IX International Symposium on

Lightning Protection
26
th
-30
th
November 2007 Foz do Iguau, Brazil


INDIRECT ESTIMATION OF LIGHTNING CURRENTS FROM REMOTE
ELECTROMAGNETIC FIELD MEASUREMENTS


Farhad Rachidi
Swiss Federal Institute of Technology (EPFL)
Email: Farhad.Rachidi@epfl.ch
EPFL-STI-LRE, EMC Group, Station 11, CH-1015 Lausanne, Switzerland


Abstract - In this lecture, the theoretical basis for the deterministic and statistical estimation of the lightning current from
distant field measurements is presented. The influence of the presence of an elevated strike object on the peak of the lightning
return stroke current determined from remote field measurements is also discussed.


1 INTRODUCTION

The problem of the estimation of lightning return stroke currents from distant electromagnetic field measurements has
received an increased attention due to the widespread use of lightning location systems (LLS). Due to the enormous
amount of data that can be gathered by means of LLS, such systems represent a significant source of experimental data
to be used for the development of standards related to the protection of power and telecommunication systems against
lightning.

Estimates of lightning peak currents from measured lightning electromagnetic fields are obtained by way of empirical
formulas [1, 2], theoretical equations -either deterministic [3] or statistical [4] - relating the electromagnetic field and
the lightning current.

In this lecture, the theoretical basis for the estimation of the lightning current from distant field measurements will be
summarized. The influence of the presence of an elevated strike object on the peak of the lightning return stroke current
determined from remote field measurements will also be discussed. Indeed, instrumented towers have been used as a
means to evaluate the performance characteristics of lightning location systems, namely, their detection efficiency, their
location accuracy and their current peak estimate (e.g. [5-8])
1
.


2 DETERMINISTIC ESTIMATION OF LIGHTNING CURRENT FROM DISTANT
ELECTROMAGNETIC FIELD

Expressions relating far electromagnetic fields and associated return stroke currents at the channel base have been
derived in the literature for various lightning return stroke models [3]. The use of such relations permits the estimation
of channel base currents of return strokes, and the estimation of not-directly-measurable parameters of the models [3].
In what follows, equations relating far electric fields to channel base currents according to various return stroke current
models will be summarized. These equations are derived assuming that the return stroke speed is constant
2
, ground is
perfectly conducting, and the return-stroke front has not reached the top of the channel). The lightning return-stroke
channel is assumed to be a vertical antenna along the vertical z axis.


1
Note that rocket-triggered lightning technique has also been used for this purpose [9].
2
The return stroke speed varies in general along the return stroke channel. However, since we are considering early times during
which current and field peaks occur, the value of the return stroke speed corresponds to the value at the bottom of the channel (below
500 m or so). Most of the measured return stroke speeds found in the literature are averaged over the lowest some hundreds of
meters.


2.1 BG (Bruce and Golde) Model

Spatial-temporal distribution [10]:


vt z t z i
vt z t i t z i
> =
=
' 0 ) , ' (
' ) , 0 ( ) , ' (
(1)

Far electric field channel base current link [11]:

=
t
far
z
o
d c r r E
vt
r c
t i
0
2
) , (
2
) , 0 (
(2)


where r is the horizontal distance between the lightning channel and the observation point, v is the return-stroke speed,
and c is the speed of light.

2.2 TL (Transmission Line) Model

Spatial-temporal distribution [12]:


vt z t z i
vt z v z t i t z i
> =
=
' 0 ) , ' (
' ) ' , 0 ( ) , ' (
(3)

Far electric field channel base current link [13]:


) , (
2
) , 0 (
2
c r t r E
v
r c
t i
far
z
o
+

= (4)

2.3 TCS (Traveling Current Source) Model

Spatial-temporal distribution [14]:


vt z t z i
vt z v z t i t z i
> =
+ =
' 0 ) , ' (
' ) ' , 0 ( ) , ' (
(5)

Far electric field channel base current link [15]:


[ ] ) , 0 ( ) , 0 (
2
1
) , ( t i kt i k
cr
c r t r E
o
far
z

= +
(6)
where ) / 1 ( c v k + = .
Note that the equation (6) is of non-conventional form, due to the simultaneous presence of terms i(0,kt) and i(0,t).
However, a solution in terms of the channel-base current can be obtained [3]

)
1
, (
1
lim 2 ) , 0 (
c
r
j
k
t
n
j
r
far
z
E
j
k n
cr
o
t i +
=

= (7)

2.4 MTLL (Modified Transmission Line, Linear Current Decay with Height) Model

Spatial-temporal distribution [16]:


vt z t z i
vt z v z t i
H
z
t z i
> =
=
' 0 ) , ' (
' ) / ' , 0 ( )
'
1 ( ) , ' (
(8)

where H is the height of the lightning channel.



After mathematical developments similar to those for MTLE model in [17], the following expression relating far
electric field to the channel base current can be derived for MTLL model [4]

= ) / (
) / (
2
) , (
2
c r t i
H
v
dt
c r t di
r c
v
dt
t r dE
o
far
z
(9)

2.5 MTLE (Modified Transmission Line, Exponential Current Decay with Height) Model

Spatial-temporal distribution [17, 18]:


vt z t z i
vt z v z t i z t z i
> =
=
' 0 ) , ' (
' ) / ' , 0 ( ) ' exp( ) , ' (
(10)

Far electric field channel base current link [17]:


c
r
v
H
t
dt
c r t r dE
v
c r t r E
r c
dt
t di
far
z
far
z
o
+ <

+
+

+
=
) / , ( 1 ) / , (
2
) , 0 (
2
(11)

2.6 DU (Diendorfer-Uman) Model

Spatial-temporal distribution [19]:

+ + =
D
v
z
t
c
z
v
z
i c z t i t z i /
'
exp )
' '
, 0 ( ) / ' , 0 ( ) , ' (
(12)

Far electric field channel base current link [17]:


[ ]

=
+
+ + ) , 0 ( ) , 0 ( ) , 0 ( ) , 0 (
2
1 ) / , (
) / , ( t i kt i
dt
d
t i kt i k
cr dt
c r t r dE
c r t r E
D
o
far
z
D
far
z (13)

where
D
is the discharge time constant.

Again, as for the case of the TCS model, the above differential equation is of non-conventional form. Rachidi and
Thottappillil [3] have shown that (13) can be reformulated into an equivalent first-order differential equation, namely,


dt
c
r
j
k
t
r
far
z
dE
D
c
r
j
k
t
n
j
r
far
z
E
j
k n
cr
o
t i
dt
d
D
t i
) , (
)
1
, (
1
lim 2 ) , 0 ( ) , 0 (
+
+ +
=

= + (14)

Rachidi and Thottappillil [3] showed additionally that truncating the infinite series in (7) and (14) at n equal to about
100 yields a calculation error less than 5%.

3 STATISTICAL ESTIMATION OF LIGHTNING CURRENT FROM DISTANT
ELECTROMAGNETIC FIELD

3.1 Theory

It can be seen from the far field-channel base current relations associated with engineering models that all these
equations involve a certain number of parameters, in particular the return stroke speed v, which, in most practical cases,
are unknown. In other words, to infer the lightning current from its associated distant electric or magnetic field, one has
to assume the value for the return stroke speed. This speed changes, however, from one stroke to another and, as a
result, exhibits significant statistical variation (e.g. [20, 21]).



It follows that an error in the estimation of return stroke speed would result in practically the same amount of error in
the inferred channel-base current (see also [16]). In fact, distant electric and magnetic field peaks are determined as
much by return-stroke speeds as they are by return-stroke current peaks.

To cope with this, Rachidi et al. [4] considered the return stroke current peak I, the return stroke speed v, and the distant
electric field peak E as random variables and assumed that the spatial-temporal distribution of the lightning current
along the channel is as predicted by the Transmission Line (TL) model [22]. The TL model has been adopted only for
its simplicity and its ability to reproduce electromagnetic field peaks with reasonable accuracy, as shown in [1, 12, 22].
However, it should be emphasized that the developments presented in the present paper are also applicable to other,
more complex return stroke models (see, for example, [23]).

According to the TL model, the current peak is related to the far-field peak and to the return stroke speed through the
following expressions

I
r c
v
E
o
2
2
=
(15)

v
E
r c I
o
2
2 = (16)

where r is the horizontal distance between the lightning channel and the observation point, and c is the speed of light.
Note the usual minus sign does not appear in equation (15) because we are interested in magnitudes only.
If the probability density functions (PDF) associated with the random variables E and v are known, then it is possible to
obtain the PDF associated with the current I. In general, given the random variables x, y, and z=g(x,y), the PDF of z can
be expressed in terms of PDFs of x and y [24]


dx
z
z x g
y z x g x f dy
z
y z g
y y z g f z f

) , (
) ), , ( , (
) , (
) ), , ( ( ) (
1
1
1
1
(17)

In many practical cases, however, the exact PDFs of the involved random variables are unknown. In those case, it is
still possible to obtain estimates of the expectation and dispersion of z, starting from those of x and y. Indeed,
expanding the function g in a series about the point (
x
,
y
), where
x
,
y
are the mean values of x and y, respectively,
the mean value and variance of z can be estimated as [24]

+


+

+
2
2
2 2
2
2
2
) , ( ) , (
2
) , (
2
1
) , (
y
y x
y x xy
y x
x
y x
y x z
y
g
y x
g
x
g
g
(18)

2
2
2 2
) , ( ) , ( ) , (
2
) , (
y
y x
y x xy
y x y x
x
y x
z
y
g
y
g
x
g
x
g

(19)
where
z y x
, , are the standard deviation of x, y and z, respectively, and
xy
is the correlation coefficient between x
and y, given by

y x
y x xy
xy


=
(20)
Applying equations (18) and (19) to the TL model equation (15), and after straightforward mathematical manipulations,
we obtain [4]

I v I v
2
I v
2
E
2
1
2
1

r c r c
o o
(21)
and

v I vI v I
2 2
2
I
2
v
2
2
v
2
I
2
2
E
2
1 1
2
1
2
1


+

r c r c r c r c
o o o o
(22)
Neglecting any correlation between current peak and return stroke speed (see [25] for a discussion of this issue),
equations (21) and (22) become, respectively [4]




I v
2
E
2
1

r c
o
(23)

2
I
2
v
2
2
v
2
I
2
2
E
2
1
2
1

r c r c
o o
(24)

It is interesting to observe that equation (23) has the same mathematical form as Equation (15), where the values for E,
v, and I are simply replaced by the respective mean values
E
,
v
and
I
. This result gives to some extent a theoretical
justification to the use of LLS to infer statistical parameters of lightning current from measured fields alone. In other
words, although it seems impossible to obtain reasonably accurate estimate of return stroke current from distant field
measurements for a single event without prior knowledge of the return stroke speed, this could be done statistically, in
terms of mean value and standard deviation, using field measurements acquired by LLS, provided that statistical data
for the return stroke speed are available from independent measurements and are not varying much within the area
covered by the LLS.

3.2 Comparison with experimental data

In order to test the validity of the derived equations, we will use simultaneous measurements of return stroke current,
electric field at 5 km, and return stroke speed associated with triggered lightning return strokes and reported by Willett
et al. in [1]. Table 1 summarizes values for return stroke current peak, electric field (essentially radiation) peak at 5 km,
and photographically-measured two-dimensional return stroke speed for 17 events. The corresponding mean values,
standard deviations, and correlation coefficients are given in Table 2.

Table 1: Simultaneous measurements of channel-base current peak, electric field peak at 5 km and return stroke speed associated
with 17 rocket-triggered lightning return strokes (adapted from Willett et al. [1]).

Flash/Stroke I
(kA)
E at 5
km
(V/m)
v
2-D

(10
8
m/s)
8705/1 8.2 76 1.8
8705/3 7.7 64 1.6
8705/5 10.3 80 1.7
8705/6 11.6 84 1.9
8715/9 33 206 1.4
8715/10 15.6 88 1.4
8725/1 20.3 109 1.6
8725/2 16.7 84 1.5
8725/3 43 196 1.7
8725/5 11.7 60 1.4
8726/2 26.9 174 1.2
8726/3 16.4 95 1.4
8726/4 22.4 143 1.4
8728/10 26.9 144 1.4
8728/11 16.1 88 1.4
8732/1 18.0 126 1.5
8732/2 16.8 97 1.6

Let us assume that the channel-base current is unknown. Inserting statistical parameters associated with field and return
stroke speed into equations (21) and (22), the following values can be calculated for the return stroke current
5 . 18
I
= kA
and
0 . 7
I
= kA
which are in very good agreement with values (see Table 2) determined from measurements presented in Table 1.
Similarly, starting from statistical parameters for current and return stroke speed, one can obtain from equations (21)
and (22) the following values for the electric field


4 . 113
E
= V/m
and
5 . 57
E
= V/m
which, again, are in excellent agreement with the values in Table 2 determined from data presented in Table 1.


Table 2: Statistical parameters for the data presented in Table 1 ([4]).

Parameter I
(kA)
v
2-D

(10
8
m/s)
E
(V/m)
Min. value 7.7 1.2 60
Max. value 43 1.9 206
Mean value 18.9 1.5 112.6
Standard
deviation
9.3 0.18 45.1
Iv

0.0144
Ev

0.0268

Note that, since the correlation coefficients
Iv
and
Ev
are very small, the use of equations (23) and (24) instead of
more general equations (21) and (22) does not introduce significant errors. This is an important observation since it
suggests that no knowledge (usually unavailable) of the correlation coefficients is necessary [4].


4 PRESENCE OF A TALL STRIKE OBJECT

4.1 Return-Stroke Models Including the Presence of an Tall Strike Object

To analyze the interaction of lightning with tall strike objects, some of the engineering return stroke models, initially
developed for the case of return strokes initiated at ground, were extended to take into account the presence of a
vertically-extended strike object e.g., [26-42]. In some of these models, it is assumed that a current pulse i
o
(t) associated
with the return-stroke process is injected at the lightning attachment point both into the strike object and into the
lightning channel, e.g., [27, 28, 31, 33-38, 43]. The upward-moving wave propagates along the channel at the return-
stroke speed v as specified by the return-stroke model. The downward-moving wave propagates at the speed of light
along the strike object, assumed to be a lossless uniform transmission line characterized by constant non-zero reflection
coefficients at its top and its bottom. As noted in [29], the assumption of two identical current waves injected into the
lightning channel and into the strike object implies that their characteristic impedances are equal to each other. This
assumption makes the models not self-consistent in that (1) there is no impedance discontinuity at the tower top at the
time of lightning attachment to the tower, but (2) there is one when the reflections from ground arrive at the tower top.

Extension of engineering models based on a distributed-source representation
Rachidi et al. [44] presented an extension of the so-called engineering return stroke models, taking into account the
presence of a vertically-extended strike object, which does not employ the assumption that identical current pulses are
launched both upward and downward from the object top. The extension is based on a distributed-source representation
of the return-stroke channel [17, 45], which allows more general and straightforward formulations of these models than
the traditional representations implying a lumped current source at the bottom of the channel.

The general equations for the spatial-temporal distribution of the current along the lightning channel and along the
strike object have been derived [44]:

( ) ( )
( )( )

+
+ +

=
+
v
h z
t u
c
nh
c
z h
t h i
c
h z
t h i
v
h z
t h i h z P t z i
o
n
n n
t t
o t o
t g
' 2 '
, 1 1
'
,
*
'
, ' , '
0
1
for
0
' H z h < < (25)



( )

=
c
nh
c
z h
c
nh
c
z h
t
c
nh
c
z h
t
t u h i
h i t z i
o
n
g
n
t
n
o
n
g
n
t t
2 ' 2 '
2 '
,
, 1 ) , ' (
1
0
for h z ' 0 (26)

In (25) and (26),
- h is the height of the tower,
-
t
and
g
are the top and bottom current reflection coefficients for upward and downward propagating waves,
respectively, given by

ch t
ch t
t
Z Z
Z Z
+

=
(27)

g t
g t
g
Z Z
Z Z
+

=
(28)

- H
0
is the height of the extending return stroke channel,
- c is the speed of light,
- P(z) is a model-dependent attenuation function,
- u(t) the Heaviside unit-step function,
- v is the return-stroke front speed, and
- v* is the current-wave speed.

Expressions for P(z) and v* for some of the most commonly used return-stroke models are summarized in Table 3, in
which is the attenuation height for the MTLE model and H
tot
is the total height of the lightning channel.
Equations (1) and (2) are based on the concept of undisturbed current i
o
(t), which represents the ideal current that
would be measured at the tower top if the current reflection coefficients at its both extremities were equal to zero.
It is assumed that the current reflection coefficients
t
and
g
are constant. In addition, any upward connecting leader
and any reflections at the return stroke wavefront [36] are disregarded.

Table 3: P(z) and v* for different return-stroke models (Adapted from [23]).

Model P(z) v*
BG 1
TCS 1 -c
TL 1 v
MTLL 1-z/H
tot
v
MTLE exp(-z/) v


Extension of engineering models based on a lumped series voltage source
Baba and Rakov [41, 46] proposed an alternative approach to Rachidi et al.s distributed source representation [44],
using a lumped series voltage source at the junction point between the channel and the strike object. They showed that
such a representation assures appropriate boundary conditions at the attachment point and is equivalent to the
distributed source representation [46]. In their representation, Baba and Rakov expressed the spatial-temporal
distribution of the current along the strike object and along the channel in terms of the short-circuit current i
sc
(t), which
is related to the undisturbed current through
3



) ( 2 ) ( t i t i
o sc
=

(29)


3
An equivalent representation in terms of the so-called reference current the current that would flow through the return-stroke
channel in the absence of the elevated struck object is also possible (see Shigihara and Piantini [47]).


Furthermore, in [46], Baba and Rakov considered in their expressions a different speed v
ref
for the upward propagating
current waves reflected from the ground and then transmitted into the lightning channel.

4.2 On the representation of the elevated strike object

In all engineering models, the elevated strike object is modeled as an ideal transmission line. To include the structural
nonuniformities of the elevated strike object, several transmission line sections in cascade have also been considered
(e.g. [30, 48]). The transmission line representation of the elevated strike object has been shown to yield reasonable
results in comparison with experimental data. However, one should bear in mind that experimental data associated with
lightning to tall structures are affected by other, less-easily controlled factors such as the variability of lightning
channel impedance and possible reflections at the return stroke wavefront [49]. In [50], Bermudez et al. presented an
experimental validation of the transmission line representation of an elevated object struck by lightning. The
experimental results were obtained using a reduced-scale model and injected signals with narrow pulse widths (down to
500 ps). The validation is performed using a reduced scale structure representing the Toronto CN Tower in Canada.
Two models consisting, respectively, of 1-section and 3-section uniform transmission lines were considered for the
comparison. It was shown that the 3 section model is able to accurately reproduce the obtained experimental data. The
overall agreement between the 1-section model and the experimental results was also satisfactory, at least for the early-
time response.

More recently, FDTD simulations performed by Baba and Rakov [51] suggest that the waveguide properties of a
biconical antenna (representing a tower) depend on the direction of propagation. Precisely, while the current pulses
suffer no attenuation while traveling from the tower apex to its base, the attenuation is significant when pulses
propagate from the base to the apex [51]. This finding might render questionable the validity of reflection coefficients
at ground level inferred from the measurements of current at the top of the tower.

4.3 Far-field Peak vs. current Peak relationship

For the case when the round-trip propagation time of the tall structure is greater than the zero-to-peak risetime t
f
, the
relation between far electric and magnetic field peaks and the associated undisturbed current peak at the top of the
elevated object is given by [40]
4


peak o
I
t
v
c
r c
o
v
far
peak z
E

= ) 2 1 ( 1
2
2
(30)


peak o
I
t
v
c
cr
v
far
peak
H

) 2 1 ( 1
2
(31)
where
peak o
I is the first peak of the undisturbed current i
o
(h,t).
Note that, because of the condition t
f
< 2h/c imposed on the current, the above are independent of the structures height
h and of the ground reflection coefficient
g
[40].

For the case of a strike to ground when the reflections at ground level are taken into account, the relationships between
far field peaks and the associated undisturbed current peak become [40]


peak o
I
g ch
v
c
r c
o
v
far
peak z
E

= 1
2
2
(32)

peak o
I
g ch
v
c
cr
v
far
peak
H

1
2
(33)

in which
( ) ( )
g ch g ch g ch
Z Z Z Z + =

is the current reflection coefficient at ground level .
Comparing (30)-(31) with (32)-(33), it can defined the far-field enhancement factor due to the presence of a tall strike
object as follows [52]

4
Note that Bermudez et al. [40] disregarded in their derivation the effect of the discontinuity at the return stroke wavefront.



( )
g ch
t
tall
v
c
v
c
k

+
+
=
1
2 1 1
(34)
Baba and Rakov [41] have derived a similar expression for the far field enhancement factor using their model
5
and they
have presented a thorough analysis of the distance dependences of electric and magnetic fields due to a lightning strike
to a tall object and due to the same lightning strike to flat ground.
In what follows, it is expressed the relationship between the far electromagnetic field peak and the peak current that
would be measured at the top of a tall strike object [40] [52]


peak
I
tall
k
r c
o
v
far
peak z
E
'
2
2
= (35)

peak
I
tall
k
cr
v
far
peak
H
'
2
=

(36)
where
'
tall
k is defined as

( )
t
t
tall
v c
k

+
=
1
2 1 1
'
(37)
and
peak
I is the peak amplitude of the current that would be measured at the top of the elevated strike object.

A short discussion on the terminology is in order [52]. Bermudez et al. [40] used the notation
tall
k instead of
'
tall
k and
called this factor the tower enhancement factor. In the present study, the notation and definition of Baba and Rakov
[41] are followed, in which the term 'tower enhancement factor' is attributed to the ratio of the far field associated with
a strike to a tall tower to the far field associated with a similar strike to ground, expressed by
tall
k , and given by (34)
within the model adopted in this study.

Figure 1 presents the variation of the factor
'
tall
k as a function of the reflection coefficient at the tower top. The
adopted value for the return-stroke speed is v = 150 m/s. It can be seen that a variation of
t
from 0 to -1 results in a
variation of
'
tall
k in the range 3 to 3.5.
Figure 2 presents the variation of the factor
'
tall
k as a function of the return-stroke speed. This time, the value of
t
has
been chosen to be equal to -0.4. Note that, in Figure 2, the factor
'
tall
k is somewhat more sensitive to the return-stroke
speed. However, it is important to realize that, in the expression relating the far electric field peak and the current peak
(see (36)), the return-stroke speed v appears not only within
'
tall
k but also as a separate proportionality factor. The
overall effect of the return-stroke speed on the far-field current relation is presented in Fig. 3. In this figure, two
typical values for the return-stroke speed have been considered, namely, 100 m/s and 200 m/s. It can be seen that the
peak E-field and H-field radiated by a lightning return stroke to a tall structure is little sensitive to the return-stroke
speed.


5
The expression for the far field enhancement factor in the model of Baba and Rakov is slightly different from (34). As shown in
[41] the difference is due to the fact the speed of current waves propagating along the lightning channel is assumed to be the speed
of light in the present study (which is based on the Rachidi et al. model [44], whereas in the model by Baba and Rakov, the current
waves along the channel are assumed to travel at the return stroke speed.




1 0.8 0.6 0.4 0.2 0
0
1
2
3
4
Top reflection coefficient
F
a
c
t
o
r

k
'
t
a
l
l

Fig. 1 - Variation of
'
tall
k for a tall tower as a function of the top reflection coefficient
t
. The return-stroke speed is assumed to be
v = 1.510
8
m/s. Adapted from [52]




1
.
10
8
1.2
.
10
8
1.4
.
10
8
1.6
.
10
8
1.8
.
10
8
2
.
10
8
0
1
2
3
4
5
Return-stroke speed (m/s)
F
a
c
t
o
r

k
'
t
a
l
l

Fig. 2 - Variation of
'
tall
k for a tall tower as a function of the return-stroke speed.
The top reflection coefficient
t
is assumed to be equal to -0.4. Adapted from [52]




0 5 10 15 20
0
10
20
30
Current Peak (kA)
E
-
F
i
e
l
d

P
e
a
k

(
V
/
m
)
0 5 10 15 20
0
0.02
0.04
0.06
Current Peak (kA)
H
-
F
i
e
l
d

P
e
a
k

(
A
/
m
)

(a) (b)
Fig. 3 Electric field peak (a) and Magnetic field peak (b) as a function of return-stroke current peak.
The observation point is located at a distance of 100 km from the strike location.
The top reflection coefficient
t
is assumed to be equal to -0.4
(Solid line: return stroke speed: v = 210
8
m/s. Dashed line: v = 10
8
m/s). Adapted from [52]

4.4 Comparison between directly-measured currents and estimates from LLS for strikes to tall towers [53]

Recently, lightning research groups were interested in evaluating the performance characteristics of lightning location
systems, namely, their detection efficiency, their location accuracy and their current peak estimate by means of
ground-truth measurements using either instrumented towers (e.g. [6, 7]) or rocket-triggered lightning [9].

Diendorfer et al. [6] compared lightning peak currents measured at the Gaisberg tower (100-m tall) with correlated
lightning peak currents reported by the Austrian lightning location system ALDIS, finding a surprisingly good
agreement, quantified in terms of a current peak estimate from ALDIS equal to 95 % the value recorded at the tower.

Jerauld et al [9] evaluated the performance characteristics of the US NLDN using rocket-triggered lightning data
acquired during the summers 2001-2003 at Camp Blanding, Florida, reporting a tendency of NLDN to underestimate
peak currents, with a median peak current estimation error of about 18 %.

More recently, the directly-measured lightning currents near the top of the CN Tower during the Summer 2005 have
been correlated and compared with the US NLDN current peak estimates [53]. Unlike the Gaisberg tower, whose
height above ground is only 100 m and hence can be considered as an electrically-short tower for typical lightning
return-stroke currents, the CN Tower with its 553 m tallness is tall enough so that the condition current risetime t
r
< h/c
is often satisfied. Therefore, it is expected that a non-negligible enhancement effect takes place for strikes to the CN
Tower
6
. In fact, a lightning detection network, composed by a certain number of field measuring stations, is not able to
distinguish between a strike to ground and a strike to a tall object. As a result, the inferred current peak for a tall tower-
initiated event might be overestimated because of the tower factor (expressed in terms of the coefficient
'
tall
k in Eq.
(37)).

This discrepancy was also observed in the past by Montandon [57] between the return-stroke current estimations
provided by the Lightning Positioning And Tracking System (LPATS) installed in Switzerland in 1990 by Swiss PTT
and the lightning currents actually recorded using two different instrumented towers: the 168-m-tall Peissenberg Tower
in Germany and the 250-m-tall St. Chrischona Tower in Switzerland, reporting an overestimation of the current peaks
by the LPATS system of 15 to 20 dB.

In Fig 4, the current peaks inferred for each return-stroke detected by US NLDN are shown as a function of the
corresponding current peaks measured at the CN Tower. The best-fit straight line of the points shows that the current

6
Note that at very close distance ranges, the effect of the tower could result in a significant decrease of the electric field [41,54,55],
and sometimes a change in its polarity [56].


estimation provided by the lightning detection network for those tower strokes is about 3.5 times higher than the actual
measurement.
0
10
20
30
40
50
0 5 10 15 20
N
L
D
N

E
s
t
i
m
a
t
e

(
k
A
)
Measured First Peak (kA)
y =x
y =3.5196x

Fig. 4 - Comparison of peak current measurements at the CN Tower with peak currents of correlated events reported by NLDN. The
solid line represents the linear curve fit. The dashed line represents the ideal relation between the directly measured currents and
NLDN estimates (NLDN data courtesy of K. Cummins, Vaisala). Adapted from [53]

This result agrees very well with the value obtained for the factor
'
tall
k discussed above. Indeed, assuming that NLDN
uses a classical TL relation [58] to infer the current peak, the presence of a tall strike object would result in an
overestimation of the inferred value by a factor
'
tall
k . For the adopted values (
t
= -0.366 and v = 120 m/s), this factor
is equal to 3.9. Consequently, the current-peak estimation of the detected CN Tower strikes has been corrected in Fig 5
plotting the current peaks inferred by the US NLDN divided by the factor
'
tall
k = 3.9, as a function of the
corresponding current peaks measured at the CN Tower. The best-fit straight line for the corrected points shows now a
very satisfactory estimation of the current peaks from the lightning detection network.
0
2
4
6
8
10
12
14
16
0 2 4 6 8 10 12 14 16
C
o
r
r
e
c
t
e
d

N
L
D
N

E
s
t
i
m
a
t
e

(
k
A
)
Measured First Peak (kA)
y =0.90199x
y =x

Fig. 5 - Comparison of peak current measurements at the CN Tower with peak currents of correlated events reported by NLDN, after
correction using the tower factor
'
tall
k
. The solid line represents the linear curve fit.
The dashed line represents the ideal relation between the directly measured currents and corrected NLDN estimates (NLDN data
courtesy of K. Cummins, Vaisala). Adapted from [53]




Note, finally, that this excellent agreement is not due to a judicious choice of the parameters
t
and v. Indeed, as it has
been shown in the previous section, the far field peaks associated with strikes to tall strike objects are very little
sensitive to
t
and v.

5 CONCLUSION

In this paper, the theoretical basis for the estimation of the lightning current from distant field measurements was
presented and discussed. First, analytical expressions relating the lightning return stroke channel base current and the
far electric field have been summarized for the engineering models. The use of such relations permits the determination
of nondirectly measurable parameters of the models, the estimation of channel base currents of return strokes, and
further testing of these models using simultaneous fields and current measurements.

There is, however, an inherent difficulty in extracting lightning current parameters from LLS-measured electromagnetic
fields since unknown parameters and in particular the return stroke speed affect the lightning current inferred from
remote electromagnetic fields. Although, due to the high variability of key parameters such as the return stroke speed, it
is impossible to determine the lightning current accurately from the remotely measured electric or magnetic field for a
given event, the statistical estimation (e.g. in terms of mean values and standard deviations) is possible. For the
Transmission Line (TL) model, the equation permitting to infer the mean value of the return stroke current from the
mean value of electric or magnetic field and the mean value of speed has the same functional form as the well-known
TL currentfar field relationship. This result gives to some extent a theoretical justification to the use of lightning
location systems (LLS) to infer parameters of lightning current statistical distributions.

Finally, the directly-measured lightning currents at the CN Tower have been correlated and compared with US NLDN
current peak estimates. It has been shown that the NLDN inferred values overestimate actual current peaks, because the
presence of the tall struck object is not included in the algorithm used to infer lightning current peaks from remote field
measurements. However, correcting the NLDN estimates using the tower factor results in an excellent estimation of
lightning current peaks [53].

Acknowledgments Most of the results reported in this lecture are obtained within the framework of an international collaboration
involving Swiss Federal Institute of Technology (Lausanne, Switzerland), Universities of Bologna and Rome (Italy), University of
Florida (USA), Universities of Toronto and Mc Master (Canada) and others. The author expresses his sincere gratitude to C.A.
Nucci, D. Pavanello, V.A. Rakov and M. Rubinstein for their valuable contributions, comments and suggestions. The author is also
grateful to Y. Baba, J.L. Bermudez, A. Borghetti, J.S. Chang, W.A. Chisholm, V. Cooray, K. Cummins, G. Diendorfer, Z. Flisowski,
F. Heidler, S. Guerrieri, A.M. Hussein, M. Ianoz, W. Janischewskyj, E.P. Krider, C. Mazzetti, R. Moini, E. Montandon, A.
Mosaddeghi, M. Paolone, E. Petrache, A. Piantini, T. Shindo, V. Shostak, W. Schulz, F.M. Tesche, N. Theethayi, R. Thottappillil,
M.A. Uman, S. Visacro and S. Yokoyama for interesting and fruitful discussions on the subject over the last two decades or so.
This work has been has been financially supported by the Swiss State Secretariat for Education and Research (Grant No C05.0149)
and by the European COST Action P18 The Physics of Lightning Flash and Its Effects.

6 REFERENCES

[1] J. C. Willett, J. C. Bailley, V. P. Idone, A. Eybert-Berard, and L. Barret, "Submicrosecond Intercomparison of Radiation
Fields and currents in triggered Lightning Return Strokes Based on the Transmission-Line Model," Journal of Geophysical
Research, vol. 94, pp. 13,275 - 13,286, 1989.
[2] V. A. Rakov, R. Thottappillil, and M. A. Uman, "On the empirical formula of Willett et al. relating lightning return-stroke
peak current and peak electric field," Journal of Geophysical Research, vol. 97, pp. 11527-33, 1992.
[3] F. Rachidi and R. Thottappillil, "Determination of lightning currents from far electromagnetic fields," Journal of Geophysical
Research, vol. 98, pp. 18315-20, 1993.
[4] F. Rachidi, J. L. Bermudez, M. Rubinstein, and V. A. Rakov, "On the estimation of lightning peak currents from measured
fields using lightning location systems," Journal of Electrostatics, vol. 60, pp. 121-129, 2004.
[5] W. Janischewskyj, V. Shostak, A. M. Hussein, and W. A. Chisholm, "Estimation of lightning location system accuracy using
CN Tower lightning data," presented at International Conference on Lightning Protection (ICLP96), Florence, Italy, 1996.
[6] G. Diendorfer, W. Hadrian, F. Hofbauer, M. Mair, and W. Schulz, "Evaluation of lightning location data employing
measurements of direct strikes to a radio tower," Session 2002 CIGRE, Paris, France 2002.
[7] W. Schulz and G. Diendorfer, "Lightning peak currents measured on tall towers and measured with lightning location
systems," presented at 18th International Lightning Detection Conference ILDC 2004, Helsinki, Finland, 2004.
[8] A. Lafkovici, A. M. Hussein, W. Janischewskyj, and K. Cummins, "Performance analysis of the North American Lightning
Detection Network using CN Tower Lightning data," presented at International Lightning Detection Conference, Tucson,
Arizona, 2006.


[9] J. Jerauld, V. A. Rakov, M. A. Uman, K. J. Rambo, and D. M. Jordan, "An evaluation of the performance characteristics of the
U.S. National Lightning Detection Network in Florida using rocket-triggered lightning," Journal of Geophysical Research, vol.
110, 2005.
[10] C. E. R. Bruce and R. H. Golde, "The lightning discharge," The journal of the institution of electrical engineers, vol. 88, pp.
487-520, 1941.
[11] M. A. Uman and D. K. McLain, "Lightning return stroke current from magnetic and radiation field measurements," Journal of
Geophysical Research, vol. 75, pp. 5143-7, 1970.
[12] M. A. Uman and D. K. McLain, "Magnetic field of lightning return stroke," Journal of Geophysical Research, vol. 74, pp.
6899-910, 1969.
[13] M. A. Uman, D. K. McLain, and E. P. Krider, "The electromagnetic radiation from a finite antenna," American Journal of
Physics, vol. 43, pp. 33-8, 1975.
[14] F. Heidler, "Traveling current source model for LEMP calculation," Electromagnetic Compatibility, vol. 626, 1985.
[15] F. Heidler, "Lightning electromagnetic pulse, Theorie und messungen," in Fak. der Elektrotech. Munich: Univ. der
Bundeswehr, 1987.
[16] V. A. Rakov and A. A. Dulzon, "Results of calculation of the electromagnetic fields of lightning discharges," Tekhnicheskaya
Elektrodinamika. no., vol. 1, pp. 87-9, 1987.
[17] F. Rachidi and C. A. Nucci, "On the Master, Uman, Lin, Standler and the Modified Transmission Line lightning return stroke
current models," Journal of Geophysical Research, vol. 95, pp. 20389-94, 1990.
[18] C. A. Nucci, C. Mazzetti, F. Rachidi, and M. Ianoz, "On lightning return stroke models for LEMP calculations," presented at
19th international conference on lightning protection, Graz, 1988.
[19] G. Diendorfer and M. A. Uman, "An improved return stroke model with specified channel-base current," Journal of
Geophysical Research, vol. 95, pp. 13621-44, 1990.
[20] V. P. Idone and R. E. Orville, "Lightning return stroke velocities in the Thunderstorm Research International Program (TRIP),"
Journal of Geophysical Research, vol. 87, pp. 4903-15, 1982.
[21] D. M. Mach and W. D. Rust, "photoelectric Return-stroke velocity and peak current estimates in natural and triggered
lightning," Journal of geophysical research, vol. 94, pp. 13,237-13,247, 1989.
[22] R. Thottappillil and M. A. Uman, "Comparison of lightning return-stroke models," Journal of Geophysical Research, vol. 98,
pp. 22903-14, 1993.
[23] V. A. Rakov and M. A. Uman, "Review and evaluation of lightning return stroke models including some aspects of their
application," IEEE Transactions on Electromagnetic Compatibility, vol. 40, pp. 403-26, 1998.
[24] A. Papoulis, Probability, random variables, and stochastic processes, Third ed: Mc Graw Hill, 1991.
[25] V. A. Rakov and M. A. Uman, Lightning: physics and effects: Cambridge University Press, 2003.
[26] T. Zundl, "Lightning current and LEMP calculations compared to measurements gained at the Peissenberg tower," presented at
22nd ICLP (International Conference on Lightning Protection), Budapest, Hungary, 1994.
[27] S. Guerrieri, F. Heidler, C. A. Nucci, F. Rachidi, and M. Rubinstein, "Extension of two return stroke models to consider the
influence of elevated strike objects on the lightning return stroke current and the radiated electromagnetic field: comparison
with experimental results," EMC '96 ROMA. International Symposium on Electromagnetic Compatibility. Univ. Rome `La
Sapienza', Rome, Italy, vol. 2, 1996.
[28] S. Guerrieri, C. A. Nucci, F. Rachidi, and M. Rubinstein, "On the influence of elevated strike objects on directly measured and
indirectly estimated lightning currents," IEEE Transactions on Power Delivery, vol. 13, pp. 1543-55, 1998.
[29] S. Guerrieri, E. P. Krider, and C. A. Nucci, "Effects of Traveling-Waves of Current on the Initial Response of a Tall Franklin
Rod," presented at ICLP2000, Rhode, Greece, 2000.
[30] I. Rusan, W. Janischewskyj, A. M. Hussein, and J.-S. Chang, "Comparison of measured and computed electromagnetic fields
radiated from lightning strikes to the Toronto CN tower," presented at 23rd International Conference on Lightning Protection
(ICLP), Florence, 1996.
[31] H. Motoyama, W. Janischewskyj, A. M. Hussein, R. Rusan, W. A. Chisholm, and J. S. Chang, "Electromagnetic field radiation
model for lightning strokes to tall structures," IEEE Transactions on Power Delivery, vol. 11, pp. 1624-32, 1996.
[32] V. A. Rakov, "Transient response of a tall object to lightning," IEEE Transactions on Electromagnetic Compatibility, vol. 43,
pp. 654-61, 2001.
[33] F. Rachidi, W. Janischewskyj, A. M. Hussein, C. A. Nucci, S. Guerrieri, B. Kordi, and J. S. Chang, "Current and
electromagnetic field associated with lightning return strokes to tall towers," IEEE Trans. on Electromagnetic Compatibility,
vol. 43, 2001.
[34] W. Janischewskyj, V. Shostak, and A. M. Hussein, "Comparison of lightning electromagnetic field characteristics of first and
subsequent return strokes to a tall tower 1. Magnetic field," presented at 24th ICLP (international conference on lightning
Protection), Birmingham, U.K., 1998.
[35] W. Janischewskyj, V. Shostak, and A. M. Hussein, "Lightning electric field characteristics of first and subsequent return
strokes to a tall tower," Eleventh International Symposium on High Voltage Engineering, vol. 467, 1999.
[36] V. Shostak, W. Janischewskyj, A. Hussein, and B. Kordi, "Electromagnetic fields of lightning strikes to a tall tower: a model
that accounts for upward-connecting discharges," presented at 25th ICLP (International Conference on Lightning Protection),
Rhodes, Greece, 2000.
[37] V. Shostak, W. Janischewskyj, A. M. Hussein, J. S. Chang, and B. Kordi, "Return-stroke current modeling of lightning striking
a tall tower accounting for reflections within the growing channel and for upward-connecting discharges," presented at 11th
International Conference on Atmospheric Electricity, Guntersville, U.S.A., 1999.


[38] H. Goshima, A. Asakawa, T. Shindo, H. Motoyama, A. Wada, and S. Yokoyama, "Characteristics of electromagnetic fields
due to winter lightning stroke current to a high stack," Transactions of the Institute of Electrical Engineers of Japan, Part B,
vol. 120, pp. 44-9, 2000.
[39] J. L. Bermudez, M. Rubinstein, F. Rachidi, F. Heidler, and M. Paolone, "Determination of Reflection Coefficients at the Top
and Bottom of Elevated Strike Objects Struck by Lightning," Journal of Geophysical Research, vol. 108, pp. 4413, doi:
10.1029/2002JD002973, 2003.
[40] J. L. Bermudez, F. Rachidi, W. Janischewskyj, V. Shostak, M. Rubinstein, D. Pavanello, A. M. Hussein, J. S. Chang, C. A.
Nucci, and M. Paolone, "Far-field - current relationship based on the TL model for lightning return strokes to elevated strike
objects," IEEE Transactions on Electromagnetic Compatibility, vol. 47, pp. 146-159, 2005.
[41] Y. Baba and V. A. Rakov, "Lightning electromagnetic environment in the presence of a tall grounded strike object," Journal of
Geophysical Research, vol. 110, 2005.
[42] F. Rachidi, "Modeling Lightning Return Strokes to Tall Structures: A Review," Journal of Lightning Research, vol. 1, pp. 16-
31, 2007.
[43] R. Rusan, W. Janischewskyj, A. M. Hussein, and J. S. Chang, "Comparison of measured and computed electromagnetic fields
radiated from lightning strikes to the Toronto CN tower," presented at 23rd ICLP (International Conference on Lightning
Protection), Florence, Italy, 1996.
[44] F. Rachidi, V. A. Rakov, C. A. Nucci, and J. L. Bermudez, "The Effect of Vertically-Extended Strike Object on the
Distribution of Current Along the Lightning Channel," Journal of Geophysical Research, vol. 107, pp. 4699, 2002.
[45] G. V. Cooray, "On the concepts used in return stroke models applied in engineering practice," IEEE Trans. on Electromagnetic
Compatibility, vol. 45, pp. 101-108, 2002.
[46] Y. Baba and V. A. Rakov, "On the use of lumped sources in lightning return stroke models," Journal of Geophysical Research,
vol. 110, 2005.
[47] M. Shigihara and A. Piantini, "Estimation of lightning currents from measurements performed on elevated objects," presented
at 28th International Conference on Lightning Protection (ICLP), Kanazawa, Japan, 2006.
[48] V. Shostak, W. Janischewskyj, A. Hussein, J. S. Chang, F. Rachidi, and J. L. Bermudez, "Modeling of the electromagnetic
field associated with lightning return strokes to a complex tall tower," presented at 26th ICLP (International Conference on
Lightning Protection), Cracow, Poland, 2002.
[49] V. Shostak, W. Janischewskyj, and A. M. Hussein, "Expanding the modified transmission line model to account for reflections
within the continuously growing lightning return stroke channel," presented at IEEE Power Engineering Society Summer
Meeting, Piscataway, USA, 2000.
[50] J. L. Bermudez, F. Rachidi, W. A. Chisholm, M. Rubinstein, W. Janischewskyj, A. M. Hussein, V. Shostak, and J. S. Chang,
"On the use of transmission line theory to represent a nonuniform vertically-extended object struck by lightning," presented at
2003 IEEE Symposium on Electromagnetic Compatibility (EMC), Boston, USA, 2003.
[51] Y. Baba and V. A. Rakov, "On the interpretation of ground reflections observed in small-scale experiments simulating
lightning strikes to towers," IEEE Transactions on Electromagnetic Compatibility, vol. 47, 2005.
[52] D. Pavanello, F. Rachidi, M. Rubinstein, J. L. Bermudez, W. Janischewskyj, V. Shostak, C. A. Nucci, A. M. Hussein, and J. S.
Chang, "On return stroke currents and remote electromagnetic fields associated with lightning strikes to tall structures: Part I:
Computational Models," Journal of Geophysical Research, vol. 112, 2007.
[53] D. Pavanello, "Electromagnetic Radiation from Lightning Return Strokes to Tall Structures," vol. Ph.D. Lausanne: Swiss
Federal Institute of Technology (EPFL), 2007, pp. 134.
[54] D. Pavanello, F. Rachidi, M. Rubinstein, N. Theethayi, and R. Thottappillil, "Electromagnetic environment in the immediate
vicinity of a tower struck by lightning," presented at EUROEM2004, Magdeburg, Germany, 2004.
[55] S. Miyazaki and M. Ishii, "Influence of elevated stricken object on lightning return-stroke current and associated fields,"
presented at International Conference on Lightning Protection ICLP, Avignon, France, 2004.
[56] A. Mosaddeghi, D. Pavanello, F. Rachidi, and M. Rubinstein, "On the Inversion of Polarity of the Electric Field at Very Close
Range from a Tower Struck by Lightning," Journal of Geophysical Research (in press), 2007.
[57] E. Montandon, "Lightning positioning and lightning parameter determination experiences and results of the Swiss PTT
research project," presented at 21st International Conference on Lightning Protection ICLP, Berlin, 1992.
[58] K. L. Cummins, E. P. Krider, and M. D. Malone, "The US National Lightning Detection Network (TM) and applications of
cloud-to-ground lightning data by electric power utilities," IEEE Transactions on Electromagnetic Compatibility, vol. 40, pp.
465-480, 1998.