AE23A-0830 1340h
Lightning Observation on the Nikaho-Kogen Wind Farm in Japan
Wind power is one of the most expected clean energy worldwide. But lightning damages of wind turbines have been reporting in many countries. The upward lightning flashes, which are initiated by the upward leader from the tall structures to the thunderstorms, are often observed in winter in Japan, because the winter thunderstorm height is very low compared to the summer thunderstorm. It is necessary to consider the probability of downward lightning flashes and upward lightning flashes for the windmills in winter. It is well known that positive lightning often occurs in winter. Positive lightning current sometimes have a long duration of several hundreds milliseconds and the total charge exceeds several hundreds coulombs. Positive lightning discharges with a large energy are threat for wind turbine blades. We started to observe the winter lightning on the Nikaho-Kogen wind farm in Akita prefecture in Japan for studying the lightning protection methods of wind turbine blades. The plant consists of fifteen windmills and the maximum height of a windmill is 93m with the power of 1650kw. The lightning observation is as follows: (1) Lightning current measurements using Rogowski coils and coaxial shunt resistors. We developed a new current measurement system for windmills. A Rogowski coil of 0.9 m in diameter was installed around the main shaft and a Rogowski coil of 4 m in diameter was installed at the tower bottom of windmill. (2) Lightning channel measurements with liquid crystal shutter cameras. The camera has a special shutter system to take the picture very quickly. If the strong luminosity like as a lightning stroke, a liquid crystal shutter is open within 5 microseconds. Main observational results are as follows: (1) A positive lightning flash composed of two strokes that a time interval between strokes is 97ms, first stroke peak current is +40.1kA and second stroke peak current is +33.5kA was observed. (2) Many pulse currents of upward leader discharges without return strokes were often measured. Since the upward discharges from the blades occur frequently under the low-altitude winter thunderstorms, there is the possibility of the partial destruction of the blades due to the cumulative stress of the blade materials by the upward discharges
AE23A-0831 1340h
Triggered Lightning in Brazil: First Results
Two induction techniques using rocket-and-wire systems were applied at the International Center for Natural and Triggered Lightning Studies (CIPRIN) in Brazil: (i) the classical method, which involves a conductive trailing wire connected to the ground and (ii) the altitude method, with an ungrounded wire. The artificial initiation of atmospheric discharges was conducted for the first time in tropical latitudes (S~$22^041.2^{\prime}$; W $44^059.0^{\prime}$; altitude: $625$~m). The $4$ campaigns analyzed in this work occurred during the summers of $2000$ to $2003$, resulting in $41$ rocket flights with success rates of $45%$ and $27%$ for the classical and the altitude techniques, respectively. Coaxial shunts, an optical sensor, electric field antennas and a high speed camera were used for characterizing the current waveforms of the classical strokes. Simultaneous results on current, luminosity and electric field, accompanied by high-speed camera images were obtained for the first time for an altitude flash. In addition, the peak currents of the strokes were determined using data from the shunts and compared to measurements from magnetic cards. The Brazilian Integrated Network (BIN), ground-based sensors that cover the southeastern area of Brazil, detected $54%$ of the triggered flashes. Compared to both artificial lightning observations performed in other countries and natural strokes detected by the BIN during the induced events, there was a tendency for peak currents to be generally higher for triggered lightning at CIPRIN . Moreover, the current pulses of the induced strokes observed in Brazil were faster than pulses measured in other sites. The higher peak currents and faster pulses observed at CIPRIN resulted in similar values of charge transferred per stroke compared to data obtained by other sites. However, more triggering experiments at CIPRIN are needed to provide better statistical results.
AE23A-0832 1340h
Modelisation of electromagnetic radiation of negative stepped leader
Experiments in laboratory and in stormy environments have shown that one of the main component of electromagnetic radiation of lightning is mainly due to the negative stepped leader in the VHF-UHF frequency band; the power radiated by the positive one is 1000 times lower than that by the negative one. This difference of radiation is due to the stepped propagation of the negative leader with which strong and fast pulses of current are associated. Just before the step, the negative discharge is composed of four zones: a resistive and conductive canal called negative leader, a high resistive zone called positive streamers zone, a resistive and conductive canal called space leader, and an active head associated with a negative corona development. The strong current pulse occurs during the connection of the negative and space leaders. The time evolution of the current along the negative discharge is performed using a transmission line model in which the distributed capacitances and inductances remain constant while the distributed resistances vary as a function of currents. A current pulse injected at the extremity of the transmission line simulates the negative corona development. The radiated field is computed from the current distribution by integrating along the discharge the electromagnetic field radiated by elementary Hertz dipoles. From this model, the power radiated by the negative leader versus the altitude is investigated.
AE23A-0833 1340h
Initial Data from Balloon Borne Slow-Antenna
We report on progress toward the goal of using Balloon-borne slow antenna measurements in conjunction with lightning mapping array data to discern details of charge transport in lightning strokes. Our field-change electronics directly descended from traditional slow antennas. Each of four 100 cm$^{2}$ electrodes connects to a charge amplifier designed to swing each output by 1 Volt for 22.5 kV/meter of external E-field change. The instrument comprises four electrodes(virtual ground) mounted to a 6" diameter sheet aluminum cylinder (which provides common for all on-board electronics). The electrodes are arranged on the cylinder such that summing and differencing of electrode charges yields signals proportional to the E-field along the symmetry axis of the cylinder and in two directions perpendicular to it. Were it possible to hold the cylinder fixed in the sky, this electrode configuration would provide vector field-change data directly. As the cylinder cannot easily be constrained from rotating under a balloon, a multi-axis magnetometer is added to allow calculation of the instantaneous orientation of sonde and electrodes relative to fixed Earth B-field. To put all measurements in proper geographic context and enable ready package recovery, GPS data are collected and transmitted to ground. All data acquisition is integrated at a 16-channel, 16-bit A/D converter which multiplexes through 8 channels every 100 $\mu$s. The A/D converter is part of a single-board computer running Linux. The A/D buffering scheme is sufficient to acquire 2 seconds worth of data continuously. Every two seconds, data acquisition pauses for some tenths of a second to write the data to a compact flash. Telemetry at 19200 kbaud is also provided to enable real-time monitoring of instrument health and maximum E-field change over each 18-second data interval. Three flights into electrical storms and two ground tests during storms provided initial data. Lightning flashes are clearly seen on the ground corresponding to measurements by other ground-based E-field instruments. In flight, the dominant signal in an electrified environment with distant lightning flashes consists of two elements. a) Continuously varying signals in the fractional hertz range caused by the transverse components of the cloud electrical field modulated by the rotation of the package. b) Impulsive signals with periods of 2-5 milliseconds corresponding to corona discharge. With nearer lightning flashes, the signal is similar to ground-based measurements to first order. Our current efforts are in data analysis to properly remove the rotation effects and impulsive discharges.
AE23A-0834 1340h
Observations of Energetic Radiation Associated with Natural Lightning
Over the past five years we have made observations of energetic radiation associated with the stepped-leaders of natural lightning. Coincident recordings of the electric field changes due to the discharges indicate that the bursts associated with stepped leaders begin a few milliseconds before and continued until the onset of the first return stroke. This radiation was associated with the approaching stepped-leaders. Our most recent observations have been made with the aid of the New Mexico Tech Lightning Mapping Array, allowing us to determine the location of the lightning channel with respect to our radiation detectors. Our data indicate that energetic radiation is not only associated only with stepped-leaders, but also with the return-strokes and dart-leaders of natural lightning as well. Through the use of the channel and ground strike locations, it appears that the bursts of radiation associated with natural dart-leaders and return-strokes occur during closest approach of their wave-fronts.
AE23A-0835 1340h
The Corona Discharge Waves in Thunderclouds and Formation of Ionic Channels
Measurements of electric field strength in the thunderclouds persistently give values by an order of magnitude lower than the breakdown threshold of the pure air. Models of lightning propagation through the thundercloud usually start with existing highly conductive channel in the cloud of rather large length and rather thin, at the end of which the field is already enhanced due to charge redistribution along the channel and thus the channel may increase its length due to streamers and leader formation. It is not clear how such long highly conductive channel may be formed. It is well known that the droplets, ice particles, hailstones or snowflakes may enhance the electric field and produce corona discharge. We assume that in a small part of the thundercloud an exceptionally high concentration of large hailstones or (and) water drops is formed and the onset-strength of the corona discharge becomes smaller than the background electric field. Polarization of this "hot spot" (with high conductivity and high ionization rates) produces charges at opposite sides of this volume. The increased electric field initiates corona discharge in other parts of the cloud with "normal" sizes of large ice particles and water drops. The small ice particles and water droplets are removed in the direction perpendicular to the axis of the channel. The high conductivity channel increases its length. The corona discharge front moves as a wave with the velocity of the order of ion drift velocity and the electric field at the ends of the channel increases, until the breakdown conditions are reached. A simple analytical model of such a wave is developed and the results are compared with some observations and data of the laboratory experiments, and the results of 3D numerical simulations of the relevant electrodynamic problem (Poisson equations are solved simultaneously with equations of motion of ions and charged particles, and electrons, and with a set of kinetic equations).
AE23A-0836 1340h
Investigations of Physical Details in Developing Lightning Flashes
The goal of this presentation is to investigate the physical details of a lightning flash, as determined by comparing data from three different lightning sensors. The sensors are a Lightning Mapping Array (LMA), a logarithmic radio frequency receiver (LogRF), and a flat plate antenna for sensing electric field changes, referred to as a `fast' antenna (FA) because of its short (100 microsecond) decay time constant. The LMA locates impulsive events in successive 100 microsecond time windows, while the FA and LogRF detect impulsive and more continuous events of the developing lightning flash. The data from these instruments are characterized by time, amplitude or power, and, for the LMA, position. FA and LogRF sensors have long been used in lightning studies while the newer LMA shows how the flash develops in space and time. In this presentation, we will compare the amplitudes of the FA and the LogRF events to each other and to the power of the associated LMA events. We will examine situations in which only one or two of the sensors are detecting radiation from a lightning flash. We will also investigate where the FA and LogRF events occur in the time-resolved LMA picture of a flash. By these comparisons we hope to gain new insights into the physics of a developing flash.
AE23A-0837 1340h
Dimensional Reduction: a Method for Retrieving Lightning Charge
A method is introduced for retrieving the locations and magnitudes of charges deposited by a lightning flash using multiple ground-based electric field change measurements. The method, called Dimensional Reduction, reduces the number of unknowns in a discrete 2-charge model from the standard of eight (x, y, z, Q, x', y', z', Q') to just four (x, y, z, Q). This reduction is accomplished by analyzing "residual measurements" that are formed by subtracting from each ground-based electric field change the contribution due to the source (x, y, z, Q). Using an improved analytic solution to the four-parameter point charge model (or "Q-model") the residual measurements are inverted to find the associated "residual source." For flash charge depositions that look approximately like two arbitrary charges, the residual source will be modeled accurately when (x, y, z, Q) is accurate. Hence, one need only minimize a chi-squared goodness-of-fit that is a function of the four variables (x, y, z, Q), rather than one that is a function of the eight variables (x, y, z, Q, x', y', z', Q'). The accuracy of the method is assessed by inverting computer-simulated electric field changes produced from known charge sources (see related presentation by Murray et al., this session). The method is also applied to analyze real lightning electric field change data derived from the NASA Kennedy Space Center (KSC) and United States Air Force (USAF) Eastern Range (ER) ground-based field mill network. It is shown that the charge retrievals compare favorably with associated ancillary ground- and satellite- based lightning measurements.
AE23A-0838 1340h
A Comparison of Charge Retrievals using Standard One- and Two-Charge Source Models and the Dimensional Reduction Method
Lightning poses a frequent threat to US space vehicle operations in Florida. In response to this threat, the Launch Pad Lightning Warning System (LPLWS) was developed in the early 1970s. Today, it consists in part of a network of 31 electric field mills situated over Kennedy Space Center (KSC) and Cape Canaveral Air Force Station (CCAFS) in east central Florida. When a lightning flash occurs in the vicinity of the LPLWS, several or all of the mills will record a change in electric field. Provided the flash is not too distant and/or too complex, the set of field changes can be analyzed to reasonably retrieve the locations and magnitudes of charges deposited by the flash. This work assesses retrieval accuracy for flashes that occur in the vicinity of the LPLWS and that are not overly complex. Two optimization methods have been used to analyze simulated electric field change data produced from computer-generated charge sources. The first method, a standard least-squares optimization procedure using the Levenberg-Marquardt algorithm, has been used to find the eight unknown parameters in a general two-charge model. The second, a Dimensional Reduction (DR) technique described by Koshak (this session), reduces the eight parameters to four by analyzing residual measurements. The simulated field change data wereas generated using known charge sources and a forward law (i.e., Coulomb's law of electrostatics with the assumptions that the Earth is an infinite plane conductor and the atmosphere a perfect vacuum). To determine the typical retrieval error of a particular source situated at a particular location relative to the LPLWS, the source was analyzed 100 times (each time the simulated field change measurement error was defined by selecting it from a random distribution). Hence, statistically significant retrieval errors were obtained for the given source location. The process was repeated for many other source locations, both over and outside the LPLWS, so that a color-coded "retrieval error map" could be obtained for each retrieved model parameter. The process was completed for three different source types (monopole, vertical 2-charge, slanted 2-charge), and the results are inter-compared and summarized.
AE23A-0839 1340h
The Possible Charge Structure of Thunderstorms and Lightning Discharge Features on the Tibetan Plateau
The Tibetan Plateau exerts a profound thermal and dynamical influence on the atmospheric circulation because of its high elevation and great dimension. The Plateau is usually covered by active cumulus convection in the summer and there are many thunderstorms, hailstorms, showers and convective clouds on the central Plateau. In this paper we present an analysis of the characteristics of thunderstorms and lightning discharges on the central Tibetan Plateau by using the ground-based observation data from field mill, slow antenna, fast antenna system, high-speed digital camera and VHF radiation pulse location technique in the summer of 2003 and 2004. It has been found that the thunderstorm on the plateau is usually accompanied with hail fall and the surface electric field underneath thunderstorm usually shows the same polarity as clear sky. The characteristics of both surface electric field and lightning discharges suggest that there was a larger-than-usual positive charge region in the bottom of thundercloud, and resulted in a tripole charge structure in thunderstorm. Precipitation particles, especially solid particles, are the main contributors of the lower positive charge region. This indicates that collision between ice crystals and rimed graupel particles is the dominant mechanism for charge separation in thunderstorms. The cloud-to-ground (CG) discharge takes about 21.6% of all lightning discharges with a positive CG ratio of about 18.8%. A long-duration intracloud discharge process between the lower positive charge region and the middle negative charge region in the cloud usually occurred just before the negative CG discharge, and it seems to be a necessary condition for the negative CG. Acknowledgments. The research was supported by the National Natural Science Foundation of China (Grant No. 40135010 and 40325013 ).
AE23A-0840 1340h
Analysis of Two Storms Observed During STEPS
The Severe Thunderstorm Electrification and Precipitation Study (STEPS), which took place in the vicinity of Goodland, KS, was conducted during the summer of 2000 in order to gain a better understanding of low precipitation (LP) supercells and positive cloud-to-ground (CG) lightning. We report on two storms observed during STEPS: one that occurred on 3 June and the other on 19 June. The 3 June storm developed just north of the Nebraska-Kansas border and moved into northwestern Kansas. It was an isolated LP supercell with one inch hail that appears to have exhibited an inverted electric dipole structure (negative over positive charge). There was frequent intracloud (IC) lightning, but no CG lightning was observed. The 19 June storm was multicellular, with surface winds exceeding 65 kts. This storm passed through the center of the STEPS domain and had predominately IC lightning, with less than one percent of all flashes reaching the ground. The polarity of the few CGs was mainly negative (93%). Data for these storms were gathered using the National Lightning Detection Network, the New Mexico Tech Lightning Mapping Array (LMA), the NCAR S-Pol radar located near Idalia, Colorado, the CSU-CHILL radar located near Burlington, Colorado, mobile GLASS soundings, and the SDSMT T28 research aircraft. The microphysical and electrical structure of these storms will be presented and compared using hydrometeor data from the T28, polarimetric variables from the S-Pol and CHILL radars, along with data from the LMA and T28 electric field mills.
AE23A-0841 1340h
Thermodynamic Conditions Favorable to Thunderclouds with Inverted Electrical Polarity
Cloud base height has recently been suggested as a key parameter in accounting for the order-of-magnitude contrast in land-ocean lightning activity (Williams and Stanfill, 2002), and the factor of 2-3 contrast in lightning activity over the Amazon and Congo River basins (Williams and Satori, 2004). The present study examines the influence of cloud base height in reversing the polarity of the main cloud dipole, for which there is increasing evidence (e.g., Krehbiel et al, 2000; Rust and MacGorman, 2002; Lang et al, 2004). Numerous laboratory studies indicate that the positive charging of large ice particles is associated with large cloud water contents (Takahashi, 1978; Saunders et al, 1991; Pereyra et al, 2000). The latter circumstance is encouraged by broad, undiluted updrafts, or by a suppression of warm rain coalescence. Both situations are favored by high cloud base heights. These qualitative ideas are supported by thermodynamic climatologies of cloud base height (dew point depression) and wet bulb potential temperature (a proxy for instability) over the CONUS, showing maximum cloud base height coincident with maximum w in the immediate lee of the Rocky Mountains, where both clustered positive ground flash activity and inverted polarity storms have been most prevalent. Clustered +CG activity is relatively scarce in thunderstorm-rich Florida and Alabama, where cloud base heights are half as high, but CAPE values are comparable to those in the Great Plains.
AE23A-0842 1340h
Are Large Volcanic Eruptions Just Dirty Thunderstorms?
A large number of volcanic eruptions are known to produce lightning, and eruption clouds in this category are comparable in size or larger than ordinary thunderclouds. Water substance is fundamental to explosive volcanism by exsolving from magma to the vapor phase, as the magma ascends beneath the volcano. A cubic meter of magma at depth contains about 100 liters of water (in condensed form). Simple calculations are performed that distribute all the available water over a spherical relaxation volume representing the explosion. The relaxation volume is given by E/Po where E is the available energy and Po is the ambient pressure against which the explosion expands. These results show that water concentrations in an eruption cloud can saturate the relaxation volume, guaranteeing the presence of condensate (liquid water and ice). In this environment, ice microphysics is applicable, volcanic ash and tephra may be ice-coated, and possibly the same viable mechanism known to be responsible for the electrification of thunderstorms is active. The gross electrical structures of large volcanic eruptions and the behavior of cloud-to-ground lightning are also consistent with the dirty thunderstorm hypothesis.
AE23A-0843 1340h
Thunderstorm Charge Studies Using a Simple Cylindrical Charge Model, Electric Field Measurements, and Lightning Mapping Observations
A simple cylindrical disk model has been developed to simulate the electrical structure and temporal evolution of thunderstorms. Lightning mapping observations are used to infer the heights, thicknesses, and diameters of the charge regions involved in lightning in an actual storm. The amounts of charge and the charging currents are determined by adjusting their values to replicate a) balloon-borne electric field mill (EFM) soundings through the storm and/or b) ground-based EFM observations beneath the storm. The occurrence of lightning is simulated using the breakeven field for runaway electron breakdown as a criterion for `triggering' discharges, and by then removing an adjustable fraction of charge from the appropriate charge regions. With the lightning mapping observations as a guide, the above approach allows one to quantitatively interpret space-time EFM soundings and ground EFM records in terms of charging rates, charge amounts, and instantaneous profiles of the electric field and electric potential in the storm. The `ground truth' information provided by the EFM data can also be used to infer and quantitatively estimate other charges not involved in lightning, in particular the screening charge around the top of the storm and corona charges from the ground. An interesting result of the simulations is that the upper screening charge is required to be dissipated or removed in some manner; otherwise lightning eventually occurs between it and the uppermost storm charge (e.g., between negative screening charge and upper positive charge in a normal polarity storm). In addition to raising the question as to how the screening charge is removed, the result suggests that upward jets could result from the lack of such removal or dissipation.
AE23A-0844 1340h
Leader Velocities of Normal and Inverted Intracloud Lightning Flashes as Observed by the Lightning Mapping Array
The New Mexico Tech lightning mapping array, LMA, has been operated during six lightning seasons. We have performed a detailed analysis of the velocities of lightning channels. The average velocity of the negative leaders have been deduced for many flashes. Normal intracloud flashes begin upward from between 6 and 10 km with velocities of about $1.5 \times 10^5$ m/s (0.5 to $3 \times 10^5$). The negative leader in inverted intracloud flashes are slower and begin downward from 9 to 12 km at about half the speed. In these flashes we isolated the branches of interest and used a cubic spline to fit each branch in X, Y, and Z versus time. As a result of the fits, we have characterized the evolution of the velocity of individual branches and the overall flash. Typical branch velocities range between $4 \times 10^4$ and $3 \times 10^5$ m/s.
http://lightning.nmt.edu
AE23A-0845 1340h
Potential Wells and Preliminary Breakdown in CG Flashes
In this study, we explore the link between preliminary breakdown in CG flashes and the presence of potential wells associated with the lower positive charge of the storms which create the CG flashes. Because E is proportional to the first derivative of the electric potential, V, the presence of a one dimensional potential well can be inferred from a change in the polarity of the vertical component of E between two altitudes. In this study, the two altitudes are typically at ground and near the altitude of lightning initiation. The measurement of E at the ground is taken by ground field mills in the vicinity of the CG flash, and the measurement of E in the cloud is typically taken by a balloon borne electric field meter. When balloon measurements of E are not available, the altitude and behavior of the first few LMA detected radiation sources of the CG are used as a proxy for the altitude of large positive fields. The time between the first LMA detected radiation source and the first return stroke as detected by the NLDN or by fast antenna data is used as a proxy for determining the presence of preliminary breakdown. {\it Beasley et al.} [1983] studied 80 CG flashes from nine storms at the NASA Kennedy Space Center and found that the "durations of stepped leaders lie most frequently in the range of 6-20 milliseconds." They also found that preliminary breakdown, or "preliminary variation," lasted from 11 to 500 ms with a mean duration of 118 ms and a median duration of 65 ms. This study shows a strong correlation between the existence of potential wells associated with the lower positive charge of storms and delays of greater than 20 ms, the longest lifetime reported by {\it Beasley et al.} for the stepped leader, between flash initiation and the first return stroke. This result is consistent with a potential well influencing the propagation of the negative leader and causing preliminary breakdown. References: Beasley, W.H., M. A. Uman, and P.L. Rustan, Electric fields preceding cloud-to-ground lightning flashes, {\it J. Geophys. Res., 87}, 4883-902, 1983
AE23A-0846 1340h
Runaway breakdown and thunderstorm and lightning electric fields
Recent measurements of x-ray emission made at the International Center for Lightning Research and Testing (ICLRT) at Camp Blanding, FL, have shown that runaway breakdown is likely a common property of lightning. However, the intensity, source location and energy spectrum of this emission does not appear to be consistent with the relativistic runaway electron avalanche (RREA) model, developed by Gurevich et al. (1992). An alternative model is the so-called cold runaway electron model. In this model, the high energy tail of the bulk free-electron population runs away and can be accelerated to relativistic energies. However, this model requires an electric field, presumably at the leader tip, much larger than the conventional breakdown field. On the other hand, recent observations of gamma-ray emission from thunderstorms appear to be consistent with the RREA model. In some cases, calculations show that the intensity of runaway electrons may be sufficiently large to limit the electric field that can be achieved inside thunderstorms. In this presentation, results of a new Monte Carlo simulation of runaway breakdown will be presented, a comparison with the x-ray observations of lightning will be made, and the implications for thunderstorm and lightning electric fields will be discussed.