TRIGGERED-LIGHTNING PROPERTIES INFERRED FROM MEASURED...
Transcript of TRIGGERED-LIGHTNING PROPERTIES INFERRED FROM MEASURED...
TRIGGERED-LIGHTNING PROPERTIES INFERRED FROM MEASURED
CURRENTS AND VERY CLOSE MAGNETIC FIELDS
By
ASHWIN B. JHAVAR
A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
2005
Copyright 2005
by
Ashwin B. Jhavar
iii
ACKNOWLEDGMENTS
I would like to thank Dr. Vladimir A. Rakov for his infinite patience, guidance, and
support throughout my graduate studies at the University of Florida. I would like to thank
Dr. Martin A. Uman and Dr. Douglas M. Jordan for their valuable suggestions during the
weekly lightning meetings. I sincerely thank Jens Schoene, Jason Jerauld, Rob Olsen,
Brian DeCarlo, and Vinod Jayakumar for helping me with the data and software, and for
other innumerable favors (without which I would not have been able to complete my
thesis). Research in my thesis was funded in part by National Science Foundation. The
data analyzed in the thesis were originally acquired with NSF and FAA funding.
iv
TABLE OF CONTENTS page
ACKNOWLEDGMENTS ................................................................................................. iii
LIST OF TABLES............................................................................................................. vi
LIST OF FIGURES .......................................................................................................... vii
ABSTRACT..................................................................................................................... xiii
CHAPTER
1 INTRODUCTION ........................................................................................................1
2 LITERATURE REVIEW .............................................................................................3
2.1 Cumulonimbus........................................................................................................3 2.2 Cloud Charge Distribution......................................................................................4 2.3 Mechanisms of Cloud Electrification .....................................................................6
2.3.1 Convection Mechanism. ...............................................................................7 2.3.2 Graupel-ice Mechanism................................................................................7
2.4 Downward Negative Lightning Discharges to Ground ........................................10 2.5 Artificial Initiation (Triggering) of Lightning Using the Rocket-and-Wire
Technique...............................................................................................................15 2.5.1 Classical Triggering....................................................................................16 2.5.2 Altitude Triggering.....................................................................................18
2.6 Previous Studies of Displacement Current Associated with Triggered Lightning................................................................................................................19
2.6.1 Theory.........................................................................................................19 2.6.2 Estimation of Displacement Current Contribution at 50 m........................21
3 CHARACTERIZATION OF EXPERIMENTAL DATA USED IN THIS STUDY .25
3.1 Magnetic Field Measuring Techniques ................................................................25 3.2 Experimental Setup...............................................................................................27
3.2.1 ICLRT Overview........................................................................................27 3.2.2 1997 Experiments .......................................................................................29 3.2.3 1999 Experiments .......................................................................................30
3.2.3.1 Instrumentation for Current Measurements .....................................32 3.2.3.2 Instrumentation for Electric Field Measurements ............................33
v
3.2.3.3 Instrumentation for Electric Field Derivative Measurement............33 3.2.4 2000 Experiments .......................................................................................33 3.2.5 2001 Experiments .......................................................................................35
3.3 Data Presentation ..................................................................................................36 3.3.1 General Information ...................................................................................36 3.3.2 Channel-base current ..................................................................................40 3.3.3 Magnetic Field............................................................................................42 3.3.4 Electric field derivative (dE/dt) ..................................................................47
4 ESTIMATION OF LEADER AND RETURN-STROKE CURRENTS FROM MEASURED MAGNETIC FIELDS..........................................................................51
4.1 Introduction...........................................................................................................51 4.2 Estimation of Currents Using Ampere’s Law ......................................................51 4.3 Discussion and Summary .....................................................................................77
5 DISPLACEMENT CURRENT ASSOCIATED WITH LEADER/RETurN STROKE SEQUENCES IN TRIGGERED LIGHTNING.........................................80
5.1 Displacement Current Estimates from Measured Magnetic Fields and Channel-Base Currents ..........................................................................................80
5.2 Displacement Current Estimates from dE/dt Signatures at 15 and 30 m .............92 5.3 Discussion and Summary ...................................................................................106
6 SUMMARY..............................................................................................................108
7 RECOMMENDATIONS FOR FUTURE RESEARCH ..........................................112
APPENDIX DISPLACEMENT CURRENT GRAPHS.................................................113
LIST OF REFERENCES.................................................................................................144
BIOGRAPHICAL SKETCH ...........................................................................................147
vi
LIST OF TABLES
Table page 2-1 Displacement currents Id estimated from measured electric field data at times
prior to and at the onset of a 20 kA peak stroke current. .........................................22
3-1 Summary of mean, standard deviation, GM (geometric mean), and sample sizes of measured peak current, peak magnetic field and electric field derivative ...........37
5-1 Displacement currents estimated using eq. 5.4 (step-wise approximation of dE/dt distance dependence) and eq. 5.5 (linear approximations of dE/dt distance dependence)..............................................................................................................95
5-1 (Contd.) Displacement currents estimated using eq. 5.4 (step-wise approximation of dE/dt distance dependence) and eq. 5.5 (linear approximations of dE/dt distance dependence)..................................................................................96
vii
LIST OF FIGURES
Figure page 2-1 An isolated thundercloud in central New Mexico, with a rudimentary indication
of how electric charge is thought to be distributed inside and around the thundercloud, as inferred from the remote and in situ observations.. ........................5
2-2 Balloon measurements of corona current and the inferred vertical electric field E versus altitude and air temperature inside a small storm in New Mexico on 16 August 1981, which produced no lightning. ..............................................................6
2-3 Illustration of the convection mechanism of cloud electrification. ............................8
2-4 Charge transfer by collision in the graupel-ice mechanism of cloud electrification. It is assumed that the reversal temperature TR is -15 oC and that it occurs at a height of 6 km. ......................................................................................9
2-5 A vertical tripole representing the idealized gross charge structure of a thundercloud such as that shown in Figure 2-1; the negative screening layer charges at the cloud top and the positive corona space charge produced at ground are ignored here............................................................................................10
2-6 Various processes comprising a negative cloud-to-ground lightning flash. ............11
2-7 Sequence of events in classical triggered lightning. The upward positive leader and initial continuous current constitute the initial stage. ........................................16
2-8 Sequence of events in altitude-triggered lightning leading to the establishment of a relatively low-resistance connection between the upward-moving positive leader tip and the ground. The processes that follow the sequence shown, an initial continuous-current and possibly one or more downward-leader--upward-return-stroke sequences, are similar to their counterparts in classical triggered lightning. The rocket speed is of the order of 102 m s-1. ..........................................18
2-9 Qualitative illustration of the shape of the time derivative of the ground level vertical electric field of a nearby return stroke . Since displacement current density is simply related to the derivative of electric field by a constant, it has the same waveshape.. ...............................................................................................24
viii
2-10 From top to bottom, magnetic field measured at 50 m during stroke 5 of Flash 96-23, the corresponding field according to Ampere’s law for magnetostatics, as applied to the measured channel-base current, and the difference between the two.. ..........................................................................................................................24
3-1 Thevenin equivalent circuit of a loop antenna .........................................................25
3-2 Norton equivalent circuit for a loop antenna............................................................26
3-3 An overview of the ICLRT at Camp Blanding, Florida, 1999-2001. Not all test objects are shown. ....................................................................................................28
3-4 Photograph of lightning flash S0012 triggered from the underground launcher. ....28
3-5 Locations of different instrumentation stations for 1997 multiple station experiment. ...............................................................................................................30
3-6 Data acquisition system used in the 1997 multiple-station experiment. ..................31
3-7 Experimental setup (placement of electric and magnetic field antennas) used in SATTLIF for 2000.. .................................................................................................34
3-8 Setup of strike rod and ring mounted over launch tubes in 2000.............................35
3-9 Measured channel-base current, Flash S9901, Stroke 3...........................................37
3-10 Magnetic field at 15 m, Flash S9901, Stroke 3. .......................................................38
3-11 Magnetic field at 30 m, Flash S9901, Stroke 3. .......................................................38
3-12 Electric field derivative (dE/dt) at 15 m, Flash S9901, Stroke 3. ............................39
3-13 Electric field derivative (dE/dt) at 30 m, Flash S9901, Stroke 3. ............................39
3-14 Return-stroke peak currents in (a) 1997, (b) 1999, (c) 2000, and (d) 2001. ............40
3-15 Return-stroke peak currents in 1997, 1999, 2000 and 2001.....................................41
3-16 Peak magnetic fields measured at (a) 5 m, (b) 10 m, (c) 20 m, and (d) 30 m in 1997. .........................................................................................................................42
3-17 Peak magnetic fields measured at (a) 15 m and (b) 30 m in 1999. ..........................43
3-18 Peak magnetic fields measured at (a) 15 m and (b) 30 m in 2000. ..........................44
3-19 Peak magnetic fields measured at (a) 15 m and (b) 30 m in 2001. ..........................45
3-20 Peak magnetic fields measured at (a) 15 m and (b) 30 m in 1997, 1999, 2000 and 2001. .........................................................................................................................46
ix
3-21 Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 1999. ................................47
3-22 Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 2000. ................................48
3-23 Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 2001. ................................49
3-24 Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 1999, 2000 and 2001. ......50
4-1 A straight current channel of infinite length and B at a distance r...........................51
4-2 A streaked-image diagram of a dart leader—return-stroke sequence in a rocket-triggered lightning flash ...........................................................................................52
4-3 Magnetic field at 15 m, Flash S9901, Stroke 3. .......................................................54
4-4 Magnetic field at 30 m, Flash S9901, Stroke 3. .......................................................54
4-5 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 5 m, (b) 10 m, (c) 20 m, and (d) 30 m in 1997......55
4-6 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 15 m and (b) 30 m in 1999. ...................................56
4-7 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 15 m and (b) 30 m in 2000. ...................................57
4-8 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 15 m and (b) 30 m in 2001. ...................................58
4-9 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 15 m in 1999, 2000, and 2001. ...................................59
4-10 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 30 m in 1997, 1999, 2000, and 2001. .........................60
4-11 Dart leader current inferred using Ampere’s Law for magnetostatics from measured magnetic fields measured at 30 m vs. that at 15 m in 1999, 2000, and 2001. .........................................................................................................................61
4-12 Dart leader current inferred using Ampere’s Law for magnetostatics from magnetic fields measured at 15 m vs. leader current inferred from dE/dt measurements in 1999, 2000, and 2001. ..................................................................62
4-13 Dart leader current inferred using Ampere’s Law for magnetostatics from magnetic fields measured at 30 m vs. leader current inferred from dE/dt measurements in 1999, 2000, and 2001. ..................................................................63
x
4-14 Return-stroke peak currents inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 5 m, (b) 10 m, (c) 20 m, and (d) 30 m in 1997. .........................................................................................................................64
4-15 Return-stroke peak currents inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 15 m and (b) 30 m in 1999. ..........................65
4-16 Return-stroke peak currents inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 15 m and (b) 30 m in 2000. ..........................66
4-17 Return-stroke peak currents inferred using Ampere’s Law for magnetostatics from measured magnetic fields at (a) 15 m and (b) 30 m in 2001. ..........................67
4-18 Return-stroke peak currents inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 15 m in 1999, 2000, and 2001. ..........................68
4-19 Return-stroke peak currents inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 30 m in 1997, 1999, 2000, and 2001. ................69
4-20 Return-stroke peak current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 30 m vs. that inferred from measured magnetic fields at 15 m in 1999, 2000, and 2001. ...................................................70
4-21 Return-stroke peak current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 15 m vs. measured return-stroke peak current in 1999, 2000, and 2001. ..........................................................................................71
4-22 Return-stroke peak current inferred using Ampere’s Law for magnetostatics from measured magnetic fields at 30 m vs. measured return-stroke peak current in 1997, 1999, 2000, and 2001. ................................................................................72
4-23 Leader vs. return-stroke currents inferred using Ampere’s Law for magnetostatics from magnetic fields measured at 15 m in 1999, 2000, and 2001...73
4-24 Leader vs. return-stroke currents inferred using Ampere’s Law for magnetostatics from magnetic fields measured at 30 m in 1999, 2000, and 2001...74
4-25 Comparison of IRS from (BL+BRS)30m and IRS from (BL+BRS)15m for 1999, 2000, and 2001. ..................................................................................................................75
4-26 Comparison of IRS from (BL+BRS)15m inferred using Ampere’s Law for magnetostatics vs. IRS measured at channel base in 1999, 2000, and 2001. ............76
4-27 Comparison of IRS from (BL+BRS)30m inferred using Ampere’s Law for magnetostatics vs. IRS measured at channel base in 1997, 1999, 2000, and 2001. ..77
5-1 Superposition of measured magnetic field and channel-base current for Flash S9903, stroke 3. ........................................................................................................81
xi
5-2 Displacement current for S9903 RS3 as inferred, using Eq. 5.3, from Imeas and Hmeas at 30 m.............................................................................................................83
5-3 Return-stroke peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 1999. ...................................................................84
5-4 Return-stroke peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 2000. ...................................................................85
5-5 Return-stoke peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 2001. ...................................................................86
5-6 Return-stroke peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 1999, 2000, and 2001..........................................87
5-7 Leader peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 1999..............................................................................88
5-8 Leader peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 2000..............................................................................89
5-9 Leader peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 2001..............................................................................90
5-10 Leader peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 1999, 2000, and 2001. ..................................................91
5-11 Distance dependences of dE/dt used in evaluating displacement current within 30 m of the lightning channel based on measured dE/dt at 15 and 30 m.................92
5-12 Return-stroke displacement current within 30 m of the lightning channel at the time of dE/dt peak at 15 m estimated using step-wise approximation (with propagation delay taken into account) in (a) 1999 and (b) 2000. ............................97
5-13 Return-stroke displacement current within 30 m of the lightning channel at the time of dE/dt peak at 15 m estimated using step-wise approximation (with propagation delay taken into account) in (a) 2001 and (b) 1999, 2000, and 2001...98
5-14 Return-stroke displacement current within 30 m of the lightning channel estimated using peak values of dE/dt at both 15 and 30 m estimated using step-wise approximation (without propagation delay taken into account) in (a) 1999 and (b) 2000. ............................................................................................................99
5-15 Return-stroke displacement current within 30 m of the lightning channel estimated using peak values of dE/dt at both 15 and 30 m estimated using step-wise approximation (without taking propagation delay into account) in (a) 2001 and (b) 1999, 2000, and 2001.................................................................................100
xii
5-16 Return-stroke displacement current within 30 m of the lightning channel estimated using peak values of dE/dt at both 15 and 30 m estimated using linear approximation (without propagation delay taken into account) in (a) 1999 and (b) 2000. .................................................................................................................101
5-17 Return-stroke displacement current within 30 m of the lightning channel estimated using peak values of dE/dt at both 15 and 30 m estimated using linear approximation (without taking propagation delay into account) in (a) 2001 and (b) 1999, 2000, and 2001.................................................................................102
5-18 Scatter plot showing displacement current estimates from dE/dt measured at 15 and 30 m from the lightning channel (step-wise approximation) vs. those estimated from measured channel base current and associated magnetic field. In the former case, the estimate corresponds to the time of peak value of dE/dt at 15 m........................................................................................................................103
5-19 Scatter plot showing displacement current estimates from dE/dt measured at 15 and 30 m from the lightning channel (step-wise approximation) vs. those estimated from measured channel base current and associated magnetic field. In the former case, the estimate is obtained using peak values of dE/dt at both 15 and 30 m. ................................................................................................................104
5-20 Scatter plot showing displacement current from dE/dt measured at 15 and 30 m from the lightning channel (linear approximation) vs. those estimated from measured channel base current and associated magnetic field. In the former case, the estimate is obtained using peak values of dE/dt at both 15 and 30 m. .............105
xiii
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science
TRIGGERED-LIGHTNING PROPERTIES INFERRED FROM MEASURED CURRENTS AND VERY CLOSE MAGNETIC FIELDS
By
Ashwin B. Jhavar
December 2005
Chair: Vladimir A. Rakov Cochair: Martin A. Uman Major Department: Electrical and Computer Engineering
Very close magnetic fields produced by rocket-triggered lightning measured in
1997, 1999, 2000, and 2001 at Camp Blanding, Florida, are examined. The leader and
return stroke contributions to the total magnetic field are estimated and used to infer
leader and return-stroke currents. The statistical characteristics of these inferred currents
are examined. The return-stroke currents inferred from measured magnetic fields are
compared with directly measured ones. Leader currents inferred from measured magnetic
fields are compared with those estimated using dE/dt measurements. The statistics of the
ratio of leader to return stroke currents are compiled. Current estimates from measured
magnetic fields are in reasonable agreement with independent measurements and
theoretical predictions found in the literature. From the statistical analysis of 97 records
obtained in 1999, 2000, and 2001 the mean leader current inferred from magnetic field at
15 m is 1.87 kA and standard deviation is 1.01 kA. For magnetic fields at 30 m (103
records obtained in 1997, 1999, 2000, and 2001), the mean leader current is 2.58 kA and
xiv
the standard deviation is 1.57 kA. Typically leader currents are in the range of few
kiloamperes.
Displacement current associated with leader/return stroke sequences in triggered
lightning is estimated using (1) measured channel-base current and current inferred from
measured magnetic field and (2) dE/dt measurements at 15 and 30 m. The displacement
currents found using the above two approaches are compared. Return-stroke
displacement current within 30 m of the lightning channel at the time of peak dE/dt at 15
m, estimated using step-wise approximation of dE/dt variation with distance from the
lightning channel (after taking propagation delay into account) is characterized by a mean
value of 3.1 kA. The minimum value is 0.5 kA and maximum is 8.0 kA.
1
CHAPTER 1 INTRODUCTION
Lightning discharges are the cause of many deaths and injuries. Electromagnetic
fields generated by lightning can have deleterious effects on sensitive electronic devices.
A detailed knowledge of electromagnetic fields generated by close lightning is needed for
developing adequate lightning protection schemes.
A review of the lightning literature is presented in Chapter 2. Cloud electrical
structure and mechanisms of cloud electrification are discussed. Salient properties of both
natural and triggered lightning are given with focus on triggered lightning. Both classical
and altitude rocket-triggered lightning discharges are considered.
Chapter 3 presents the characteristics of measured current, magnetic field, and
electric field derivative (dE/dt) waveforms due to rocket-triggered lightning. It also
contains a description of the instrumentation used to measure these quantities in 1997,
1999, 2000 and 2001.
Leader and return-stroke currents inferred from measured magnetic fields are
presented in Chapter 4. These are compared with the directly measured return-stroke
currents and with the leader currents inferred from close electric field derivative (dE/dt)
measurements.
Chapter 5 deals with the displacement currents associated with leader/return stroke
sequences in triggered lightning. Displacement currents are estimated from measured
magnetic fields and current records using Maxwell’s equations. Also included is an
estimation of peak displacement currents from dE/dt signatures measured at 15 and 30 m.
2
Chapter 6 summarizes the data and analysis of the leader and return-stroke currents
and the displacement currents associated with triggered lightning.
Recommendations for future research are given in Chapter 7.
The Appendix has the waveforms of the displacement currents.
3
CHAPTER 2 LITERATURE REVIEW
2.1 Cumulonimbus
The primary source of lightning is the cloud type termed cumulonimbus or
thundercloud. A thundercloud develops from a small fair-weather cloud called a cumulus,
which is formed when parcels of warm, moist air rise and cool by adiabatic expansion,
that is, without the transfer of heat or mass across the boundaries of the air parcels. When
the relative humidity in a rising and cooling parcel exceeds saturation, moisture
condenses on airborne particulate matter within it to form the many small water particles
that constitute the visible cloud. The height of the condensation level, which determines
the height of the visible cloud base, increases with decreasing relative humidity at the
ground. This is why cloud bases in Florida are generally lower than in arid locations,
such as New Mexico or Arizona. Parcels of warm, moist air can only continue to rise to
form a cumulus and eventually a cumulonimbus if the atmospheric temperature lapse
rate, that is, the decrease in temperature with increasing height, is larger than the moist-
adiabatic lapse rate, about of 0.6 oC per 100 m.
The convection of buoyant moist air is usually confined to the troposphere, the
layer of the atmosphere that extends from the Earth’s surface to the tropopause. The
height of tropopause varies from approximately 18 km in the tropics in the summer to 8
km in high latitudes in the winter. The tropopause is a narrow layer that separates the
troposphere from the next layer of the atmosphere, the stratosphere, which extends from
tropopause to a height of 50 km. In the troposphere the temperature decreases with
4
increasing altitude, while in the stratosphere the temperature at first becomes roughly
independent of altitude and then increases with altitude. A zero or positive temperature
gradient in the stratosphere serves to suppress convection and, therefore, hampers the
penetration of cloud tops into the stratosphere.
Lightning is usually associated with convective cloud systems ranging from 3 to 20
km in vertical extent. The horizontal dimensions of active air-mass thunderstorms range
from about 3 km to >50 km.
2.2 Cloud Charge Distribution
The distribution and motion of thunderstorm electric charges, most resides on
hydrometeors but some of which is free ions, is complex and changes continuously as the
cloud evolves. Hydrometeors whose motion is predominantly influenced by gravity (with
fall speeds > 0.3 m s-1) are called precipitation particles. All other hydrometeors are
called cloud particles. The basic features of the cloud charge structure include a net
positive charge near the top, a net negative charge below it, and an additional positive
charge at the bottom of the cloud. These features are illustrated in Figure 2-1.
In-situ measurements of electric fields inside the cloud have been made using free
balloons carrying instruments to measure those fields. In situ measurements are superior
to remote measurements in that a relatively accurate charge height can be determined.
However, since the balloon can sense the field only along a more or less straight vertical
path and it samples different portions of that path at different times, the charge magnitude
can be estimated only if assumptions regarding the size and shape of individual charge
regions and the charge variation with time are made. The average volume charge density,
ρv, in the cloud is generally found by assuming that the charge (i) is horizontally uniform
and (ii) does not vary in time.
5
Figure 2-1. An isolated thundercloud in central New Mexico, with a rudimentary
indication of how electric charge is thought to be distributed inside and around the thundercloud, as inferred from the remote and in situ observations. Adapted from Krehbiel (1986).
Then according to Gauss’s law in point form ρv = εo(dEz/dz), that is, ρv is
proportional to the rate at which the vertical electric field Ez increases or decreases with
increasing altitude z as the balloon ascends. Figure 2-2 shows the results of a vertical
sounding of the electric field in a small New Mexico storm that produced no lightning.
This electric field profile was obtained up to a height of 10 km above mean sea level
using a balloon-borne instrument that measured the corona current from a 1-m-long
vertical wire. The corona current and the corresponding vertical electric field reversed
sign twice, between 6 and 7 km and above 9 km. The charge structure in Figure 2-2, a
negative charge between -5 and -15 oC with positive charges above and below it, appears
to be consistent with the “classical” tripolar charge structure.
6
Figure 2-2. Balloon measurements of corona current and the inferred vertical electric
field E versus altitude and air temperature inside a small storm in New Mexico on 16 August 1981, which produced no lightning. Adapted from Byrne et al. (1983)
2.3 Mechanisms of Cloud Electrification
Any cloud electrification mechanism involves (i) a small-scale process that
electrifies individual hydrometeros and (ii) a process that spatially separates these
charged hydrometeors by their polarity, the resultant distances between the charged cloud
regions being of the order of kilometers. Since most charges reside on hydrometeors of
relatively low mobility, the cloud is a relatively good electrical insulator and leakage
7
currents between the charged regions are thought to have a small effect on the charge
separation process.
2.3.1 Convection Mechanism.
In this mechanism the electric charges are supplied by external sources: fair-
weather space charge and corona near the ground and cosmic rays near the cloud top.
Organized convection provides large-scale separation. According to this mechanism,
illustrated in Figure 2-3, warm air currents (updrafts) carry positive fair-weather space
charge to the top of the growing cumulus. Negative charge, produced by cosmic rays
above the cloud, is attracted to the cloud’s boundary by the positive charges within it.
The negative charge attaches, within a second or so, to cloud particles to form a negative
screening layer. These charged cloud particles carry much more charge per unit volume
of cloudy air than is carried by precipitation particles. Downdrafts, caused by cooling and
convective circulation, assumed to carry the negative charge down the sides of the cloud
toward the cloud base, this negative charge serving to produce positive corona at the
Earth’s surface. Corona generates additional positive charge under the cloud and, hence,
provides a positive feedback to the process. The convective mechanism results in a
positive cloud-charge dipole, although it seems unlikely that the negative charge region
formed by this mechanism would lie in a similar temperature range for different types of
thunderstorms, as suggested by observations. Note that in convection model there is no
role for precipitation in forming the dipole charge structure.
2.3.2 Graupel-ice Mechanism.
In this mechanism the electric charges are produced by collisions between
precipitation particles (graupel) and cloud particles (small ice crystals). Precipitation
particles are generally larger than cloud particles, although there is no absolute
8
Figure 2-3. Illustration of the convection mechanism of cloud electrification. Adapted
from MacGorman and Rust (1998)
demarcation in size to distinguish precipitation particles, which are falling out of the
cloud, from cloud particles, which remain essentially suspended or move upward in
updrafts. The large-scale separation of charged particles is provided by the action of
gravity. In the graupel-ice mechanism, which appears to be capable of explaining the
“classical” tripolar cloud charge structure, the electrification of individual particles
involves collisions between graupel particles and ice crystals in the presence of water
droplets. The presence of water droplets is necessary for significant charge transfer, as
shown by the laboratory experiments. A simplified illustration of this mechanism is given
in Figure 2-4.
The heavy graupel particles (two of which are shown in Figure 2-4) fall through a
suspension of smaller ice crystals (hexagons) and supercooled water droplets (dots). The
droplets remain in a supercooled liquid state until they contact an ice surface, whereupon
they freeze and stick to the surface in a process called riming. Laboratory experiments
show that when the temperature is below a critical value called the reversal temperature,
TR, the falling graupel particles acquire a negative charge in collision with the ice
particles. At temperature above TR they acquire positive charge. The charge sign reversal
9
temperature TR is generally thought to be between -10 and -20 oC, the temperature range
characteristic of the main negative charge region found in thunderclouds.
Figure 2-4. Charge transfer by collision in the graupel-ice mechanism of cloud
electrification. It is assumed that the reversal temperature TR is -15 oC and that it occurs at a height of 6 km. Taken from Rakov and Uman (2003).
It is believed that the polarity of the charge that is separated in ice-graupel
collisions is determined by the rates at which the ice and graupel surfaces are growing.
The surface that is growing faster acquires a positive charge.
It is possible that the primary electrification mechanism changes once a storm
becomes strongly electrified. For example, collisions between ice crystals and graupel
10
could initiate the electrification, and then the larger convective energies of the storm
could continue it.
2.4 Downward Negative Lightning Discharges to Ground
The source of lightning is usually a cumulonimbus, whose idealized charge
structure is shown in the Figure 2-5 as three vertically stacked regions labeled P and LP
for the main positive and the lower positive charge regions and N for the main negative
charge region.
Figure 2-5. A vertical tripole representing the idealized gross charge structure of a
thundercloud such as that shown in Figure 2-1; the negative screening layer charges at the cloud top and the positive corona space charge produced at ground are ignored here.
Downward negative lightning discharges, that is, discharges that are initiated in the
cloud, initially develop in an overall downward direction, and transport negative charge
to ground, probably account for about 90 percent of all cloud-to-ground discharges. The
overall cloud-to-ground lightning discharge, termed a flash, is composed of a number of
processes, some of which involve channels that emerge from the cloud while others
involve channels that are confined to the cloud volume. Only processes occurring in the
channels outside the cloud render themselves to optical observations that can be used to
11
determine channel geometry, extension speed and other pertinent features of those
channels. The sequence of the processes involved in a typical negative downward
lightning flash is shown in Figure 2-6.
Figure 2-6. Various processes comprising a negative cloud-to-ground lightning flash.
Adapted from Uman (1987, 2001)
The stepped leader is preceded by an in-cloud process called the preliminary or
initial breakdown. There is no consensus on the mechanism of this process. It may be a
discharge bridging the main negative and the lower positive charge regions, as shown in
Figure 2-5. The initial breakdown may last from a few milliseconds to some tens of
milliseconds and serves to provide conditions for the formation of the stepped leader. The
latter is a negatively charged plasma channel extending toward the ground at an average
speed of 2 x 105 m s-1 in a series of discrete steps. From high-speed time resolved
CLOUD CHARGE DISTRIBUTION PRELIMINARY
BREAKDOWNSTEPPED LEADER
FIRST RETURN STROKE
K AND J PROCESSES DART
LEADER
SECOND RETURN STROKE
ATTACHMENT PROCESS
t = 0 1.10 ms1.00 ms 1.15 ms 1.20 ms
19 ms 20 ms 20.10 ms 20.15 ms 20.20 ms
40 ms 60 ms 61 ms 62 ms 62.05 ms
12
photographs, each step is typically 1 µs in duration and tens of meters in length, the time
interval between steps being 20 to 50 µs. The peak value of the current pulse associated
with an individual step has been inferred to be 1 kA or greater. The stepped leader serves
to form a conducting path or channel between the cloud charge source and ground.
Several coulombs of negative charge are distributed along this path, including downward
branches. Thus the leader may be viewed as a process removing negative charge from the
source and depositing this charge onto the downward extending channel. The stepped-
leader duration is typically some tens of milliseconds, and the average leader current is
some hundreds of amperes.
The electric potential difference between a downward-moving stepped-leader tip
and ground is probably some tens of megavolts which is comparable to or a considerable
fraction of that between the cloud charge source and ground. The magnitude of the
potential difference between two points, one at the cloud charge source and the other on
ground, is the line integral of the electric field intensity between those points.
The upper and lower limits for the potential difference between the lower boundary
of the main negative charge region and ground can be estimated by multiplying,
respectively, the typical observed electric field in the cloud, 105 V m-1, by the height of
the lower boundary of the negative charge center above ground, 5 km or so. The resultant
range is 50 to 500 MV. As the leader approaches ground, the electric field at the ground
surface, particularly at objects or relief features protruding above the surrounding terrain,
increases until it exceeds the critical value for the initiation of one or more upward-
connecting leaders. The initiation of an upward connecting leader from ground in
response to the descending stepped leader marks the beginning of the attachment process.
13
This process ends when contact is made between the downward and upward moving
leaders, probably some tens of meters above ground (more above a tall structure),
whereafter the first return stroke begins. The return stroke serves to neutralize the leader
charge, in other words, to transport the negative charges stored on the leader channel to
the ground. It is worth noting that the return-stroke process may not neutralize all the
leader charge or may deposit some excess positive charge onto the leader channel and
into the cloud charge source region. The final stage of the attachment process and the
initial stage of the return-stroke process are complex. The net result of those stages is a
fully formed return stroke, which is somewhat similar to the potential discontinuity that
would travel upward along a vertical, negatively charged transmission line if the lower
end of the line were connected to the ground. The first return-stroke current measured at
ground rises to an initial peak of about 30 kA in some microseconds and decays to half-
peak value in some tens of microseconds while exhibiting a number of subsidiary peaks,
probably associated with the branches. The return stroke effectively lowers to ground the
several coulombs of charge originally deposited on the stepped-leader channel, including
that on all the branches.
The high-current return-stroke wave rapidly heats the channel to a peak
temperature near or above 30 000 K and creates a channel pressure of 10 atm or more,
resulting in channel expansion, intense optical radiation, and an outward propagating
shock wave that eventually becomes the thunder (sound wave) we hear at a distance.
When the first return stroke, including any associated in-cloud discharge activity, ceases,
the flash may end. In this case, the lightning is called a single-stroke flash. However,
more often the residual first-stroke channel is traversed downwards by a leader that
14
appears to move continuously, a dart leader. During the time interval between the end of
the first return stroke and the initiation of a dart leader, J (for junction) and K processes
occur in the cloud. K-process can be viewed as transients occurring during the slower J-
process. The J-processes amount to a redistribution of cloud charge on a time scale of
tens of milliseconds, in response to the preceding return stroke. The J-process is often
viewed as a relatively slow positive leader extending from the flash origin into the
negative charge region, the K-process then being a relatively fast “recoil streamer” that
begins at the tip of the positive leader and propagates toward the flash origin. Both the J-
process and the K-process in cloud-to-ground flashes serve to transport additional
negative charge into and along the existing channel, although not all the way to the
ground. In this respect, K-process may be viewed as attempted dart leaders. The
processes that occur after the only stroke in single stroke flashes and after the last stroke
in multiple-stroke flashes are sometimes termed F (final) processes. These are similar, if
not identical, to J-processes.
The dart leader progresses downward at a typical speed of 107 m s-1, typically
ignores the first stroke branches and deposits along the channel a total charge of the order
of 1 C. The dart-leader current peak is about 1 kA. Some leaders exhibit stepping near
ground while propagating along the path traversed by the preceding return stroke, these
leaders being termed dart-stepped leaders. When a dart leader or dart-stepped leader
approaches the ground, an attachment process similar to that described for the first stroke
takes place, although it probably occurs over a shorter distance and consequently takes
less time, the upward connecting-leader length being of the order of some meters. Once
the bottom of the dart or the dart-stepped leader channel is connected to the ground, the
15
second return-stroke wave is launched upward and again serves to neutralize the leader
charge. The subsequent return-stroke current at ground typically rises to a peak value to
10 to 15 kA in less than a microsecond and decays to half-peak value in a few tens of
microseconds. The upward propagation speed of such a subsequent return stroke is
similar to that of the first return stroke, although due to the absence of branches the speed
variation along the channel does not exhibit abrupt drops.
The impulsive component of the current in a subsequent return stroke is often
followed by a continuing current that has a magnitude of tens to hundreds of amperes and
duration up to hundreds of milliseconds. Continuing currents with a duration in excess of
40 ms are traditionally termed long continuing currents. The source for the continuing
current is the cloud charge, as opposed to the charge distributed along the leader channel,
the latter charge contributing to at least the initial few hundred microseconds of the
return-stroke current observed at ground. The time interval between successive return
strokes in a flash is usually several tens of milliseconds, although it can be as large as
many hundreds of milliseconds if a long continuing current is involved and as small as
one millisecond or less. The total duration of a flash is typically some hundreds of
milliseconds, and the total charge lowered to ground is some tens of coulombs.
2.5 Artificial Initiation (Triggering) of Lightning Using the Rocket-and-Wire Technique
Two techniques for triggering lightning with a small rocket that extends a thin wire
in the gap between a thundercloud and the ground are “classical” triggering and
“altitude” triggering. These descriptions primarily apply to triggering negative lightning.
These two techniques are discussed in the next two sections. Figure 2.7 and Figure 2.8
show the sequence of events for these two techniques.
16
2.5.1 Classical Triggering
The most effective technique for triggering lightning involves the launching of a
small rocket trailing a thin grounded wire toward a charged cloud overhead. This
triggering method is usually called classical triggering and is illustrated in Figure 2-7.
Figure 2-7. Sequence of events in classical triggered lightning. The upward positive
leader and initial continuous current constitute the initial stage. Adapted from Rakov et al. (1998)
The triggering success rate is generally relatively low during very active periods of
thunderstorms, one reason being that during such periods the electric field is more likely
to be reduced by a natural lightning discharge before the rocket rises to a height sufficient
for triggering.
When the rocket, ascending at about 200 m s-1, is about 200 to 300 m high, the
enhanced field near the rocket tip results in a positively charged leader that propagates
upward toward the cloud. This upward positive leader (UPL) vaporizes the trailing wire,
bridges the gap between the cloud and ground, and establishes an initial continuous
current (ICC) with a duration of some hundreds of milliseconds that effectively transports
17
negative charge from the cloud charge source to the triggering facility. The ICC can be
viewed as a continuation of the UPL when the latter has reached the main negative
charge region in the cloud. At that time the upper extremity of the UPL is likely to
become heavily branched. The UPL and ICC constitute the initial stage (IS) of a classical
triggered-lightning discharge. After cessation of the initial continuous current, one or
more downward dart-leader--upward-return-stroke sequence may traverse the same path
to the triggering facility. The dart leaders and the following return strokes in triggered
lightning are similar to dart-leader – return-stroke sequences in natural lightning,
although the initial processes in natural downward and in classical triggered lightning are
distinctly different. In summer, the triggering success rate for positive lightning is
apparently lower than for the negative lightning.
There is contradictory information regarding whether the height H of the rocket at
the time of lightning triggering depends on the electric field intensity E at ground at the
time of launching the rocket. A strong correlation (with correlation coefficient – 0.82)
between H and E for triggered lightning in New Mexico was given by H = 3900E-1.33
where H is in meters and E in kV m-1.
In Florida it was found that lightning can be initiated with grounded triggering
wires approximately 400 m long when the ambient fields aloft are as small as 13 kV m-1.
When lightning occurred, ambient potentials with respect to earth at the triggering-rocket
altitude were 3.6 MV (negative with respect to earth). These potentials are referred to as
triggering potentials. The first measurable current pulses at the bottom of the triggering
wires were observed at similar fields aloft but at wire heights only about half as large, the
corresponding potential being 1.3 MV.
18
2.5.2 Altitude Triggering
A stepped leader followed by a first return stroke in natural downward lightning
can be reproduced to some degree by triggering lightning via a metallic wire not attached
to the ground. This ungrounded-wire technique is usually called altitude triggering. In
this type of lightning, illustrated in Figure 2-8, a bidirectional (positive charge up and
negative charge down) leader process is involved in initiating the first return stroke from
ground.
Figure 2-8. Sequence of events in altitude-triggered lightning leading to the establishment
of a relatively low-resistance connection between the upward-moving positive leader tip and the ground. The processes that follow the sequence shown, an initial continuous-current and possibly one or more downward-leader--upward-return-stroke sequences, are similar to their counterparts in classical triggered lightning. The rocket speed is of the order of 102 m s-1. Adapted from Rakov et al. (1998)
Note that the “gap”, in this case, the length of the insulating Kevlar cable, between
the bottom end of the upper (triggering) wire and the top end of the grounded
19
(intercepting) wire is some hundreds of meters. Altitude triggering can also be
accomplished without using an intercepting wire, whose only function is to increase the
probability of lightning attachment to instrumented rocket-launching facility.
2.6 Previous Studies of Displacement Current Associated with Triggered Lightning
Schnetzer et al. (1998) in a study reported only in a conference proceeding, carried
out a numerical evaluation of Maxwell’s integral equation relating azimuthal magnetic
field to its sources to investigate the relative contributions of conduction and
displacement currents to the total ground level azimuthal magnetic field at a distance of
50 m from the base of a triggered lightning stroke. The results revealed differences
between observed magnetic fields measured at 50 m from expectations based on
Ampere’s law for magnetostatics. It is concluded that Ampere’s law for magnetostatics,
which neglects displacement current contributions, provides an inadequate representation
of the total magnetic field due to a lightning ground stroke at distances beyond
approximately 30 m. The rest of this section contains an overview of Schnetzer et al.
(1998), which is the only publication, as of today, on the subject of displacement current
in triggered lightning.
2.6.1 Theory
The following simple expression based on Ampere’s law for magnetostatics is
often used to estimate the azimuthal magnetic field intensity at the ground due to nearby
lightning:
H(r,t)=I(t)/2πr (2.1)
where I(t) is the channel current, H(r,t) is the magnetic field intensity, and r is the radial
distance from the strike point on the earth. Schnetzer et al. (1998) observed that beyond
30 m, peak amplitudes of the measured magnetic fields are somewhat lower than their
20
counterparts predicted by equation (2.1) when measured currents are substituted in it. To
understand the sources of this observed behavior, a first-order numerical evaluation was
carried out using as a starting point Maxwell’s integral equation relating the azimuthal
magnetic field to its arbitrary, time-varying sources,
ol S S
EH dl J d S d St
ε ∂⋅ = ⋅ + ⋅
∂∫ ∫ ∫
oS
EI d St
ε ∂= + ⋅
∂∫ (2.2)
Here the first term on the right corresponds to I(t) in Eq. 2.1. The second term
constitutes the contribution to the azimuthal magnetic field due to displacement current
through the integration surface S bounded by l. When applied to a lightning stroke to
earth with S being defined as a flat, circular area of integration lying on the surface of the
ground and centered at the strike point, the first terms accounts for the effects of
conduction currents flowing normally through the designated surface. These would
include corona and upward leader currents prior to the attachment of the descending
leader, and the return-stroke current after that time. The second term accounts for the
effects of time variations of the vertical electric field associated with the changing charge
density distribution along the lightning channel during both the approach of the leader
and the resultant return stroke.
On the basis of circular symmetry considerations,
( ) 1 ( , )( )2 2
zo
S
I t E r tH t dStr r
επ π
∂= + ⋅
∂∫ (2.3)
To the degree that an average zEt
∂∂
can be treated as constant over the area of integration,
H can be approximated as
21
( ) ( )( )2 2
o zI t r dE tH tr dt
επ
= + = Hc + Hd (2.4)
where Hc and Hd are the contribution to the total magnetic field H from conduction and
displacement current, respectively. Alternately, average values of dEz/dt can be defined
for different ring-shaped parts of S, in which case the second term on the right side of Eq.
2.4 becomes the sum of all individual contributing subareas of S. The first term on the
right of Eq. 2.4 is recognized as corresponding to Ampere’s law for magnetostatics, and
therefore any effects due to the current variation with height and radiation field must be
accounted for by the second, displacement current term. Since the descending leader does
not penetrate the surface of integration, during the leader phase I(t) is approximately zero,
except for assumed-to-be-negligibly-small corona or upward leader currents emitted from
the prospective attachment point. During that time, therefore, the magnetic field intensity
is related solely to the displacement current. Commencing with the onset of the return
stroke, both terms contribute. However, close to the channel, where r → 0, the
displacement current contribution tends to zero, and the total field is dominated by the
contribution from the channel conduction current term.
2.6.2 Estimation of Displacement Current Contribution at 50 m
In order to investigate the distance at which the conduction current term alone
ceases to provide an adequate representation of the total magnetic field, Schnetzer et al
(1998) carried out a numerical evaluation of the relative contributions of both terms in
Eq. 2.4 for a typical 20-kA stroke using available recorded channel-base current and
associated magnetic field at 50 m and electric field data at two distances, 5 and 20 m,
from the strike point. The first step was to derive estimates of the displacement current
for r ranging from 0 to 10 m and from 10 to 50 m from electric fields measured at 5 and
22
20 m, respectively. Because electric fields were not recorded beyond 20 m, estimates of
the displacement current contributed by the annular area between 10 and 50 m were
obtained by assuming that the electric field over that area was, on average, the same as
the value measured at 20 m. Electric field waveforms were numerically differentiated to
obtain dEz/dt waveforms to be substituted in Eq. (2.4). The resulting estimated
amplitudes of displacement currents calculated at four time points prior to and at the
return-stroke onset are listed in Table 2.1.
Table 2-1. Displacement currents Id estimated from measured electric field data at times prior to and at the onset of a 20 kA peak stroke current. Taken from Schnetzer et al. (1998).
A qualitative representation of the electric field derivative is given in Figure 2-9.
According to Table 2-1, prior to the return stroke, while there is no appreciable
conduction current flowing through the integration surface at ground level, the
magnitudes of displacement currents out to 10 m at the various time steps are negligible
relative to the prospective returns stroke peak current of 20 kA. On the other hand, during
23
that same period the displacement current contribution from the area between 10 and 50
m becomes increasingly significant in comparison to 20 kA. Consistent with Figure 2-10,
the displacement current increases in amplitude with time and is negative in polarity up to
the point of onset of the return stroke. Upon initiation of the return stroke, the local
electric field abruptly changes sign and its derivative reaches its peak. The corresponding
magnitude of the total displacement current at that point becomes comparable to that of
the channel current, which then begun to flow but has not reached its peak.
Contribution of the displacement current in addition to the return-stoke channel
conduction current contribution clearly alters both the waveshape and amplitude of the
return-stroke portion of the total magnetic field compared to that predicted by Ampere’s
law for magnetostatics. This is demonstrated in Figure 2-10. Here, the magnetic field
measured at 50 m (labeled Measured) during stroke 5 of flash 96-23 is plotted along with
the magnetic field (labeled Ampere’s Law) that would be predicted by Eq. 2.1 when
applied to the recorded channel-base current of that stroke. Finally, the difference
between the two, which is qualitatively consistent with the results of Table 2-1 and
apparently represents the contribution due to the displacement current term of Eq. 2.4, is
also plotted. As can be seen in Figure 2-10, the shape of the difference curve corresponds
well to the anticipated shape of the electric field derivative, and hence of displacement
current density, illustrated in Figure 2-9. Prior to the return stroke, while no channel
current is flowing at ground level, the effect of the displacement current is apparent in the
initial ramp of the measured magnetic field. Further effects of the displacement current
component on the total field are the rounding and reduction in amplitude of the peak of
the magnetic field waveform.
24
Figure 2-9. Qualitative illustration of the shape of the time derivative of the ground level
vertical electric field of a nearby return stroke . Since displacement current density is simply related to the derivative of electric field by a constant, it has the same waveshape. Taken from Schnetzer et al. (1998).
Figure 2-10. From top to bottom, magnetic field measured at 50 m during stroke 5 of
Flash 96-23, the corresponding field according to Ampere’s law for magnetostatics, as applied to the measured channel-base current, and the difference between the two. Taken from Schnetzer et al. (1998).
25
CHAPTER 3 CHARACTERIZATION OF EXPERIMENTAL DATA USED IN THIS STUDY
3.1 Magnetic Field Measuring Techniques
To measure the magnetic field from lightning a loop of wire can be used as an
antenna. According to Faraday’s Law a changing magnetic field passing through an open
circuited loop of wire will induce a voltage at the terminals of the wire. The voltage at the
terminals of the wire is
dtdBAv n
out =
where A is the area of the antenna and B is the magnetic flux density passing through the
loop, perpendicular to the plane of the loop. Since the output voltage is proportional to
the time derivate of the magnetic field, this voltage will have to be integrated to obtain
the signal proportional to the field. The Thevenin equivalent circuit of a loop antenna is
the open circuit source voltage dBAdt
in series with the source impedance (primarily
inductive). The Thevenin equivalent is shown in Figure 3-1.
Figure 3-1. Thevenin equivalent circuit of a loop antenna
26
A Norton equivalent circuit model for the antenna can be derived from the
Thevenin equivalent with thV Aj B ABIj L j L L
ωω ω
= = = . The Norton equivalent circuit is
shown in Figure 3-2
Figure 3-2. Norton equivalent circuit for a loop antenna.
Since the loop antenna has an inductance L associated with it, the impedance of the
antenna will change with frequency. In the frequency domain that impedance is ωL
where ω is the angular frequency. This frequency-dependent impedance will cause
distortion in the derivative signal. To eliminate the distortion a resistor can be placed in
series with the antenna with the resistive impedance R much higher than the inductive
impedance ωL of the antenna at the highest frequency of interest. The decay time
constant of the overall circuit in Figure 3-1 and Figure 3-2 will be τ = RC as long as R is
much smaller then the input resistance of the recorder and C is much larger then the input
capacitance of the recorder, both conditions being usually met. In the frequency domain
the output voltage across the capacitor is
1( )( )
1out
A j Bj CV
R j Lj C
ωω
ωω
=+ +
27
If we choose R >> jωL and R >> 1j Cω
for the highest and lowest frequencies of
interest respectively, as discussed above, then outABVRC
= . In this way magnetic field from
lightning can be measured from a loop antenna. Using the output voltage, Vout, area of the
loop antenna A, and choosing appropriate resistance R the magnetic field B can be
estimated.
3.2 Experimental Setup
3.2.1 ICLRT Overview
In this section, an overview of International Center for Lightning Research and
Testing (ICLRT) is presented. The ICLRT is located at Camp Blanding, Florida, at
coordinates 29°56’ N, 82° 02’ W, 8 km east of Starke. It was constructed by Power
Technologies in 1993 to study the effect of lightning on power lines. It has been operated
by the University of Florida since 1994. The rocket-and-wire technique (e.g., Rakov et
al., 1998) was used to artificially initiate (trigger) lightning from natural thunderclouds.
An overview of the ICLRT during the 1999, 2000, and 2001 experiments is found in
Figure 3-3 and a photograph of Flash S0012, triggered in 2000, in Figure 3-4. Flash
S0012 was triggered using the underground launcher. The triggered-lightning
experiments are usually conducted from May through September. As can be seen from
Figure 3-3, the ICLRT includes a tower launcher 11 m in height. Other launchers placed
at different positions on the site were also used in different years. Apart from the
launchers, the ICLRT includes test overhead power lines, under-ground cables, four
instrumentation stations, a test house, test runway, other test objects and systems and
various office and storage trailers.
28
Figure 3-3. An overview of the ICLRT at Camp Blanding, Florida, 1999-2001. Not all
test objects are shown.
Figure 3-4. Photograph of lightning flash S0012 triggered from the underground launcher
(see Figure 3-3).
29
3.2.2 1997 Experiments
In this section, a brief discussion of instrumentation used in the 1997 multiple
station experiment is discussed. The multiple station experiment involved a 5 m rocket
launcher located at northeast corner of the ICLRT and grounded through three 8 foot
(2.43 m) galvanized steel rods driven into sandy soil. Seven field measuring stations were
located (see Figure 3-5) south of the rocket launcher with distances ranging from 5 to 500
m. Each station had electric field and magnetic field antennas. The lightning channel base
current was measured using 0.5 milliohm CVR (Current Viewing Resistor) located at the
launcher. Three digitizing systems (see Figure 3-6) were used in this experiment, two
were provided by the University of Florida and were located at SATTLIF trailer (see
Figure 3-5) and launch control, the third digitizing system was provided by Sandia and
was located at SATTLIF trailer. When the current measured by the CVR located at
SATTLIF launcher exceeded 4 kA a digital pulse was generated by a current threshold
triggering circuit, which triggered all three digitizing systems. The digitizing system at
SATTLIF trailer consisted of a single Nicolet Multipro 150 digitizer and five LeCroy
model RM9400 digital oscilloscopes. Nicolet Multipro consisted of 8 data acquisition
cards of 4 channels each, but for this experiment only two cards were used of which one
was faulty. Each channel had 12-bit resolution with a sampling rate of 10 MHz for a total
record length of 51.2 ms. The LeCroy oscilloscopes were segmented for multiple triggers
and provided 8-bit resolution with a sampling rate of 25 MHz for a total record length of
100 µs.
At launch control five Nicolet model pro90 digital oscilloscopes were used. Pro90
consists of 4 channels; channels 1 and 2 had 8-bit resolution and a sampling rate up to
200 MHz, and channels 3 and 4 had 12-bit resolution and a sampling rate up to10 MHz.
30
The first channel had a pretrigger delay of 2 ms and second channel had a 23.6 ms
pretrigger delay that allowed continuous recording of the same quantity for 51.2 ms.
Nicolet Isobe 3000 fiber optic links were used to transmit analog data from antennas to
digitizing systems. More detailed description of 1997 instrumentation can be found in
Chapter 4 of David E. Crawford’s thesis (Crawford 1998).
N
W E
S
SATTLIFLAUNCHER
UF LAUNCHCONTROL
TRAILER
100m
OFFICE ANDSTORAGE
UF TOWER LAUNCHER
SATTLIFTRAILER ANTENNA FIELD
TEST POWERLINE
Figure 3-5. Locations of different instrumentation stations for 1997 multiple station
experiment. Taken from Crawford (1998).
3.2.3 1999 Experiments
The rocket launcher consisted of six metallic tubes aligned vertically from which
rockets were launched. The rocket launcher was mounted on insulating fiberglass legs
and placed underground, with the top of the launcher flush with ground, in a 4 m × 4 m ×
4 m pit. The pit and the launcher were located in the center of a 70 m × 70 m buried
metallic grid (see Figure 3-3) intended to simulate a perfectly conducting ground. This
configuration eliminates ground surface arcing (e.g., Rakov et al., 1998) and minimizes
field propagation effects due to a finite ground conductivity. The low frequency, low
current grounding resistance of the buried grid was measured to be 6 Ω. A hollow metal
rod with an outside diameter of 3.8 cm and a wall thickness of 0.6 cm protruding 1 m
31
ANTENNAS
10 m
UF LAUNCH CONTROL
REC INNICOLET
DSOs10 Channels256 Ksamples/ch10 Msamples/sec12 bit res.
TRIG IN
ANALOG FIELD DATA
110 m
500 m
20 m
30 m
50 m
LEGEND:
ELECTRIC FIELDANTENNA (.155 m²)
MAGNETIC FIELDANTENNA (.27 m²)
DSO = Digital Storage Oscilloscope
SATTLIFLAUNCHER
5 m
TRIG IN10 Channels4 Ksamples/seg25 Msamples/sec8 bit res.
4 Channels512 Ksamples/ch10 Msamples/sec12 bit res.
REC IN
ANALOG FIELD DATAREC IN
NICOLETDIGITIZER
TRIG OUT
TRIG IN
ANALOGMAG TAPERECORDER
GEN. (±4 kA Thresh.)
CHANNEL BASE CURRENT SHUNT
REC IN
FIRE CONTROL
CONTROL
LECROYDSOs
TRIGGER
SATTLIF
LAUNCH
Figure 3-6. Data acquisition system used in the 1997 multiple-station experiment. Taken
from Crawford (1998).
above the ground surface was used as the strike object for flashes S9901–S9918 in 1999.
For the other flashes triggered in 1999 (flashes S9925 - S9935), a 2 m rod was used in
order to increase the probability of lightning attachment to the rod. During 1999, the
underground launcher was connected via four metal straps to the buried grid and the base
of the launcher was connected via two metal straps to a 16.5 m long vertical ground rod
whose low frequency, low current grounding resistance was measured to be 40 Ω.
The electric field, magnetic field, and their time derivatives produced by lightning
strokes were measured 15 and 30 m from the strike rod. A TTL-level digital pulse trigger
signal was generated when the magnetic field sensor located 15 m from the rocket
launcher detected a magnetic field that corresponded to a current of at least 5 kA (using
32
Ampere’s law of magnetostatics). The TTL signal was transmitted to the external trigger
input of the oscilloscopes located in the SATTLIF trailer (see Figure 3-3) to trigger the
digitizing system. The oscilloscopes for the field and current measurements had a pre-
trigger (stored data recorded prior to the trigger pulse) that ranged from 10% to 50% of
the total record length. Fiber optic transmitters (FOT) converted the analog output signals
from the antennas to optical signals and transmitted those signals via fiber optic cables to
fiber optic receivers (FOR). Meret and Nanofast fiber optic links (FOL) were used in the
experiment. The bandwidth of the Meret and Nanofast FOL was dc to 35 MHz and 5 Hz
to 175 MHz, respectively. The FOT were powered with 12 V dc lead-acid batteries. RG-
58 or RG-223 coaxial cables (both 50 Ω) connected the FOT to the antennas, which were
located in metal boxes near the antennas. The FORs in the SATTLIF trailer were
powered with 120 V ac uninterruptible power supplies (UPS). RG-58 or RG-223 coaxial
cables connected the FOR to the digitizing system. The optical fibers transmitting the
signal from the FOT to the FOR were 200 µm glass, Kevlar-reinforced, duplex cables.
3.2.3.1 Instrumentation for Current Measurements
Six P110A current transformers (CTs) with a lower frequency response of 1 Hz and
an upper frequency response of 20 MHz were used to measure the current at the lightning
channel base, two measured the current flow to the vertical ground rod, and four
measured the current flow to the buried grid. The current amplitude range of each sensor
is from a few amperes to approximately 20 kA. A passive combiner summed the two
signals from the ground rod CTs to obtain a total ground rod current, and another one
summed the four signals from the buried grid CTs to obtain a total grid current. The total
ground rod current signal and the total grid current signal were then each transmitted via
33
separate Meret FOL (35 MHz bandwidth) to the SATTLIF trailer. Both signals were
filtered with a 20 MHz, 3 dB anti-aliasing filters and then digitized at 50 MHz. The total
current at the lightning channel base was obtained by numerically summing the ground
rod current and the buried-grid current.
3.2.3.2 Instrumentation for Electric Field Measurements
Electric fields were measured using flat plate antennas with an area of 0.16 m2. The
output of each electric field antenna was connected to an integrating capacitor of value
105 nF at 15 m and 55 nF at 30 m. The input impedance of the fiber optic transmitter was
about 1 MΩ. Meret fiber optic links with a bandwidth of 35 MHz were used to transmit
the signal to the SATTLIF trailer where the 15-m electric field signals were filtered using
10 MHz, 3 dB anti-aliasing filters and digitized at 25 MHz, and the 30-m electric field
signals were filtered using 20 MHz, 3 dB anti-aliasing filters and digitized at 25 MHz.
3.2.3.3 Instrumentation for Electric Field Derivative Measurement
Electric field derivatives were measured using flat plate antennas with an area of
0.16 m2. The output of each electric field derivative antenna was terminated using 50 Ω,
and the input impedance of the fiber optic transmitter was about 1 MΩ. Meret fiber optic
links with a bandwidth of 35 MHz were used to transmit the 15-m signal, and a Nanofast
FOL was used to transmit the 30-m signal to SATTLIF trailer. Signals of dE/dt at 15 and
30 m were filtered using 20 MHz, 3 dB anti-aliasing filters and digitized at 250 MHz.
3.2.4 2000 Experiments
The strike object was a 2 m vertical rod surrounded by a 3 m diameter horizontal
ring elevated to 1.5 m height and electrically connected to the base of the rod (see Figure
3-8). Most of the 2000 instrumentation was the same as the 1999 instrumentation except
for the two major changes made in 2000 and listed below: (1) The triggering signal was
34
generated when the current measured at the strike rod base exceeded 3 kA. (2) Current at
the channel base was measured simultaneously by two different methods (a) the total
lightning current was measured using a single current viewing resistor installed just
below the strike object (new measurement), and (b) the current components flowing to
the vertical ground rod and to the buried grid were measured individually and added
numerically to obtain the total lightning current. The current to the ground rod was
measured as in 1999, while the currents to the buried grid were measured with two
current viewing resistors in the two connections rather than with the four current
transformers and four connections used in 1999. Electric field and electric field derivative
instrumentation in 2000 was the same as in 1999.
SATTLIF Pit & Launch Tubes
B(15)
B(30)
E(15)
E(30) dB(30)
dB(15)
dE(15)
dE(30)
M-E15
M-E30
M-B15
M-B30
sh(15)
E(5)
sh(30)
Camcorder
Camcorder
Video & 35-mm Cameras
Edge of Ground Screen
Pockel Cell sh(3)
30 o 30 o
o o 15 15
Figure 3-7 Experimental setup (placement of electric and magnetic field antennas) used
in SATTLIF for 2000 [Courtesy G. Schnetzer].
35
Figure 3-8 Setup of strike rod and ring mounted over launch tubes in 2000.
3.2.5 2001 Experiments
The experiment set-up was the same as in 2000, except for the strike ring and the
ground rod connection were removed. Additionally, the 2 m strike rod was replaced with
4.5 m section of gas pipeline for the experiments on August 18, 2001. The pipeline was
mounted directly onto the base of the incident current CVR (current viewing resistor) and
nylon fishing line was used to support the structure. The pipeline consisted of four
sections of different diameters, the sections being connected by three different insulating
joints. Current was measured in the same way as in 2000, but was digitized at 25 MHz
(instead of 50 MHz as in 1999 and 2000). The electric field and electric field derivative
instrumentation in 2001 was the same as in 1999 and 2000.
36
3.3 Data Presentation
3.3.1 General Information
In this section the statistical distributions of measured peak magnetic fields, return-
stroke peak currents and peak dE/dt fields are presented. Corresponding waveforms are
shown in Figures 3-9 to 3-13.
For measured return-stroke current which are not saturated the sample size is 88.
For 1997, 1999, 2000, and 2001 data combined, the mean value of peak current is 15.3
kA and standard deviation is 7.50 kA.
The magnetic fields are used in Chapter 4 to estimate the leader and return-stroke
currents. For measured peak magnetic fields which are not saturated at 15 m for 1999,
2000 and 2001 the mean value is 201 µWb/m2 and the standard deviation is 96.8
µWb/m2. At 30 m for 1997, 1999, 2000 and 2001 the mean value is 104 µW/m2 and the
standard deviation is 49.2 µW/m2. The sample size for 15 m data is 97, and for 30 m data
the sample size is 103. Histograms are shown in Figures 3.17-3.22.
For measured dE/dt waveforms which are not saturated the sample size is 50 for
both 15 and 30 m. At 15 m for 1999, 2000 and 2001 the mean value is 355 kV/m/µs and
the standard deviation is 154 kV/m/µs. At 30 m for those three years the mean value is
118 kV/m/µs and the standard deviation is 56 kV/m/µs.
Table 3-1 summarizes the mean, standard deviation, geometric mean and sample
sizes for 1999, 2000 and 2001 combined.
37
Table 3-1. Summary of mean, standard deviation, GM (Geometric Mean), and sample sizes of measured peak current, peak magnetic field and electric field derivative
Parameter Mean St. Dev. GM Sample size
Peak current, kA 15.3 7.5 13.4 88
Peak magnetic field at 15 m, µWb/m2
201 96.8 179 97
Peak magnetic field at 30 m, µWb/m2
104 49.2 92.9 103
Peak dE/dt field at 15 m, kV/m/µs
355 145 315 50
Peak dE/dt field at 30 m, kV/m/µs
118 56 103 50
Figure 3-9. Measured channel-base current, Flash S9901, Stroke 3.
38
Figure 3-10. Magnetic field at 15 m, Flash S9901, Stroke 3.
Figure 3-11. Magnetic field at 30 m, Flash S9901, Stroke 3.
39
Figure 3-12. Electric field derivative (dE/dt) at 15 m, Flash S9901, Stroke 3.
Figure 3-13. Electric field derivative (dE/dt) at 30 m, Flash S9901, Stroke 3.
40
3.3.2 Channel-base Current
0 4 8 12 16 20 24 28 32 36 40
Num
ber
0
2
4
6
8
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32 36 40
Num
ber
0
2
4
6
8
10
12
Return-Stroke Peak Current [kA] 0 4 8 12 16 20 24 28 32 36 40
Num
ber
0
1
2
3
4
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32 36 40
Num
ber
0
1
2
3
4
Return-Stroke Peak Current [kA]
(a) (b)
(c) (d)
1997 1999
20012000
Mean 13.1 kASt. Dev. 5.60 kAMin 5.70 kAMax 23.0 kAGM 12.1 kASample Size 11
Mean 16.7 kASt. Dev. 6.80 kAMin 2.80 kAMax 30.0 kAGM 15.1 kASample Size 24
Mean 13.2 kASt. Dev. 6.90 kAMin 1.10 kAMax 36.8 kAGM 11.3 kASample Size 40
Mean 21.1 kASt. Dev. 9.30 kAMin 9.20 kAMax 38.9 kAGM 19.3 kASample Size 13
Figure 3-14. Return-stroke peak currents in (a) 1997, (b) 1999, (c) 2000, and (d) 2001.
41
0 4 8 12 16 20 24 28 32 36 40
Num
ber
0
2
4
6
8
10
12
14
16
18
20
1999 n = 24 Mean = 16.7 kA St. Dev. = 6.80 kA2000 n = 40 Mean = 13.2 kA St. Dev. = 6.90 kA2001 n = 13 Mean = 21.1 kA St. Dev. = 9.30 kA1997 n = 11 Mean = 13.1 kA St. Dev. = 5.60 kA
Return-Stroke Peak Current [kA]
1997, 1999, 2000, & 2001
Mean 15.3 kASt. Dev. 7.54 kAMin 1.05 kAMax 38.9 kAGM 13.4 kASample Size 88
Figure 3-15. Return-stroke peak currents in 1997, 1999, 2000 and 2001.
42
3.3.3 Magnetic Field
0 200 400 600 800 1000
Num
ber
0
1
2
3
4
Peak Magnetic Field [µWb/m2]
Distance 5 m, 1997
Mean 511 µWb/m2
St. Dev. 274 µWb/m2
Min 176 µWb/m2
Max 916 µWb/m2
GM 443 µWb/m2
Sample Size 7
0 100 200 300 400 500
Num
ber
0
1
2
3
4
5
Peak Magnetic Field [µWb/m2]
Distance 10 m, 1997
Mean 247 µWb/m2
St. Dev. 101 µWb/m2
Min 100 µWb/m2
Max 395 µWb/m2
GM 227 µWb/m2
Sample Size 9
0 50 100 150 200 250
Num
ber
0
1
2
3
4
5
Peak Magnetic Field [µWb/m2]
Distance 20 m, 1997
Mean 111 µWb/m2
St. Dev. 47.3 µWb/m2
Min 51.3 µWb/m2
Max 183 µWb/m2
GM 103 µWb/m2
Sample Size 7
0 25 50 75 100 125 150
Num
ber
0
1
2
3
4Distance 30 m, 1997
Peak Magnetic Field [µWb/m2]
Mean 79.8 µWb/m2
St. Dev. 33.7 µWb/m2
Min 37.8 µWb/m2
Max 118 µWb/m2
GM 73.4 µWb/m2
Sample Size 6
(a) (b)
(c) (d)
Figure 3-16. Peak magnetic fields measured at (a) 5 m, (b) 10 m, (c) 20 m, and (d) 30 m
in 1997.
43
0 50 100 150 200 250 300 350 400 450 500
Num
ber
0
2
4
6
8
10
12
Peak Magnetic Field [µWb/m2]
Distance 15 m, 1999
Mean 223 µWb/m2
St. Dev. 96.5 µWb/m2
Min 74.1 µWb/m2
Max 434 µWb/m2
GM 202 µWb/m2
Sample Size 39
(a)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
2
4
6
8
10
12Distance 30 m, 1999
Peak Magnetic Field [µWb/m2]
Mean 118 µWb/m2
St. Dev. 48.2 µWb/m2
Min 46.4 µWb/m2
Max 225 µWb/m2
GM 108 µWb/m2
Sample Size 39
(b)
Figure 3-17. Peak magnetic fields measured at (a) 15 m and (b) 30 m in 1999.
44
0 50 100 150 200 250 300 350 400 450 500
Num
ber
0
2
4
6
8
10
12
14
Peak Magnetic Field [µWb/m2]
Distance 15 m, 2000
Mean 180 µWb/m2
St. Dev. 92.0 µWb/m2
Min 49.5 µWb/m2
Max 467 µWb/m2
GM 159 µWb/m2
Sample Size 47
(a)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
2
4
6
8
10
12Distance 30 m, 2000
Peak Magnetic Field [µWb/m2]
Mean 93.1 µWb/m2
St. Dev. 47.2 µWb/m2
Min 25.4 µWb/m2
Max 235 µWb/m2
GM 82.5 µWb/m2
Sample Size 47
(b)
Figure 3-18. Peak magnetic fields measured at (a) 15 m and (b) 30 m in 2000.
45
0 50 100 150 200 250 300 350 400 450 500
Num
ber
0
1
2
3
4
5
Peak Magnetic Field [µWb/m2]
Distance 15 m, 2001
Mean 212 µWb/m2
St. Dev. 107 µWb/m2
Min 97.6 µWb/m2
Max 470 µWb/m2
GM 192 µWb/m2
Sample Size 11
(a)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
1
2
3
4Distance 30 m, 2001
Peak Magnetic Field [µWb/m2]
Mean 113 µWb/m2
St. Dev. 56.7 µWb/m2
Min 48.8 µWb/m2
Max 248 µWb/m2
GM 102 µWb/m2
Sample Size 11
(b)
Figure 3-19. Peak magnetic fields measured at (a) 15 m and (b) 30 m in 2001.
46
0 50 100 150 200 250 300 350 400 450 500
Num
ber
0
5
10
15
20
25
30
1999 n = 39 Mean = 223 µWb/m2 St. Dev. = 96.5 µWb/m2
2000 n = 47 Mean = 180 µWb/m2 St. Dev. = 92.0 µWb/m2
2001 n = 11 Mean = 212 µWb/m2 St. Dev. = 107 µWb/m2
Peak Magnetic Field [µWb/m2]
Distance 15 m (1999, 2000, & 2001)
Mean 201 µWb/m2
St. Dev. 96.8 µWb/m2
Min 49.5 µWb/m2
Max 470 µWb/m2
GM 179 µWb/m2
Sample Size 97
(a)
Distance 30 m (1997, 1999, 2000, and 2001)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
5
10
15
20
25
30
1999 n = 39 Mean = 118 µWb/m2 St. Dev. = 48.2 µWb/m2
2000 n = 47 Mean = 93.1 µWb/m2 St. Dev. = 47.2 µWb/m2
2001 n = 11 Mean = 112 µWb/m2 St. Dev. = 56.7 µWb/m2
1997 n = 6 Mean = 79.8 µWb/m2 St. Dev. = 33.7 µWb/m2
Peak Magnetic Field [µWb/m2]
Mean 104 µWb/m2
St. Dev. 49.2 µWb/m2
Min 25.4 µWb/m2
Max 248 µWb/m2
GM 92.9 µWb/m2
Sample Size 103
(b)
Figure 3-20. Peak magnetic fields measured at (a) 15 m and (b) 30 m in 1997, 1999, 2000 and 2001.
47
3.3.4 Electric Field Derivative (dE/dt)
0 100 200 300 400 500 600 700 800
Num
ber
0
1
2
3
4
5
Peak dE/dt [kV/m/µs]
Distance 15 m, 1999 Mean 304 kV/m/µsSt Dev 107 kV/m/µsMin 72.0 kV/m/µsMax 450 kV/m/µsGM 280 kV/m/µsSample Size 15
(a)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
1
2
3
4
5
6
7Distance 30 m, 1999
Peak dE/dt [kV/m/µs]
(b) Mean 92.6 kV/m/µsSt Dev 31.0 kV/m/µsMin 23.0 kV/m/µsMax 145 kV/m/µsGM 86.1 kV/m/µsSample Size 15
Figure 3-21. Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 1999.
48
0 100 200 300 400 500 600 700 800
Num
ber
0
2
4
6
8
10
Peak dE/dt [kV/m/µs]
Distance 15 m, 2000
Mean 350 kV/m/µsSt Dev 156 kV/m/µsMin 50.0 kV/m/µsMax 673 kV/m/µsGM 308 kV/m/µsSample Size 31
(a)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
2
4
6
8Distance 30 m, 2000
Peak dE/dt [kV/m/µs]
Mean 178 kV/m/µsSt Dev 56.0 kV/m/µsMin 16.0 kV/m/µsMax 235 kV/m/µsGM 102 kV/m/µsSample Size 31
(b)
Figure 3-22. Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 2000.
49
0 100 200 300 400 500 600 700 800
Num
ber
0
1
2
3
4
Peak dE/dt [kV/m/µs]
Distance 15 m, 2001
Mean 590 kV/m/µsSt Dev 61.0 kV/m/µsMin 519 kV/m/µsMax 644 kV/m/µsGM 579 kV/m/µsSample Size 4
(a)
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
1
2
3
4Distance 30 m, 2001
Peak dE/dt [kV/m/µs]
Mean 210 kV/m/µsSt Dev 27.0 kV/m/µsMin 181 kV/m/µsMax 241 kV/m/µsGM 209 kV/m/µsSample Size 4
(b)
Figure 3-23 Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 2001.
50
0 100 200 300 400 500 600 700 800
Num
ber
0
2
4
6
8
10
12
14
1999 n = 15 Mean = 304 kV/m/µs St. Dev. = 107 kV/m/µs 2000 n = 31 Mean = 350 kV/m/µs St.Dev. = 156 kV/m/µs2001 n = 4 Mean = 590 kV/m/µs St. Dev. = 61.0 kV/m/µs
Peak dE/dt [kV/m/µs]
Distance 15 m (1999, 2000, & 2001)
(a) Mean 355 kV/m/µsSt Dev 154 kV/m/µsMin 50.0 kV/m/µsMax 673 kV/m/µsGM 315 kV/m/µsSample Size 50
0 25 50 75 100 125 150 175 200 225 250
Num
ber
0
2
4
6
8
10
12
1999 n = 15 Mean = 92.6 kV/m/µs St. Dev. = 31.0 kV/m/µs2000 n = 31 Mean = 118 kV/m/µs St. Dev. = 56.0 kV/m/µs2001 n = 4 Mean = 211 kV/m/µs St. Dev. = 27.0 kV/m/µs
Distance 30 m (1999, 2000, & 2001)
Peak dE/dt [kV/m/µs]
(b) Mean 118 kV/m/µsSt Dev 56.0 kV/m/µsMin 16.0 kV/m/µsMax 241 kV/m/µsGM 103 kV/m/µsSample Size 50
Figure 3-24 Peak dE/dt fields measured at (a) 15 m and (b) 30 m in 1999, 2000 and 2001.
51
CHAPTER 4 ESTIMATION OF LEADER AND RETURN-STROKE CURRENTS FROM
MEASURED MAGNETIC FIELDS
4.1 Introduction
Magnetic fields measured at 15 and 30 m are available for 1999, 2000, and 2001.
In 1997, magnetic field records were measured at 5, 10, 20, and 30 m.
4.2 Estimation of Currents Using Ampere’s Law
According to Ampere’s law for magnetostatics, enc0Iµ=•∫ dlB , the line integral
of magnetic flux density, B, around a closed path is proportional to the enclosed current,
Ienc. The constant of proportionality µo is the permeability of free space. In SI units µo =
4π x 10-7 H/m.
For a straight wire of infinite length the magnetic flux density external to the wire
at a distance r from its axis is given by
r2IB o
πµ
=φ (4.1)
Figure 4-1. A straight current channel of infinite length and B at a distance r.
52
Here we have used the fact that the magnetic field is constant and tangential at any point
on the circular integration path. Equation (4.1) is essentially the same as equation (2.1).
If we assume that earth is a perfect conductor, and consider a very close
observation point (a distance 15 or 30 m compared to the lightning channel of some
kilometers) (4.1) will apply to a lightning channel above ground. Thus, knowing the
magnetic field at distance r, from the lightning channel and using (4.1) one can estimate
the current flowing in the channel if the magnetostatic approximation is valid.
Typical magnetic field waveforms with respect to time are shown in Figures 3.10
and 3.11. They usually exhibit a relatively slow initial front followed by a fast rise to
peak and then decay, as discussed below.
Figure 4-2. A streaked-image diagram of a dart leader—return-stroke sequence in a
rocket-triggered lightning flash
53
As the downward leader tip approaches ground surface the magnetic field on
ground associated with the leader current increases with time. When the leader attaches to
the ground, return stroke is initiated. Hence the current increases abruptly and there is a
fast rise in magnetic field.
The initial rising portion (up to the peak) of each magnetic field record can be
divided into two parts: a slow rise part corresponding to leader and steep rise part
corresponding to the return stroke. The division is illustrated in Figures 4-3 and 4-4,
where the magnitude of the leader part is labeled BL and that of the return stroke part BRS.
For each part the corresponding current was estimated using equation 4.1. It is important
to note that the leader current corresponds to the final stage of the dart-leader process,
when leader channel attaches to the ground, and the return-stroke current represents the
peak value based on the assumption that the leader current continues to flow in the
channel during the return-stroke process. Thus, the total peak current flowing in the
channel during the return-stroke stage is the sum of leader current inferred from BL and
return stroke current estimated from BRS.
Using this approach, leader and the return stroke currents for the years of 1997,
1999, 2000 and 2001 are evaluated. The histograms showing the statistics of leader and
return-stroke currents for 1997, 1999, 2000 and 2001 are shown in Figure 4-5 through
Figure 4-9.
The scatter plots (see Figures 4-13 and 4-14) show comparisons of leader currents
obtained from dE/dt (Kodali et al. 2005) vs. those obtained from magnetic fields (BL) and
of measured return stroke currents (see section 3.3) vs. those obtained from magnetic
fields (BRS).
54
Figure 4-3 Magnetic field at 15 m, Flash S9901, Stroke 3.
Figure 4-4. Magnetic field at 30 m, Flash S9901, Stroke 3.
55
Distance 5 m, 1997
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5
6
Distance 10 m, 1997
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5
6
Distance 20 m, 1997
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5
6Distance 30 m, 1997
Leader Current [kA]
0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5
6
(a) (b)
(c) (d)
Mean 1.80 kASt Dev 1.08 kAMin 0.94 kAMax 3.95 kAGM 1.58 kASample Size 7
Mean 1.80 kASt Dev 0.70 kAMin 1.05 kAMax 2.84 kAGM 1.69 kASample Size 9
Mean 2.02 kASt Dev 1.00 kAMin 1.10 kAMax 3.93 kAGM 1.84 kASample Size 7
Mean 2.06 kASt Dev 0.72 kAMin 1.34 kAMax 3.11 kAGM 1.95 kASample Size 6
Figure 4-5. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields at (a) 5 m, (b) 10 m, (c) 20 m, and (d) 30 m in 1997.
56
Distance 15 m, 1999
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
2
4
6
8
10
12
14
16
18Mean 2.25 kASt. Dev. 1.02 kAMin 0.52 kAMax 4.29 kAGM 2.01 kASample Size 39
(a)
Distance 30 m, 1999
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
2
4
6
8
10
12
14
16Mean 3.21 kASt. Dev. 1.61 kAMin 0.71 kAMax 7.87 kAGM 2.82 kASample Size 39
(b)
Figure 4-6. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields at (a) 15 m and (b) 30 m in 1999.
57
Distance 15 m, 2000
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
5
10
15
20
Mean 1.53 kASt. Dev. 0.82 kAMin 0.38 kAMax 3.83 kAGM 1.32 kASample Size 47
(a)
Distance 30 m, 2000
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
2
4
6
8
10
12
14
16Mean 2.13 kASt. Dev. 1.30 kAMin 0.40 kAMax 6.36 kAGM 1.80 kASample Size 47
(b)
Figure 4-7. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields at (a) 15 m and (b) 30 m in 2000.
58
Distance 15 m, 2001
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5Mean 2.03 kASt. Dev. 1.28 kAMin 0.85 kAMax 5.45 kAGM 1.76 kASample Size 11
(a)
Distance 30 m, 2001
Leader Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5Mean 2.57 kASt. Dev. 2.18 kAMin 0.87 kAMax 8.63 kAGM 2.04 kASample Size 11
(b)
Figure 4-8. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields at (a) 15 m and (b) 30 m in 2001.
59
Distance 15 m, 1999,2000,&2001
Leader Current [kA]
0 1 2 3 4 5 6 7 8 9
Num
ber
0
10
20
30
40
50
1999 n = 39 Mean = 2.25 kA St. Dev. = 1.02 kA2000 n = 47 Mean = 1.53 kA St. Dev. = 0.82 kA2001 n = 11 Mean = 2.03 kA St. Dev. = 1.28 kA
Mean 1.87 kASt. Dev. 1.01 kAMin 0.38 kAMax 5.40 kAGM 1.62 kASample Size 97
Figure 4-9. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields at 15 m in 1999, 2000, and 2001.
60
Distance 30 m, 1997, 1999, 2000, and 2001
Leader Current [kA]
0 1 2 3 4 5 6 7 8 9
Num
ber
0
5
10
15
20
25
30
35
1999 n = 39 Mean = 3.21 kA St. Dev. = 1.61 kA2000 n = 47 Mean = 2.13 kA St. Dev. = 1.30 kA2001 n = 11 Mean = 2.57 kA St. Dev. = 2.18 kA1997 n = 6 Mean = 2.10 kA St. Dev. = 0.70 kA
Mean 2.58 kASt. Dev. 1.57 kAMin 0.40 kAMax 8.63 kAGM 2.17 kASample Size 103
Figure 4-10. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields at 30 m in 1997, 1999, 2000, and 2001.
61
IL from B15 [kA]
0 2 4 6 8 10
I L fro
m B
30 [k
A]
0
2
4
6
8
10
199920002001
IL(30 m) = -0.10 + 1.44 IL(15 m) R2 = 0.83 Sample Size = 97
Figure 4-11. Dart leader current inferred using Ampere’s Law for magnetostatics from
measured magnetic fields measured at 30 m vs. that at 15 m in 1999, 2000, and 2001.
62
IL from dE/dt [kA]
0 2 4 6 8 10 12
I L fro
m B
15 [k
A]
0
2
4
6
8
10
12
199920002001
IL(Ampere’s Law) = 1.23 + 0.32 IL(dE/dt)R2 = 0.51 Sample Size = 39
IL(dE/dt) [kA] IL(Amperes Law) [kA] Mean 3.51 2.36 StDev 2.39 1.07
Figure 4-12. Dart leader current inferred using Ampere’s Law for magnetostatics from
magnetic fields measured at 15 m vs. leader current inferred from dE/dt measurements [Kodali et al. 2005] in 1999, 2000, and 2001.
63
IL from dE/dt [kA]
0 2 4 6 8 10
I L fro
m B
30 [k
A]
0
2
4
6
8
10
19992000 2001
IL(Ampere’s Law) = 3.6 + 0.44 IL(dE/dt) R2 = 0.56 Sample Size = 6
IL(dE/dt) [kA] IL(Amperes Law) [kA] Mean 4.98 5.86 StDev 2.90 1.71
Figure 4-13. Dart leader current inferred using Ampere’s Law for magnetostatics from
magnetic fields measured at 30 m vs. leader current inferred from dE/dt measurements [Kodali et al. 2005] in 1999, 2000, and 2001.
64
Distance 5 m, 1997
Return-Stroke Peak Current [kA]0 4 8 12 16 20 24 28 32
Num
ber
0
1
2
3
4
Distance 10 m, 1997
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
1
2
3
4
Distance 20 m, 1997
Return-Stroke Peak Current [kA]0 4 8 12 16 20 24 28 32
Num
ber
0
1
2
3
4
5
Distance 30 m, 1997
Return-Stroke Peak Current [kA]0 4 8 12 16 20 24 28 32
Num
ber
0
1
2
3
4
5
(a) (b)
(c) (d)
Mean 11.0 kASt Dev 5.95 kAMin 3.33 kAMax 19.0 kAGM 9.37 kASample Size 7
Mean 10.5 kASt Dev 4.44 kAMin 3.69 kAMax 17.0 kAGM 9.57 kASample Size 9
Mean 9.13 kASt Dev 3.83 kAMin 4.03 kAMax 14.4 kAGM 8.42 kASample Size 7
Mean 9.91 kASt Dev 4.57 kAMin 4.34 kAMax 15.9 kAGM 8.97 kASample Size 6
Figure 4-14. Return-stroke peak currents inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at (a) 5 m, (b) 10 m, (c) 20 m, and (d) 30 m in 1997.
65
Distance 15 m, 1999
Return-Stroke Peak Current [kA]0 4 8 12 16 20 24 28 32
Num
ber
0
2
4
6
8
10
12
14
(a) Mean 14.5 kASt Dev 6.40 kAMin 4.88 kAMax 28.7 kAGM 13.1 kASample Size 39
Distance 30 m, 1999
Return-Stroke Peak Current [kA]0 4 8 12 16 20 24 28 32
Num
ber
0
2
4
6
8
10
12
14(b) Mean 14.5 kA
St Dev 5.80 kAMin 5.70 kAMax 26.7 kAGM 13.4 kASample Size 39
Figure 4-15. Return-stroke peak currents inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at (a) 15 m and (b) 30 m in 1999.
66
Distance 15 m, 2000
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
2
4
6
8
10
12
14
Mean 12.0 kASt. Dev. 6.19 kAMin 3.34 kAMax 31.8 kAGM 10.6 kASample Size 47
(a)
Distance 30 m, 2000
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
2
4
6
8
10
12
14
16Mean 11.8 kASt. Dev. 6.91 kAMin 3.09 kAMax 28.9 kAGM 10.5 kASample Size 47
(b)
Figure 4-16. Return-stroke peak currents inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at (a) 15 m and (b) 30 m in 2000.
67
Distance 15 m, 2001
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
1
2
3
4
5Mean 13.8 kASt. Dev. 6.90 kAMin 6.47 kAMax 29.8 kAGM 12.5 kASample Size 11
(a)
Distance 30 m, 2001
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
1
2
3
4
5Mean 14.3 kASt. Dev. 6.57 kAMin 6.45 kAMax 28.6 kAGM 13.1 kASample Size 11
(b)
Figure 4-17. Return-stroke peak currents inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at (a) 15 m and (b) 30 m in 2001.
68
Distance 15 m, 1999, 2000, & 2001
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
5
10
15
20
25
30
1999 n = 30 Mean = 14.5 kA St. Dev. = 6.40 kA2000 n = 47 Mean = 12.0 kA St. Dev. = 6.19 kA2001 n = 11 Mean = 13.8 kA St. Dev. = 6.90 kA
Mean 13.2 kASt. Dev. 6.41 kAMin 3.34 kAMax 31.8 kAGM 11.7 kASample Size 97
Figure 4-18. Return-stroke peak currents inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at 15 m in 1999, 2000, and 2001.
69
Distance 30 m, 1999, 2000, 2001 & 1997
Return-Stroke Peak Current [kA]
0 4 8 12 16 20 24 28 32
Num
ber
0
5
10
15
20
25
30
35
1999 n = 39 Mean = 14.9 kA St. Dev. = 6.30 kA2000 n = 47 Mean = 11.8 kA St. Dev. = 6.91 kA2001 n = 11 Mean = 14.3 kA St. Dev. = 6.57 kA1997 n = 6 Mean = 9.90 kA St. Dev. = 4.60 kA
Mean 13.0 kASt. Dev. 6.00 kAMin 3.09 kAMax 28.9 kAGM 11.7 kASample Size 103
Figure 4-19. Return-stroke peak currents inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at 30 m in 1997, 1999, 2000, and 2001.
70
IRS from B15 [kA]
0 5 10 15 20 25 30 35
I RS
from
B30
[kA]
0
5
10
15
20
25
30
35
199920002001
IRS(30 m) = 0.99 + 0.93 IRS(15 m)R2 = 0.97 Sample Size = 97
Figure 4-20. Return-stroke peak current inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at 30 m vs. that inferred from measured magnetic fields at 15 m in 1999, 2000, and 2001.
71
Measured IRS [kA]
0 5 10 15 20 25 30 35 40
I RS
from
B15
[kA]
0
5
10
15
20
25
30
35
40
199920002001
IRS(Ampere’s Law) = 0.08 + 0.83 IRS(Meas.) R2 = 0.86 Sample Size = 71
IRS(Meas.) [kA] IRS(Amperes Law) [kA] Mean 15.3 13.2 St. Dev. 6.70 6.10
Figure 4-21. Return-stroke peak current inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at 15 m vs. measured return-stroke peak current in 1999, 2000, and 2001.
72
Measured IRS [kA]
0 5 10 15 20 25 30 35 40
I RS
from
B30
[kA]
0
5
10
15
20
25
30
35
40
1999200020011997
IRS(Ampere’s Law) = 0.36 + 0.83 IRS(Meas.) R2 = 0.90 Sample Size = 76
IRS(Meas.) [kA] IRS(Ampere’s Law) [kA]Mean 15.4 13.1 St. Dev. 6.87 5.97
Figure 4-22. Return-stroke peak current inferred using Ampere’s Law for magnetostatics
from measured magnetic fields at 30 m vs. measured return-stroke peak current in 1997, 1999, 2000, and 2001.
73
IRS from B15 [kA]
0 5 10 15 20 25 30 35
I L fro
m B
15 [k
A]
0
1
2
3
4
5
6
7
8
9
199920002001
IL(15 m) = 0.16 + 0.13 IRS(15 m) R2 = 0.68 Sample Size = 97
Figure 4-23. Leader vs. return-stroke currents inferred using Ampere’s Law for
magnetostatics from magnetic fields measured at 15 m in 1999, 2000, and 2001.
74
IRS from B30 [kA]
0 5 10 15 20 25 30 35
I L fro
m B
30 [k
A]
0
1
2
3
4
5
6
7
8
9
199920002001
IL(30 m) = -0.29 + 0.21 IRS(30 m) R2 = 0.72 Sample Size = 97
Figure 4-24. Leader vs. return-stroke currents inferred using Ampere’s Law for
magnetostatics from magnetic fields measured at 30 m in 1999, 2000, and 2001.
75
IRS from (BL+BRS)15m [kA]
0 5 10 15 20 25 30 35 40
I RS fr
om (B
L+B R
S) 30
m [k
A]
0
5
10
15
20
25
30
35
40
IRS from (BL+BRS)30m = 0.53 + 1.01 IRS from (BL+BRS)15m R2 = 0.98 Sample size = 97
Figure 4-25. Comparison of IRS from (BL+BRS)30m and IRS from (BL+BRS)15m for 1999,
2000, and 2001.
76
Measured IRS [kA]
0 10 20 30 40 50
I RS fr
om (B
L+B R
S) 15
m [k
A]
0
10
20
30
40
50
199920002001
IRS from (BL+BRS)15m = 0.07 + 0.95 IRS(Meas.) R2 = 0.87 Sample Size = 71
Figure 4-26. Comparison of IRS from (BL+BRS)15m inferred using Ampere’s Law for
magnetostatics vs. IRS measured at channel base in 1999, 2000, and 2001.
77
Measured IRS [kA]
0 10 20 30 40 50
I RS fr
om (B
L+B R
S) 30
m [k
A]
0
10
20
30
40
50
1999200020011997
IRS from (BL+BRS)30m = -0.02 + 1.01 IRS(Meas.) R2 = 0.89 Sample Size = 76
Figure 4-27. Comparison of IRS from (BL+BRS)30m inferred using Ampere’s Law for
magnetostatics vs. IRS measured at channel base in 1997, 1999, 2000, and 2001.
4.3 Discussion and Summary
From the statistical analysis of 97 records obtained in 1999, 2000, and 2001 the
mean dart-leader current (at the time of its attachment to ground) inferred from leader
magnetic fields measured at 15 m is 1.87 kA and standard deviation is 1.01 kA. The
minimum and maximum dart-leader currents are 0.38 and 5.40 kA, respectively. From
78
magnetic fields measured at 30 m (103 records obtained in 1997, 1999, 2000, and 2001),
the mean leader current is 2.58 kA and standard deviation is 1.57 kA. The minimum and
maximum dart-leader currents are 0.40 and 8.63 kA, respectively. Typically leader
currents are in the range of a few kiloamperes. Idone and Orville (1985) estimated dart-
leader peak currents for 22 leaders in two rocket-triggered flashes using the relation
between return-stroke peak current IR and return-stroke peak relative light intensity LR in
each of two flashes LR = 1.5 (IR)1.6 and LR = 6.4(IR)1.1 to the dart leader relative light
intensities in those flashes. The mean dart-leader current was 1.8 kA and the range was
0.1 to 6.0 kA. Results of this study are in reasonably good agreement with those of Idone
and Orville (1985).
Return-stroke peak currents estimated from magnetic fields (as the global magnetic
field peak minus the leader contribution) measured at 15 m in 1999, 2000, and 2001 have
a mean value of 13.2 kA and the standard deviation is 6.4 kA. For the individual years the
mean return-stroke current varies from 14.5 kA in 1999 to 12.0 kA in 2000 and 13.8 kA
in 2001. Return stroke peak currents estimated from magnetic fields measured at 30 m in
1997, 1999, 2000, and 2001 have a mean value of 13.0 kA and the standard deviation is
6.0 kA. For the individual years the mean return-stroke peak current varies from 9.9 kA
in 1997, to 14.9 kA in 1999, to 11.8 kA in 2000, and to 14.3 kA in 2001. These inferred
return-stroke peak currents are slightly lower than the directly measured return-stroke
peak currents whose mean is 15.3 kA and standard deviation is 7.5 kA. Depasse (1994)
reported an arithmetic mean of 14.3 kA (maximum value of 60 kA, standard deviation of
9 kA) for 305 peak current values directly measured at the Kennedy Space Center (KSC),
Florida and an arithmetic mean of 11 kA (maximum value 49.9 kA, and standard
79
deviation of 5.6 kA) for 54 values directly measured at Saint-Privat d’Allier, France.
Rakov et al. (1998), Crawford (1998), and Uman et al. (2000), reported an arithmetic
mean of 15.1 kA (sample size = 37, maximum = 44.4 kA, and st. dev. = 9 kA), 12. 8 kA
(sample size = 11, maximum = 22.6 kA, and st. dev. = 5.6 kA), and 14.8 kA (sample size
= 25, maximum = 33.2 kA, and st. dev. = 7 kA) from direct current measurements at
Camp Blanding in 1993, 1997, and 1998, respectively.
The return-stroke peak currents obtained using Ampere’s law for magnestotatics
and (BL+BRS), where BL and BRS are the leader and return-stroke contributions to the total
magnetic field (as done, for example, by Schoene et al. (2003)), measured at 15 m are
about 14 % higher than those obtained from BRS alone (excluding the leader
contribution). The return-stroke peak currents obtained using Ampere’s law for
magnestotatics and (BL+BRS) measured at 30 m are about 20 % higher than those
obtained from BRS alone.
Dart-leader current to return-stroke current ratio obtained from magnetic fields
measured at 15 m has a mean of 0.14 and the correlation coefficient between these
currents is 0.82. Leader to return-stroke current ratio obtained from magnetic fields at 30
m has a mean of 0.20 and the correlation coefficient is 0.85 (note that the determination
coefficient, R2, given in Figure 4-23 and Figure 4-24 is the square of the correlation
coefficient, R). Idone and Orville (1985) obtained the ratio of dart leader to return-stroke
current for 22 events having a mean of 0.17. Results of this study are consistent with
those of Idone and Orville (1985).
80
CHAPTER 5 DISPLACEMENT CURRENT ASSOCIATED WITH LEADER/RETURN STROKE
SEQUENCES IN TRIGGERED LIGHTNING
Maxwell’s integral equation relating azimuthal magnetic field to its sources has
been used to investigate the relative contributions of conduction and displacement
currents to the total ground level azimuthal magnetic field at 15 and 30 m from the base
of a triggered lightning stroke. The analysis is based on a comparison of channel base
current and corresponding magnetic fields measured at 15 and 30 m. It is concluded that
Ampere’s law for magnetostatics, which neglects the displacement current contribution,
is inadequate for presentation of the total magnetic field due to lightning strokes at
distances beyond approximately 30 m.
5.1 Displacement Current Estimates from Measured Magnetic Fields and Channel-Base Currents
The following simple expression (same as Eq. (2.1)) based on Ampere’s law for
magnetostatics is used here to estimate the azimuthal magnetic field intensity at the
ground due to nearby lightning:
( , ) ( ) / 2H r t I t rπ= (5.1)
where I(t) is the channel current, assumed to be the same in all contributing channel
sections, H(r,t) is the magnetic field intensity, and r is the radial distance from the
channel termination point on ground. This relationship has been found to describe
measured magnetic fields quite closely, in both magnitude and waveshape, at distances
up to 15 meters from the triggered-lightning channel base. Over the period from 1999 to
2001, magnetic field measurements have been made for nearly 70 return strokes in
81
Florida at a distance of 30 m from the triggered-lightning strike point. It is found that
amplitudes of the measured fields at both 15 and 30 m are somewhat lower then predicted
by Eq. 5.1, and the early portions of their waveforms exhibit slow fronts not seen in
measured current waveforms.
Figure 5-1. Superposition of measured magnetic field and channel-base current for Flash
S9903, stroke 3.
In order to examine the discrepancy, we consider as a starting point Maxwell’s
integral equation relating the azimuthal magnetic field to its arbitrary, time-varying
sources,
. . .ol S S
EH dl J d S d St
ε ∂= +
∂∫ ∫ ∫
.oS
EI d St
ε ∂= +
∂∫ = Ic + Id (5.2)
82
where Ic and Id are the channel-base conduction current and displacement current
respectively. As already described by Eq. 2.4 in section 2.6, the first term on the right
side of Eq. 5.2 corresponds to I(t) in Eq. 5.1. The second term is the displacement current
whose density is normal to the integration surface S bounded by the integration path l, on
the left-hand side of Eq. 5.2. When applied to a lightning stroke to earth with S being
defined as flat, circular area of integration lying on the surface of the ground and centered
at the strike point, the first term accounts for the effect of conduction current flowing
normally through the integration surface. The second term on right side of Eq. 5.2
accounts for the effects of time variation of the vertical electric field associated with the
changing charge density distribution along the lightning channel during both the
approach of the leader and the resultant return stroke.
Equivalently Eq. 5.2 can be written as
2d meas measI rH Iπ= − (5.3)
where Imeas and Hmeas are measured channel-base current and measured magnetic field at
distance r.
In order to investigate the peak displacement current, the following procedure was
used. As a first step, the measured magnetic field converted to current using Eq. 5.1. and
the measured conduction current, Imeas, at the channel base were superimposed. Then the
difference between the current obtained from magnetic field using Eq. 5.1 and directly
measured current was interpreted as the displacement current. The displacement current
appears to increase with time and has negative polarity up to the onset of the return
stroke. Upon initiation of the return stroke, the total displacement current at that point
changes polarity (becomes positive) and attains a peak of a few kA.
83
This is illustrated in Figure 5-2.
Figure 5-2. Displacement current for S9903 RS3 as inferred, using Eq. 5.3, from Imeas and
Hmeas at 30 m.
As seen in Fig. 5.2, the shape of the inferred displacement current waveform
resembles the shape of the electric field derivative (see Fig 3-13). Prior to the return-
stroke, while no conduction current is flowing at ground level, the effect of the
displacement current is apparent in the initial slow front of the measured magnetic field.
After the start of the return stroke, an abrupt increase in the slope of the magnetic field
takes place. Further effects of the displacement current component on the total magnetic
field are the reduction in amplitude of the peak of the magnetic field waveform.
Figures 5.3 – 5.6 show histograms of positive peak displacement currents (during
the return-stroke stage) for the years 1999, 2000, and 2001, estimated from waveforms
84
similar to that shown in Fig. 5-2. Similarly figures 5.7 – 5.10 show histograms of
negative peak displacement currents during the leader stage.
Distance 15 m, 1999
Return-Stroke Peak Displacement Current [kA]0 3 6 9 12 15
Num
ber
0
2
4
6
8
10
12
(a) Mean 6.09 kASt. Dev. 3.29 kAMin 1.10 kAMax 15.0 kAGM 5.25 kASample Size 23
Distance 30 m, 1999
Return-Stroke Peak Displacement current [kA]
0 3 6 9 12 15
Num
ber
0
2
4
6
8
10
(b) Mean 6.41 kASt. Dev. 3.11 kAMin 2.70 kAMax 15.0 kAGM 5.76 kASample Size 23
Figure 5-3. Return-stroke peak displacement current estimated, using Eq. 5.3, from Imeas and Hmeas at (a) 15 m and (b) 30 m in 1999.
85
Distance 15 m, 2000
Return-Stroke Peak Displacement Current [kA]
0 3 6 9 12 15
Num
ber
0
2
4
6
8
10
12
14
16(a)
Mean 3.91 kASt. Dev. 2.15 kAMin 1.40 kAMax 9.80 kAGM 3.43 kASample Size 31
Distance 30 m, 2000
Return-Stroke Peak Displacement Current [kA]
0 3 6 9 12 15
Num
ber
0
2
4
6
8
10
12
14
16
18(b)
Mean 4.93 kASt. Dev. 2.65 kAMin 1.69 kAMax 12.0 kAGM 4.33 kASample Size 31
Figure 5-4. Return-stroke peak displacement current estimated, using Eq. 5.3, from Imeas
and Hmeas at (a) 15 m and (b) 30 m in 2000.
86
Distance 15 m, 2001
Return-Stroke Peak Displacement Current [kA]
0 3 6 9 12 15
Num
ber
0
1
2
3
4
5
(a)Mean 4.07 kASt. Dev. 1.55 kAMin 2.30 kAMax 6.60 kAGM 3.82 kASample Size 8
Distance 30 m, 2001
Return-Stroke Peak Displacement Current [kA]
0 3 6 9 12 15
Num
ber
0
1
2
3
4
5
6
(b)Mean 5.08 kASt. Dev. 1.86 kAMin 2.40 kAMax 7.70 kAGM 4.76 kASample Size 8
Figure 5-5. Return-stoke peak displacement current estimated, using Eq. 5.3, from Imeas
and Hmeas at (a) 15 m and (b) 30 m in 2001.
87
Distance 15 m
Return-Stroke Peak Displacement Current [kA]
0 3 6 9 12 15
Num
ber
0
5
10
15
20
25
30
1999 n = 23 Mean = 6.09 kA St. Dev = 3.29 kA 2000 n = 31 Mean = 3.91 kA St. Dev = 2.15 kA2001 n = 8 Mean = 4.07 kA St. Dev = 1.55 kA
(a) Mean 4.74 kASt. Dev. 2.75 kAMin 1.10 kAMax 15.0 kAGM 4.07 kASample Size 62
Distance 30 m
Return-Stroke Peak Displacement Current [kA]
0 3 6 9 12 15
Num
ber
0
5
10
15
20
25
30
35
1999 n = 23 Mean = 6.41 kA St. Dev. = 3.11 kA2000 n = 31 Mean = 4.93 kA St. Dev. = 2.65 kA2001 n = 8 Mean = 5.08 kA St. Dev. = 1.86 kA
(b) Mean 5.50 kASt. Dev. 2.80 kAMin 1.69 kAMax 15.0 kAGM 4.87 kASample Size 62
Figure 5-6. Return-stroke peak displacement current estimated, using Eq. 5.3, from Imeas
and Hmeas at (a) 15 m and (b) 30 m in 1999, 2000, and 2001.
88
Distance 15 m, 1999
Leader Peak Displacement Current [kA]
0 2 4 6 8 10
Num
ber
0
2
4
6
8
10
12
14(a)
Mean 2.42 kASt. Dev. 1.37 kAMin 0.50 kAMax 6.50 kAGM 2.08 kASample Size 23
Distance 30 m, 1999
Leader Peak Displacment Current [kA]
0 2 4 6 8 10
Num
ber
0
2
4
6
8
10
12(b) Mean 3.07 kA
St. Dev. 1.60 kAMin 0.50 kAMax 7.50 kAGM 2.66 kASample Size 23
Figure 5-7. Leader peak displacement current estimated, using Eq. 5.3, from Imeas and
Hmeas at (a) 15 m and (b) 30 m in 1999.
89
Distance 15 m, 2000
Leader Peak Displacement Current [kA]
0 2 4 6 8 10
Num
ber
0
5
10
15
20(a)
Mean 2.15 kASt. Dev. 1.70 kAMin 0.30 kAMax 9.30 kAGM 1.73 kASample Size 31
Distance 30 m, 2000
Leader Peak Displamement Current [kA]
0 2 4 6 8 10
Num
ber
0
5
10
15
20
25(b)
Mean 1.99 kASt. Dev. 1.26 kAMin 0.50 kAMax 5.50 kAGM 1.70 kASample Size 31
Figure 5-8. Leader peak displacement current estimated, using Eq. 5.3, from Imeas and
Hmeas at (a) 15 m and (b) 30 m in 2000.
90
Distance 15 m, 2001
Leader Peak Displacement Current [kA]
0 2 4 6 8 10
Num
ber
0
1
2
3
4
5
6(a)
Mean 1.92 kASt. Dev. 0.67 kAMin 1.10 kAMax 3.00 kAGM 1.82 kASample Size 8
Distance 30 m, 2001
Leader Peak Displacement Current [kA]
0 2 4 6 8 10
Num
ber
0
1
2
3
4
5(b)
Mean 1.96 kASt. Dev. 0.54 kAMin 1.30 kAMax 3.00 kAGM 1.90 kASample Size 8
Figure 5-9. Leader peak displacement current estimated, using Eq. 5.3, from Imeas and
Hmeas at (a) 15 m and (b) 30 m in 2001.
91
Figure 5-10. Leader peak displacement current estimated, using Eq. 5.3, from Imeas and
Hmeas at (a) 15 m and (b) 30 m in 1999, 2000, and 2001.
92
5.2 Displacement Current Estimates from dE/dt Signatures at 15 and 30 m
Although variation of dE/dt as a function of radial distance from the lightning
channel is needed for estimating displacement current, a rough estimate of this current
can be obtained using dE/dt waveforms measured only at two distances, 15 and 30 m. In
the following, we will use two approximations to the expected distance dependence of
dE/dt, the latter being shown by a dashed line in Fig. 5-11.
In the first approximation, the dE/dt inside the circle of radius 15 m centered at the
channel attachment point is assumed to be constant and approximately equal to that at 15
m, and the dE/dt between 15 and 30 m is assumed to be constant and approximately equal
to that at 30 m. Clearly, this step-wise approximation of dE/dt distance dependence
(illustrated in Fig. 5-11) results in an underestimation of the displacement current.
Figure 5-11. Distance dependences of dE/dt used in evaluating displacement current within 30 m of the lightning channel based on measured dE/dt at 15 and 30 m.
The dE/dt waveform has a characteristic shape which is negative until the onset of
the return stroke. At the start of the return stroke, the polarity is reversed and its
Step-wise approximation
15 m 30 m
dEdt
Expected variation ofdEdt
Distance from the channel
Linear approximation
Measured at 30 mdEdt
Measured at 15 mdEdt
93
magnitude increases rapidly. This characteristic waveshape is seen both at 15 and 30 m.
At 30 m, the magnitude of dE/dt is smaller than at 15 m. Assuming that the dE/dt
waveform is similar at different distances ranging from 15 to 30 m from the lightning
channel, the next step was to determine the value of dE/dt at 30 m corresponding to dE/dt
peak at 15 m. Propagation over a distance of 15 m corresponds to 15/(3·108) = 5·10-8 s =
50 ns. dE/dt waveforms were digitized at 250 MHz. Since the sampling interval is 4 ns at
a rate of 250 MHz, the peak at 30 m will be observed after a delay of about 12 samples
after the occurrence of peak at 15 m. Thus, the dE/dt value at 30 m corresponding to the
peak at 15 m is 12 samples prior to the peak at 30 m. This gives us the dE/dt values at 15
and 30 m at the same instant on time but at different locations. The dE/dt values at 30 m
corresponding to dE/dt at 15 m along with peak dE/dt values at both 15 and 30 m are
given in Table 5-1.
Now, the displacement current based on the step-wise approximation of dE/dt distance
dependence can be estimated as
dI = .oS
E d St
ε ∂∂∫
15 30
0 15
(15 ) (30 )2 2o or r
E m E mrdr rdrt t
ε π ε π= =
∂ ∂= +
∂ ∂∫ ∫
2 2 2(15 ) (30 )2 (15 / 2) 2 (30 / 2 15 / 2)o oE m E m
t tε π ε π∂ ∂
= + −∂ ∂
(5.4)
Note that the displacement current estimated using Eq. 5.4 corresponds to the time of
peak dE/dt at 15 m. Displacement currents estimated in a manner described above are
summarized in Table 5-1, which also contains estimates obtained using peak values of
dE/dt at both 15 and 30 m (that is, without propagation delay taken into account).
The second approximation to the distance dependence of dE/dt, which has been used
here, is a linear approximation (also illustrated in Fig. 5-11),
94
dI = .oS
E d St
ε ∂∂∫
30
0
(15 ) ( 15) 2or
E m m r rdrt
ε π=
∂⎡ ⎤= + −⎢ ⎥∂⎣ ⎦∫
( ) ( )2 3 2(15 )2 (30 / 2) (30 / 3) (15).(30 / 2)oE m m m
tε π ⎡ ∂ ⎤⎛ ⎞= + −⎜ ⎟⎢ ⎥∂⎝ ⎠⎣ ⎦
i i i (5.5)
where
(15 ) (30 )
15 30
E m E mt tm
∂ ∂−
∂ ∂=−
is the slope of the slanted line shown in Fig. 5-11 for
linear approximation. Displacement current values calculated using Eq. 5.5 are given in
the last column of Table 5-1.
Figures 5.12 – 5.13 are the histograms of the displacement currents at the time of
dE/dt peak at 15 m, estimated using Eq. 5.4, for 1999, 2000 and 2001. Figures 5.14 –
5.15 are the histograms of the displacement currents, estimated using Eq. 5.4 and peak
values of dE/dt at both 15 and 30 m, that is, without accounting for propagation delay.
Figures 5.16 – 5.17 are histograms of the displacement currents, estimated using Eq. 5.5
and peak values of dE/dt at both 15 and 30 m.
95
Table 5-1 Displacement currents estimated using Eq. 5.4 (step-wise approximation of dE/dt distance dependence) and Eq. 5.5 (linear approximations of dE/dt distance dependence).
Flash and stroke ID
Peak dE/dt at
15m [kV/m/µs]
Peak dE/dt at
30m [kV/m/µs]
dE/dt at 30 m
corres-ponding
to peak at 15 m
[kV/m/µs]
Id at the time of
peak dE/dt at 15 m,
step-wise approx.
[kA]
Id using peak values of dE/dt at both 15 and
30 m, step-wise approx.
[kA]
Id using peak
values of dE/dt at both 15
and 30 m, linear
approx. [kA]
S9901 RS2 423 145 27 3.15 5.36 8.26S9901 RS3 172 55 14 1.34 2.11 3.33S9901 RS4 280 90 70 3.06 3.44 5.42S9915 RS4 198 74 34 1.88 2.63 3.92S9918 RS1 450 140 83 4.37 5.44 8.67S9918 RS2 235 80 15 1.75 2.97 4.59S9918 RS4 426 118 72 4.01 4.88 8.09S9918 RS6 362 96 47 3.15 4.06 6.84S9932 RS1 418 111 103 4.55 4.70 7.89S9932 RS2 307 80 32 2.52 3.42 5.79S9935 RS1 330 104 12 2.29 4.01 6.37S9935 RS3 300 92 10 2.06 3.60 5.77S9935 RS4 72 23 13 0.69 0.88 1.39S9935 RS5 250 77 3 1.62 3.01 4.81S9935 RS6 330 104 0 2.06 4.01 6.37S0006 RS1 182 73 70 2.45 2.51 3.64S0006 RS4 80 26 23 0.93 0.99 1.55S0006 RS5 50 16 12 0.54 0.61 0.97S0008 RS3 250 78 57 2.63 3.03 4.82S0008 RS4 495 166 82 4.63 6.21 9.64S0008 RS5 347 126 11 2.38 4.53 6.84S0008 RS6 258 82 8 1.76 3.15 4.99S0008 RS7 673 235 201 7.98 8.62 13.2S0012 RS1 545 200 76 4.83 7.16 10.8S0013 RS1 161 48 43 1.81 1.91 3.08S0013 RS3 336 105 54 3.11 4.07 6.48S0013 RS4 520 165 111 5.33 6.35 10.0S0013 RS5 345 107 41 2.93 4.16 6.64S0013 RS6 447 138 76 4.22 5.38 8.60
96
Table 5-1(Contd.) Displacement currents estimated using Eq. 5.4 (step-wise approximation of dE/dt distance dependence) and Eq. 5.5 (linear approximations of dE/dt distance dependence).
Flash and stroke ID
Peak dE/dt at
15m [kV/m/µs]
Peak dE/dt at
30m [kV/m/µs]
dE/dt at 30 m
corres-ponding
to peak at 15 m
[kV/m/µs]
Id at the time of
peak dE/dt at 15 m,
step-wise approx.
[kA]
Id using peak values of dE/dt at both 15 and
30 m, step-wise approx.
[kA]
Id using peak
values of dE/dt at both 15
and 30 m, linear
approx. [kA]
S0015 RS2 264 91 80.4 3.16 3.36 5.16S0015 RS4 456 166 31 3.43 5.96 8.99S0015 RS5 204 58 20 1.65 2.36 3.88S0015 RS6 349 113 76 3.61 4.30 6.76S0016 RS1 520 206 130 5.69 7.12 10.4S0016 RS5 227 95 18 1.76 3.20 4.58S0022 RS1 228 80 73 2.79 2.93 4.47S0022 RS2 406 130 33 3.16 4.98 7.85S0022 RS3 334 112 6 2.20 4.19 6.50S0023 RS1 339 111 11 2.33 4.20 6.58S0023 RS2 559 207 43 4.30 7.38 11.0S0023 RS3 656 216 88 5.75 8.15 12.7S0025 RS1 397 122 91 4.19 4.77 7.64S0025 RS3 229 67 25 1.90 2.69 4.38S0027 RS1 217 64 39 2.09 2.56 4.15S0027 RS2 355 120 23 2.65 4.47 6.92S0027 RS3 413 127 27 3.09 4.96 7.94S0105 RS5 644 241 5 4.12 8.55 12.7S0107 RS1 638 224 141 6.63 8.19 12.5S0123 RS2 560 181 7.5 3.64 6.90 10.8S0123 RS4 519 196 61 4.39 6.92 10.3
97
1999
Displacement Current [kA]
0 1 2 3 4 5 6 7 8
Num
ber
0
1
2
3
4
5(a) Mean 2.57 kA
St. Dev. 1.13 kAMin 0.69 kAMax 4.55 kAGM 2.31 kASample Size 15
2000
Displacement Current [kA]
0 1 2 3 4 5 6 7 8
Num
ber
0
2
4
6
8
10(b) Mean 3.20 kA
St. Dev. 1.59 kAMin 0.54 kAMax 7.98 kAGM 2.82 kASample Size 31
Figure 5-12. Return-stroke displacement current within 30 m of the lightning channel at
the time of dE/dt peak at 15 m estimated using step-wise approximation (with propagation delay taken into account) in (a) 1999 and (b) 2000.
98
2001
Displacement Current [kA]
0 1 2 3 4 5 6 7 8
Num
ber
0
1
2
3
(a) Mean 4.70 kASt. Dev. 1.33 kAMin 3.64 kAMax 6.63 kAGM 4.57 kASample Size 4
1999, 2000, and 2001
Displacement Current [kA]
0 1 2 3 4 5 6 7 8
Num
ber
0
2
4
6
8
10
12
14
1999 n = 15 Mean = 2.57 kA St. Dev = 1.13 kA2000 n = 31 Mean = 3.20 kA St. Dev = 1.59 kA2001 n = 4 Mean = 4.70 kA St. Dev = 1.33 kA
(b) Mean 3.13 kASt. Dev. 1.52 kAMin 0.54 kAMax 7.98 kAGM 2.76 kASample Size 50
Figure 5-13. Return-stroke displacement current within 30 m of the lightning channel at
the time of dE/dt peak at 15 m estimated using step-wise approximation (with propagation delay taken into account) in (a) 2001 and (b) 1999, 2000, and 2001.
99
1999
Displacement Current [kA]
0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5
6(a) Mean 3.63 kA
St. Dev. 1.23 kAMin 0.88 kAMax 5.44 kAGM 3.37 kASample Size 15
2000
Displacement Current [kA]
0 1 2 3 4 5 6 7 8 9
Num
ber
0
2
4
6
8
10
12
(b) Mean 4.40 kASt. Dev. 2.01 kAMin 0.61 kAMax 8.62 kAGM 3.85 kASample Size 31
Figure 5-14. Return-stroke displacement current within 30 m of the lightning channel
estimated using peak values of dE/dt at both 15 and 30 m estimated using step-wise approximation (without propagation delay taken into account) in (a) 1999 and (b) 2000.
100
2001
Displacement Current [kA]
0 1 2 3 4 5 6 7 8 9
Num
ber
0
1
2
3
4
5(a) Mean 7.64 kA
St. Dev. 0.86 kAMin 6.90 kAMax 8.55 kAGM 7.55 kASample Size 4
1999, 2000, and 2001
Displacement Current [kA]0 1 2 3 4 5 6 7 8 9
Num
ber
0
2
4
6
8
10
12
14
16
1999 n = 15 Mean = 3.63 kA St. Dev. = 1.23 kA2000 n = 31 Mean = 4.40 kA St. Dev. = 2.01 kA2001 n = 4 Mean = 7.64 kA St. Dev. = 0.86 kA
(b) Mean 4.43 kASt. Dev. 2.00 kAMin 0.61 kAMax 8.62 kAGM 3.91 kASample Size 50
Figure 5-15. Return-stroke displacement current within 30 m of the lightning channel
estimated using peak values of dE/dt at both 15 and 30 m estimated using step-wise approximation (without taking propagation delay into account) in (a) 2001 and (b) 1999, 2000, and 2001.
101
1999
Displacement Current [kA]
0 2 4 6 8 10 12 14
Num
ber
0
1
2
3
4
5
6(a) Mean 5.83 kA
St. Dev. 2.02 kAMin 1.39 kAMax 8.67 kAGM 5.38 kASample Size 15
2000
Displacement Current [kA]
0 2 4 6 8 10 12 14
Num
ber
0
2
4
6
8
10
12(b) Mean 6.81 kA
St. Dev. 3.05 kAMin 0.97 kAMax 13.2 kAGM 6.00 kASample Size 31
Figure 5-16. Return-stroke displacement current within 30 m of the lightning channel
estimated using peak values of dE/dt at both 15 and 30 m estimated using linear approximation (without propagation delay taken into account) in (a) 1999 and (b) 2000.
102
2001
Displacement Current [kA]
0 2 4 6 8 10 12 14
Num
ber
0
1
2
3
4
5(a) Mean 11.6 kA
St. Dev. 1.21 kAMin 10.3 kAMax 12.7 kAGM 11.5 kASample Size 4
1999, 2000, and 2001
Displacement Current [kA]0 2 4 6 8 10 12 14
Num
ber
0
2
4
6
8
10
12
14
16
1999 n = 15 Mean = 5.83 kA St. Dev. = 2.02 kA2000 n = 31 Mean = 6.81 kA St. Dev. = 3.05 kA2001 n = 4 Mean = 11.6 kA St. Dev. = 1.21 kA
(b) Mean 6.90 kASt. Dev. 3.02 kAMin 0.87 kAMax 13.2 kAGM 6.12 kASample Size 50
Figure 5-17. Return-stroke displacement current within 30 m of the lightning channel
estimated using peak values of dE/dt at both 15 and 30 m estimated using linear approximation (without taking propagation delay into account) in (a) 2001 and (b) 1999, 2000, and 2001.
103
Figure 5-18 shows the scatter plot comparing displacement currents obtained using
Imeas and Hmeas (Eq. 5.3) vs. those obtained using dE/dt at 15 and 30 m at the time of peak
value of dE/dt at 15 m (step-wise approximation).
Displacement current from measured channel base current and associated magnetic field [kA]
0 2 4 6 8 10 12 14
Dis
plac
emen
t cur
rent
from
dE/
dt [k
A]
0
2
4
6
8
10
12
14
199920002001
Id(dE/dt) = 1.4 + 0.37 Id(Imeas and Hmeas) R2 = 0.43 Sample size = 21
Id(dE/dt) [kA] Id(Imeas and Hmeas) [kA] Mean 3.13 4.65 St. Dev. 1.52 2.50 Figure 5-18. Scatter plot showing displacement current estimates from dE/dt measured at
15 and 30 m from the lightning channel (step-wise approximation) vs. those estimated from measured channel base current and associated magnetic field. In the former case, the estimate corresponds to the time of peak value of dE/dt at 15 m.
104
Figure 5-19 is similar to Figure 5-18, but displacement currents on the vertical axis
are estimated using peak values of dE/dt at both 15 and 30 m (step-wise approximation),
that is, without taking into account propagation delay from 15 to 30 m.
Displacement current from measured channel base current and associated magnetic field [kA]
0 2 4 6 8 10 12 14
Dis
plac
emen
t cur
rent
from
dE
/dt [
kA]
0
2
4
6
8
10
12
14
199920002001
Id(dE/dt) = 2.67 + 0.38 Id(Imeas and Hmeas) R2 = 0.31 Sample size = 21
Id(dE/dt) [kA] Id(Imeas and Hmeas) [kA] Mean 4.43 4.65 St. Dev. 2.00 2.50 Figure 5-19. Scatter plot showing displacement current estimates from dE/dt measured at
15 and 30 m from the lightning channel (step-wise approximation) vs. those estimated from measured channel base current and associated magnetic field. In the former case, the estimate is obtained using peak values of dE/dt at both 15 and 30 m.
105
Figure 5-20 is similar to Figure 5-19, but displacement currents on the vertical axis
are estimated using linear approximation (see Fig. 5-11).
Displacement current from measured channel basecurrent and associated magnetic field [kA]
0 2 4 6 8 10 12 14
Dis
plac
emen
t cur
rent
from
dE
/dt [
kA]
0
2
4
6
8
10
12
14
199920002001
Id(dE/dt) = 4.22 + 0.57 Id(Imeas and Hmeas) R2 = 0.32 Sample size = 21
Id(dE/dt) [kA] Id(Imeas and Hmeas) [kA] Mean 6.90 4.65 St. Dev. 3.02 2.50 Figure 5-20. Scatter plot showing displacement current from dE/dt measured at 15 and 30
m from the lightning channel (linear approximation) vs. those estimated from measured channel base current and associated magnetic field. In the former case, the estimate is obtained using peak values of dE/dt at both 15 and 30 m.
106
5.3 Discussion and Summary
The displacement current waveform resembles the electric field derivative
waveform. The displacement current has negative polarity up to the onset of the return-
stroke. Upon initiation of the return-stroke, the total displacement current at that point
changes polarity (becomes positive) and attains a peak of a few kiloamperes.
The return-stroke peak displacement current inferred from channel-base currents,
Imeas, and associated magnetic fields, Hmeas, measured at 15 m in 1999, 2000, and 2001
has a mean value of 4.7 kA and standard deviation of 2.7 kA, while return-stroke peak
displacement current from Imeas and Hmeas at 30 m has a mean value of 5.5 kA and
standard deviation of 2.8 kA. The sample size for both 15 m and 30 m data is 62. The
leader peak displacement current from Imeas and Hmeas at 15 m has a mean of 2.2 kA and
standard deviation of 1.5 kA, whereas for 30-m measurements the mean is 2.4 kA and
standard deviation is 1.4 kA.
The return-stroke stage displacement current within 30 m of the lightning channel
at the time of dE/dt peak at 15 m estimated using step-wise approximation of dE/dt
variation with distance from the lightning channel (after taking propagation delay into
account) is characterized by a mean value of 3.1 kA. The minimum and maximum
values are 0.5 kA and 8.0 kA, respectively. The return-stroke displacement current within
30 m of the lightning channel estimated using peak values of dE/dt at both 15 and 30 m
(i.e. without taking propagation delay into account) and step-wise approximation of dE/dt
variation with distance from the lightning channel has a mean value of 4.4 kA. The
minimum value is 0.6 kA, and the maximum value is 8.6 kA. The return-stroke
displacement current within 30 m of the lightning channel estimated using peak values of
dE/dt at both 15 and 30 m and linear approximation of dE/dt variation with distance from
107
the lightning channel has a mean value of 6.9 kA. The minimum value is 1.0 kA, and the
maximum value is 13.2 kA. Note that the displacement current during the return-stroke
stage flows in the direction opposite to that of the conduction current in the lightning
channel.
Schnetzer et al. (1998) have obtained the displacement current for flash 96-23,
stroke 5, which had a return-stroke peak current directly measured at the channel base of
approximately 20 kA. The return-stroke peak displacement current was 13.5 kA within
50 m of the lightning channel and 0.7 kA within 10 m of the lightning channel. Our
results match reasonably well those obtained by Schnetzer et al. (1998).
The correlation coefficient of 0.66 is obtained between displacement current
estimates obtained from dE/dt measured at 15 and 30 m from the lightning channel (using
step-wise approximation and at the time of peak value of dE/dt at 15 m) and those
obtained from measured channel base currents and associated magnetic fields at 30 m.
108
CHAPTER 6 SUMMARY
From the statistical analysis of 97 records obtained in 1999, 2000, and 2001 the
mean dart-leader current (at the time of its attachment to ground) inferred from leader
magnetic fields measured at 15 m is 1.87 kA and standard deviation is 1.01 kA. The
minimum and maximum dart-leader currents are 0.38 and 5.40 kA, respectively. From
magnetic fields measured at 30 m (103 records obtained in 1997, 1999, 2000, and 2001),
the mean leader current is 2.58 kA and standard deviation is 1.57 kA. The minimum and
maximum dart-leader currents are 0.40 and 8.63 kA, respectively. Typically leader
currents are in the range of a few kiloamperes. Idone and Orville (1985) estimated dart-
leader peak currents for 22 leaders in two rocket-triggered flashes using the relation
between return-stroke peak current IR and return-stroke peak relative light intensity LR in
each of two flashes LR = 1.5 (IR)1.6 and LR = 6.4(IR)1.1 to the dart leader relative light
intensities in those flashes. The mean dart-leader current was 1.8 kA and the range was
0.1 to 6.0 kA. Results of this study are in reasonably good agreement with those of Idone
and Orville (1985).
Return-stroke peak currents estimated from magnetic fields (as the global magnetic
field peak minus the leader contribution) measured at 15 m in 1999, 2000, and 2001 have
a mean value of 13.2 kA and the standard deviation is 6.4 kA. For the individual years the
mean return-stroke current varies from 14.5 kA in 1999 to 12.0 kA in 2000 and 13.8 kA
in 2001. Return stroke peak currents estimated from magnetic fields measured at 30 m in
1997, 1999, 2000, and 2001 have a mean value of 13.0 kA and the standard deviation is
109
6.0 kA. For the individual years the mean return-stroke peak current varies from 9.9 kA
in 1997, to 14.9 kA in 1999, to 11.8 kA in 2000, and to 14.3 kA in 2001. These inferred
return-stroke peak currents are slightly lower than the directly measured return-stroke
peak currents whose mean is 15.3 kA and standard deviation is 7.5 kA. Depasse (1994)
reported an arithmetic mean of 14.3 kA (maximum value of 60 kA, standard deviation of
9 kA) for 305 peak current values directly measured at the Kennedy Space Center (KSC),
Florida and an arithmetic mean of 11 kA (maximum value 49.9 kA, and standard
deviation of 5.6 kA) for 54 values directly measured at Saint-Privat d’Allier, France.
Rakov et al. (1998), Crawford (1998), and Uman et al. (2000), reported an arithmetic
mean of 15.1 kA (sample size = 37, maximum = 44.4 kA, and st. dev. = 9 kA), 12. 8 kA
(sample size = 11, maximum = 22.6 kA, and st. dev. = 5.6 kA), and 14.8 kA (sample size
= 25, maximum = 33.2 kA, and st. dev. = 7 kA) from direct current measurements at
Camp Blanding in 1993, 1997, and 1998, respectively.
The return-stroke peak currents obtained using Ampere’s law for magnestotatics
and (BL+BRS), where BL and BRS are the leader and return-stroke contributions to the total
magnetic field (as done, for example, by Schoene et al. (2003)), measured at 15 m are
about 14 % higher than those obtained from BRS alone (excluding the leader
contribution). The return-stroke peak currents obtained using Ampere’s law for
magnestotatics and (BL+BRS) measured at 30 m are about 20 % higher than those
obtained from BRS alone.
Dart-leader current to return-stroke current ratio obtained from magnetic fields
measured at 15 m has a mean of 0.14 and the correlation coefficient between these
currents is 0.82. Leader to return-stroke current ratio obtained from magnetic fields at 30
110
m has a mean of 0.20 and the correlation coefficient is 0.85. Idone and Orville (1985)
obtained the ratio of dart leader to return-stroke current for 22 events having a mean of
0.17.
The displacement current waveform resembles the electric field derivative
waveform. The displacement current has negative polarity up to the onset of the return-
stroke. Upon initiation of the return-stroke, the total displacement current at that point
changes polarity (becomes positive) and attains a peak of a few kiloamperes.
The return-stroke peak displacement current inferred from channel-base currents,
Imeas, and associated magnetic fields, Hmeas, measured at 15 m in 1999, 2000, and 2001
has a mean value of 4.7 kA and standard deviation of 2.7 kA, while return-stroke peak
displacement current from Imeas and Hmeas at 30 m has a mean value of 5.5 kA and
standard deviation of 2.8 kA. The sample size for both 15 m and 30 m data is 62. The
leader peak displacement current from Imeas and Hmeas at 15 m has a mean of 2.2 kA and
standard deviation of 1.5 kA, whereas for 30-m measurements the mean is 2.4 kA and
standard deviation is 1.4 kA.
The return-stroke stage displacement current within 30 m of the lightning channel
at the time of dE/dt peak at 15 m estimated using step-wise approximation of dE/dt
variation with distance from the lightning channel (after taking propagation delay into
account) is characterized by a mean value of 3.1 kA. The minimum and maximum
values are 0.5 kA and 8.0 kA, respectively. The return-stroke displacement current within
30 m of the lightning channel estimated using peak values of dE/dt at both 15 and 30 m
(i.e. without taking propagation delay into account) and step-wise approximation of dE/dt
variation with distance from the lightning channel has a mean value of 4.4 kA. The
111
minimum value is 0.6 kA, and the maximum value is 8.6 kA. The return-stroke
displacement current within 30 m of the lightning channel estimated using peak values of
dE/dt at both 15 and 30 m and linear approximation of dE/dt variation with distance from
the lightning channel has a mean value of 6.9 kA. The minimum value is 1.0 kA, and the
maximum value is 13.2 kA. Note that the displacement current during the return-stroke
stage flows in the direction opposite to that of the conduction current in the lightning
channel.
Schnetzer et al. (1998) have obtained the displacement current for flash 96-23,
stroke 5, which had a return-stroke peak current directly measured at the channel base of
approximately 20 kA. The return-stroke peak displacement current was 13.5 kA within
50 m of the lightning channel and 0.7 kA within 10 m of the lightning channel. Our
results match reasonably well those obtained by Schnetzer et al. (1998).
The correlation coefficient of 0.66 is obtained between displacement current
estimates obtained from dE/dt measured at 15 and 30 m from the lightning channel (using
step-wise approximation and at the time of peak value of dE/dt at 15 m) and those
obtained from measured channel base currents and associated magnetic fields at 30 m.
112
CHAPTER 7 RECOMMENDATIONS FOR FUTURE RESEARCH
Additional data are needed to verify the findings presented here. In future
measurements, magnetic field records should be synchronized to better than 40-50 ns
with return stroke current records, so that the onset of return stroke current can be
accurately identified in the magnetic field waveform. This will allow one to better
distinguish the leader and return-stroke parts of the magnetic field waveform and hence
obtain more accurate estimates of the dart-leader current. Additional measurements of
dE/dt waveforms are needed and at more distances from the lightning channel. This will
give a better approximation of dE/dt variation with distance and, as a result, more
accurate estimates of the displacement current. Data collected during 1997, 1999, 2000,
and 2001 should be analyzed further to infer additional properties of rocket-triggered
lightning. For example, multiple-station electric and magnetic field measurements can be
used for estimating distributions of charge and current along the lightning channel.
113
APPENDIX DISPLACEMENT CURRENT GRAPHS
In this appendix, the displacement current waveforms inferred from measured
channel-base currents, Imeas, and associated magnetic fields, Hmeas, measured at 15 and 30
m for strokes in lightning flashes triggered in 1999 (23 strokes), 2000 (31 strokes), and
2001 (8 strokes) are shown. The displacement currents were estimated as
2d meas measI rH Iπ= − .
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
LIST OF REFERENCES
Ben Rhouma, A., Auriol, A.P., Eybert-Berard, A., Berlandis, J.-P., and Bador, B.; "Nearby Lightning Electromagnetic Fields," in Proc. of the 11th Int. Zurich Symp. on Electromagn. Compat., Zurich, Switzerland, 1995, pp. 423-428
Byrne, C.J., Few, A.A., and Weber, M.E.; "In situ Balloon Electric Field Measurements," Geophys. Res. Lett., 10:39-42, 1983
Crawford, D.E.; "Multiple-Station Measurements of Triggered Lightning Electric and Magnetic Fields," Master’s thesis, University of Florida; 1998
Crawford, D.E., Rakov, V.A., Uman, M.A., Schnetzer, G.H., Rambo K.J., Stapleton, M.V., and Fisher, R.J.; "The Close Lightning Electromagnetic Environment: Dart-Leader Electric Field Change Versus Distance," J. Geophys. Res., 106, 14,909-14,917, 2001
Depasse, P., "Statistics on Artificially Triggered Lightning," J. Geophys. Res., 99, 18,515 – 18,522, 1994
Idone, V.P., Orville, R.E.; "Correlated peak relative light intensity and peak current in triggered lightning subsequent return strokes," J. Geophys. Res., 90, 6159– 6164, 1985
Jordan, D.M., Rakov, V.A., Besley, W.H., and Uman, M.A.; "Luminosity Characteristics of Dart Leaders and Return Strokes in Natural Lightning," J. Geophys. Res., 102, 22,025-22,032, 1997
Kodali, V.; "Characterization and Analysis of Close Lightning Electromagnetic Fields," Master’s thesis, University of Florida; 2003
Krehbiel, P.R.; The Earth’s Electrical Environment; National Academy Press, Washington DC; 1986
MacGorman, D.R., and Rust, W.D.; The Electrical Nature of Thunderstorms; Oxford University Press, New York; 1998
Rakov, V.A.; "Some Inferences on the Propagation Mechanisms of Dart Leaders and Return Strokes," J. Geophys. Res., 103, 1879-1887, 1995
145
Rakov, V.A., Uman, M.A., Rambo, K.J., Fernandez, M.I., Fisher, R.J., Schnetzer, G.H., Thottappillil, R., Eybert-Berard, A., Berlandis, J.P., Lalande, P., Bonamy, A., Laroche, P., and Bondiou-Clergerie, A.; " New Insights into Lightning Processes Gained from Triggered-Lightning Experiments in Florida and Alabama," J. Geophys. Res., 103, 14,117-14,130, 1998
Rakov, V.A.; "Lightning Discharges Triggered using Rocket-and-Wire Techniques," J. Geophys. Res., 100, 25,711-25,720, 1999
Rakov, V.A., Uman, M.A., Wang, D., Rambo, K.J., Crawford, D.E., and Schnetzer, G.H.; "Lightning Properties from Triggered-Lightning Experiments at Camp Blanding, Florida (1997-1999)," in Proc. of the 25th Int. Conf. on Lightning Protection, Rhodes, Greece, 2000, pp. 54-59
Rakov, V.A.; "Characterization of Lightning Electromagnetic Fields and Their Modeling," in Proc. of the 15th Int. Zurich Symp. on EMC, Zurich, Switzerland, 2001, pp. 545-550
Rakov, V.A, Uman, M.A., Crawford, D.E., Schoene, J., Jerauld, J., Rambo, K.J., Schnetzer, G.H., DeCarlo, B.A., and Miki, M.; "Close Lightning Electromagnetic Environment: Triggered-Lightning Experiments," in Proc. of the 15th Int. Zurich Symp. on EMC, Zurich, Switzerland, 2003, pp. 545-550
Rakov, V.A., and Uman, M.A.; Lightning: Physics and Effects; Cambridge University Press, Cambridge, UK; 2003
Rakov, V.A.; EEL 5490 "Lightning," class notes; University of Florida, Spring Term 2005
Schnetzer, G.H., Fisher, R.J., Rakov, V.A., and Uman, M.A.; "The Magnetic Field Environment of Nearby Lightning," in Proc. of the 24th Int. Conf. on Lightning Protection, Birmingham, United Kingdom, September 14-18, 1998, pp. 346-349
Schoene, J.; "Analysis of Parameters of Rocket-Triggered Lightning Measured During the 1999 and 2000 Camp Blanding Experiment and Modeling of Electric and Magnetic Field Derivatives using the Transmission Line Model," Master’s thesis, University of Florida, 2002
Schoene, J., Uman, M.A., Rakov, V.A., Kodali, V., Rambo, K.J., Schnetzer, G.H.; "Statistical Characteristics of the Electric and Magnetic Fields and Their Time Derivatives 15 m and 30 m from Triggered Lightning," J. Geophys. Res., 108, D6, 4192, 10.1029/2002JD002698, 2003
Simpson, G. and Scrase, F.J.; "The Distribution of Electricity in the Thunderclouds," Proc. R. Soc. London Ser. A, 161, 309-352, 1937
Thottappillil, R., and Rakov, V.A.; "On Different Approaches to Calculating Lightning Electric Fields," J. Geophys. Res., 106, 14,191-14,205, 2001
146
Uman, M.A.; The Lightning Discharge; Academic Press, San Diego, California; 1987
Uman, M.A., Rakov, V.A., Rambo, K.J., Vaught, T.W., Fernandez, M.I., Cordier, D.J, Chandler, R.M., Bernstein, R., and Golden, C.; "Triggered-Lightning Experiments at Camp Blanding, Florida (1993-1995)," Trans. of IEE Japan, Special Issue on Artificial Rocket Triggered Lightning, 117-B, No. 4, 446-452, 1997
Uman, M. A.; The Lightning Discharge; Dover, Mineola, New York; 2001
Uman, M.A., Schoene, J., Rakov, V.A., Rambo, K.J., and Schnetzer, G.H.; "Correlated Time Derivatives of Current, Electric Field Intensity, and Magnetic Flux Density for Triggered Lightning at 15 m," J. Geophys. Res., 107, D13, 4160, 10.1029/2000JD000249, 2002
Wang, D., Rakov, V.A., Uman, M.A., Fernandez, M.I., Rambo, K.J., Schnetzer, G.H., and Fisher, R.J.; "Characterization of the Initial Stage of Negative Rocket-Triggered Lightning," J. Geophys. Res., 104, 4213-4222, 1999
Wang, D., Rakov, V.A., Uman, M.A., Takagi, N., Watanabe, T., Crawford, D.E., Rambo, K.J., Schnetzer, G.H., Fisher, R.J., and Kawasaki, Z.I.; "Attachment Process in Rocket-Triggered Lightning Strokes," J. Geophys. Res., 104, 2141-2150, 1999
147
BIOGRAPHICAL SKETCH
Ashwin Jhavar was born in Akola, India, in 1979. He graduated with a bachelor’s
degree in electrical and electronics engineering from Birla Institute of Technology and
Science, India, in 2002. In 2003, he went to the USA to pursue graduate studies at the
University of Florida.