in terms of thin channels of ionization called streamers, which exhibit acceleration, expansion and branching [Liu and Pasko, JGR, 109, A04301, 2004]. The expanding streamers can reach a branching state for a wide range of applied electric fields and an extremely high peak field can be generated in the streamer tip immediately preceding the branching [Liu and Pasko, 2004]. The acceleration of electrons in tips of highly overvolted streamers [Babich, Sov. Phys. Dokl., 27, 215, 1982] has been proposed for interpretation of X-ray radiation observed in experiments reported by Tarasova et al. [Sov. Phys. Tech. Phys., 19, 351, 1974]. It has also been proposed that the high electric fields in streamer tips can accelerate electrons to energies of several keV, initiating electron runaway in relatively low ambient electric fields E$\sim$Ek [Pasko et al., GRL, 25, 2123, 1998], where Ek is the conventional breakdown threshold field. The recently reported X-ray emissions observed during the approach to ground stage of natural [Moore et al., GRL, 28, 2141, 2001] and triggered [Dwyer et al., Science, 299, 694, 2003; GRL, 31, L05118, 2004] lightning leaders may be related to the enhancement of electric field in the leader streamer zone [Dwyer et al., GRL, 31, L12102, 2004] leading to the generation of runaway electrons in streamer tips. In this talk we report results from a Monte Carlo model, which is capable of describing electron dynamics in air (including the runaway phenomena) under influence of an external electric field of arbitrary strength. The model is similar in technical details to the model previously developed for N2 by Tzeng and Kunhardt [Phys. Rev. A, 34, 2148, 1986] and Kunhardt and Tzeng [Phys. Rev. A, 34, 2158, 1986], and incorporates the following features: (1) the null collision method to determine time between collisions; (2) the remapping of the electron assembly to improve statistics for the high-energy tail of the electron distribution; (3) the differential ionization and scattering cross sections for realistic description of energy spectrum of secondary electrons and the forward scattering properties of electrons at high energies. The elastic collision cross sections of electrons on N2, O2 and Ar are determined from the most recent total cross sections available in the literature and forty-three inelastic collision cross sections obtained from A.V. Phelps [http://jilawww.colorado.edu/www/research/colldata.html]. At high electric fields ($>$5Ek) the model results are validated by comparisons with studies conducted for N2 by Tzeng and Kunhardt [1986] and more recently by Bakhov et al. [IEEE Trans. Plasma Sci., 28, 1254, 2000], and at low fields ($<$5Ek) by comparisons with available data from swarm experiments in air and solutions of Boltzmann equation based on two-term spherical harmonic expansion of the electron distribution function [Morgan and Penetrante, Comp. Phys. Comm., 58, 127, 1990].
AE31A-0159 0800h
First Principles Calculations of Excitation Rates in the Presence of an Electric Field
In this paper we describe detailed 4.5 dimensional (two spatial dimensions and 2.5 phase space dimensions) kinetic calculations for the electron distribution function resulting from the injection of energetic electrons into air. Various pressures and applied electric field intensities are investigated. From the electron distribution function and measured excitation cross-sections we then compute the optical efficiencies for a large number of nitrogen and oxygen lines across the electromagnetic spectrum from 320.0 nm to 800.0 nm. The fluorescence efficiencies used in calculating the optical emissions produced by energetic particles and penetrating radiation in air are derived for the most part from the measurements of Davidson and O'neil ({\it J. Chem. Phys.}, {\bf 41}, No. 12, 3946, 1964), Mitchell (J. Chem. Phys., 53, No. 5, 1795, 1970) and Hartman ({\it Planet. Spa. Sci.}, {\bf 16}, 1315, 1968). These efficiencies were obtained from experiments conducted at various air pressures and in the absence of an applied electric field. We choose beam parameters and dimensions that are directly relevant to the original Davidson and O'neil experiments and present comparisons to these measurements. The computed fluorescence efficiencies can be used to determine the optical emissions associated with high-altitude discharges driven by runaway air breakdown and results are discussed in a separate presentation at this meeting (Triplett {\it et al.}). Simulations are also run for the acceleration of `Thermal' electrons in the presence of an applied electric field and we have recalculated optical emission rates that are applicable to discharges dominated by conventional breakdown for comparison with Taranenko et al. ({\it GRL}, {\bf 19}, No. 18, 1815, 1992).
AE31A-0160 0800h
Heating and luminous emission of upper atmosphere due to an applied DC E-field
TLEs are transient and luminous events occurring between thunderstorm top and the lower ionosphere. The optical emission typically lasts for several milliseconds and is from the radiative emission of excited air molecules (mostly nitrogen and oxygen). The excitation of air molecules is due to the residual electric field above thundercloud that accelerates electrons and the resultant inelastic collisions of the accelerating electrons with molecules. Based on a Boltzmann-type distribution of electron energy, heating and luminous emission of air molecules in an applied DC electric filed was computed. The computed quantities are further compared with the electron-swarm properties of air, and the average electron energy as well as the strength of the causative electric field deduced from the TLE events observed by ROCSAT-2 ISUAL and by the 2004 Taiwan TLEs campaign. * Works performed at National Cheng Kung University were supported in part by grants from NSPO (93-NSPO(B)-ISUAL-FA09-01) and NSC (NSC93-2112-M-006-007, NSC93-2111-M-006-001) in Taiwan.
AE31A-0161 0800h
Influence of Gravity Wave Density Perturbations on the Sprite Generation Mechanism
This paper is part of a new initiative towards understanding the coupling of mechanical/thermal, electrical energy deposition in the atmospheric system by gravity waves and by lightning, which are signaled by the occurrence of lightning induced optical phenomena, e.g. sprites. We have investigated the effects of neutral density perturbations specifically caused by gravity waves in determining electric breakdown locations in the mesosphere that initiate sprites. We adapted the model of {\it S\~{a}o Sabbas} [2003] to simulate the spatio-temporal evolution of the lightning induced quasi-electrostatic field in the mesosphere when gravity waves are present. Gravity wave density perturbations were computed for a number of representative convective plume geometries. Sprites are one of several optical observable components of electrical energy deposition in the middle/upper atmosphere by lightning. Multiple sprites are often observed to occur simultaneously, laterally displaced from the underlying causative cloud-to-ground (CG) discharge. {\it S\~{a}o Sabbas} [2003] showed that pockets of neutral density depletions have a lower electric breakdown threshold. This facilitates the development of the streamer channels constituting sprites at multiple locations offset from the causative discharge, in good agreement with observations. Gravity waves are the principal source of density inhomogeneities in the mesosphere. They transport energy and momentum from the troposphere into the middle and upper atmosphere being a major forcing in the general circulation of the middle atmosphere [{\it Fritts and Alexander}, 2003]. The vertical motion of strong convective thunderstorm plumes generates gravity waves with potentially steep phase slopes and density perturbations which are likely most important for sprite initiation.