High-speed video observations of positive ground flashes produced by

intracloud lightning

Marcelo M. F. Saba,1 Leandro Z. S. Campos,1,2 E. Philip Krider,3 and Osmar Pinto Jr.1

Received 20 April 2009; revised 18 May 2009; accepted 28 May 2009; published 26 June 2009.

[1] High-speed video recordings of two lightning flashes
confirm that positive cloud-to-ground (CG) strokes can be
produced by extensive horizontal intracloud (IC) discharges
within and near the cloud base. These recordings constitute
the first observations of CG leaders emanating from IC
discharges of either polarity. In one case, the discharge
began with a negative leader that propagated horizontally,
went upward and produced an IC discharge. After the
beginning of the IC discharge, a positive leader emanated
from the lowest portion of the IC discharge, and initiated a
positive return stroke. In the other case, the IC discharge
began with a positive leader and then initiated a downward-
propagating positive leader that contained recoil processes
and produced a bright return stroke followed by a long
continuing luminosity. These observations help to
understand the complex genesis of positive CG flashes,
why IC lightning commonly precedes them and why
extensive horizontal channels are often involved.
Citation: Saba, M. M. F., L. Z. S. Campos, E. P. Krider, and

O. Pinto Jr. (2009), High-speed video observations of positive

ground flashes produced by intracloud lightning, Geophys. Res.

Lett., 36, L12811, doi:10.1029/2009GL038791.

1. Introduction

[2] Electric field records have shown that positive light-
ning flashes to ground are often preceded by significant
intracloud (IC) discharge activity lasting, on average, more
than 100 ms [Fuquay, 1982] or 200 ms [Rust et al., 1981].
Optical measurements also show that positive discharges to
ground often involve long, horizontal channels, up to tens of
kilometers in length [e.g., Fuquay, 1982; Kong et al., 2008;
Saba et al., 2008]. It is now generally accepted that one
cause of the red sprites (and other transient luminous
events) that occur at high altitudes are horizontally exten-
sive flashes that transfer large amounts of positive charge to
ground [Boccippio et al., 1995; Rodger, 1999; Moudry et
al., 2003; Pasko, 2007; Lyons et al., 2008; Mika and
Haldoupis, 2008]. According to Valdivia et al. [1997] the
electromagnetic pulses from the IC portion of a positive
cloud-to-ground +CG lightning may cause the nucleation of
sprites and according to Yashunin et al. [2007] M compo-
nents have the potential to initiate transient luminous events

during the continuing current stage and explain their delay
from the parent lightning.
[3] Another characteristic of positive flashes to ground is

that multiple strokes (i.e., strokes subsequent to the first) are
very infrequent and usually do not follow the same path to
ground as the first stroke. For example, all the subsequent
strokes in multiple-stroke positive flashes observed in Japan
(M. Ishii et al., Termination of multiple-stroke flashes
observed by electromagnetic field, paper presented at the
24th International Conference on Lightning Protection,
Staffordshire University, Birmingham, U. K., 1998) and in
Brazil (M. M. F. Saba et al., High-speed video observations
of positive lightning, paper presented at IX International
Symposium on Lightning Protection, Institute of Electro-
technology and Energy, Foz do Iguaçú, Brazil, 2007) created
new ground terminations.
[4] Based on the above facts, Rakov and Uman [2003]

have suggested that if the cloud conditions are right, a
positive discharge to ground can be initiated by a branch of,
or otherwise produced by, an extensive IC discharge. Here,
we will describe video observations of two (+CG) flashes,
one in Brazil and one in the U.S., that confirm this
hypothesis and show the close relationship between IC
and +CG discharges.

2. Instrumentation

[5] A RedLake MotionScope 8000S high-speed digital
video camera (1000 frames per second) was used to record
images of cloud-to-ground flashes in Southern Brazil be-
tween December 2006 and March 2007. During August
2007, the same camera, together with a Photron PCI-512
high-speed digital camera (4000 frames per second) was
used to record several flashes in Tucson, Arizona. The high-
speed video images were GPS time-stamped and stored in the
cameras until a signal was received from an external source,
and the trigger point could be set to determine how many
video frames were read out before the event of interest. Each
trigger pulse was initiated manually by an operator when
a flash was observed within the camera field-of-view. For
more details on the operation and accuracy of high-speed
cameras for lightning observations, see Saba et al. [2006].
[6] In order to determine the stroke location, polarity, and

an estimate of the peak current, we used data from two
Vaisala lightning location systems, BrasilDat in Brazil and
the NLDN in the United States. More information on the
characteristics of these networks is given by O. Pinto et al.
(Recent upgrades to the Brazilian Integrated Lightning
Detection Network, paper presented at 19th International
Lightning Detection Conference, Vaisala, Tucson, Arizona,
2006) and Cummins et al. [1998], respectively. The match-
ing of strokes between cameras and networks was done

GEOPHYSICAL RESEARCH LETTERS, VOL. 36, L12811, doi:10.1029/2009GL038791, 2009

1INPE, National Institute for Space Research, S. José dos Campos, São
Paolo, Brazil.

2Departamento de Fı́sica e Quı́mica, UNESP, Guaratinguetá, São Paulo,
Brazil.

3Institute of Atmospheric Physics, University of Arizona, Tucson,
Arizona, USA.

Copyright 2009 by the American Geophysical Union.
0094-8276/09/2009GL038791

L12811 1 of 5



using GPS time-synchronization (to an accuracy better than
1 millisecond).

3. Observations

[7] A total of 39 +CG flashes were recorded in both
campaigns (28 in Southern Brazil and 11 in Tucson, AZ). IC
discharge activity was nearly always recorded before and
after the +CG flashes, and in the two cases that will be
described here, IC channels were clearly visible at or near
the cloud base with characteristics that indicated they
initiated positive downward leaders and +CG strokes (as
reported by the lightning locating networks) followed by
long continuing current. The luminosity produced by the
continuing current was recorded by the camera in both
flashes and will be referred in this work as continuing
luminosity (CL).

3.1. Case 1: Southern Brazil

[8] Figure 1 shows a sketch and timeline of the first event
that was a bipolar CG flash in Southern Brazil (14:58 UT,
March 26, 2007). According to BrasilDAT, the positive CG
stroke struck ground at a distance of 13 km from the camera
and had an estimated peak current of +21 kA. The overall
discharge began with a sequence of five negative cloud-to-
ground (�CG) strokes (see Table 1), each of which
appeared to have a different ground termination and no
continuing luminosity (CL). 222 ms after the last negative
stroke, a new negative leader emanated from the same
region of the cloud and propagated horizontally and to the
right in the camera field-of-view. After crawling along
the cloud base to the left (between the heights of 2.5 and
4.0 km) and then going upward toward higher parts of the
cloud (Figure 2a), this leader initiated a very bright IC
discharge that was followed by a CL that lasted for 365 ms

(Figure 2b). The CL contained several intensifications of
luminosity that were similar to M-components in CG
flashes [Campos et al., 2007, 2009]. 58 ms after the
beginning of the IC discharge, a new leader emanated
from the lowest portion of the IC channel (Figure 2c) and
created a positive stroke to ground and a new ground
termination (Figure 2d). Both the IC and the +CG chan-
nels remained luminous for hundreds of milliseconds
(Figure 2e), but the latter lasted longer and had several
M-components (more details on the M-components in this
and other positive flashes and the methodology used to
analyze them are given by Campos et al. [2007, 2009]).
[9] The height of the positive charge center in Case 1 could

be estimated by adding the height of the leader before it left
the camera field-of-view to an estimate of the distance it
traveled during the next three video frames (i.e., before the
initiation of the IC discharge). The latter distance was esti-
mated by multiplying the average vertical speed (Table 2) by
the duration of the last three video frames (3 ms). The hori-
zontal distance from the camera to the negative charge center
was assumed to be the average range of the�CG strokes, and
the horizontal distance to the positive center was assumed to
be the range of the +CG stroke. Because of the different
distances, the negative charge center appears to be below the
positive center in Figure 1.
[10] It should be noted that all images are two-dimensional

projection of the channel. Therefore, all heights inferred in
this case and in the case described below are rough estimates.
However, a good agreement was obtained between these
rough estimates and the information given by radiosondes for
the heights of the �10 to �25 �C temperature levels where
the negative charge centers are usually located [Rakov and
Uman, 2003].

3.2. Case 2: Arizona

[11] Case 2 occurred during a large, monsoon thunder-
storm in Southeastern Arizona, (14:58 UT, July 31, 2007),
and a sketch and timeline of the development of this flash is
shown in Figure 3. According to the NLDN data, this +CG
stroke struck ground 6.4 km from the camera and had an
estimated peak current of +23 kA. The horizontal distance
to both charge centers depicted in Figure 3 was assumed to
be the range of the +CG stroke, and the height of the charge
centers was estimated using the same method employed in
Case 1.
[12] The flash started with a leader that propagated

horizontally to the left in the camera field-of-view. After
crawling along the cloud base (between the heights of 2.2
and 2.7 km), and then going upward into higher parts of the

Figure 1. Timeline and a schematic representation of the
sequence of events that were recorded on video for Case 1.
The +CG struck 13 km from the camera and had an
estimated peak current of +21 kA.

Table 1. Some Characteristics of the Discharges in Case 1

Time (ms)
Discharge

Type
Estimated Peak
Current (kA)

Preceding Time
Interval (ms)

Distance From
the Camera (km)

0 �CG �30 – 19
49 �CG �17 49 24
94 �CG �20 45 24
147 �CG – 53 –
254 �CG �12 107 33
578 IC - 324 –
636 +CG +21 58 after IC;

382 after last �CG
13

L12811 SABA ET AL.: POSITIVE CG FLASHES FROM IC LIGHTNING L12811

2 of 5



cloud (Figure 4a), it initiated an IC discharge (Figure 4b).
During the last 22 ms of leader development (that lasted a
total of 149 ms), several recoil leaders (RLs) were observed
that indicated this leader had positive polarity [Saba et al.,
2008]. After the IC discharge, several new leader channels
emanated from its region and propagated toward ground
(Figure 4c). One of these leaders initiated a very bright
positive stroke that followed the right-hand portion of the IC
channel into the cloud (Figure 4d). A long CL followed this

single stroke for 216 ms (Figure 4e) and contained several
M-components. Contrary to the first case (discussed above),
the IC channel disappeared immediately after the beginning
of the +CG stroke, but several RLs retraced the left portion
of the IC channel (that disappeared after the return stroke)
and other previously-formed leader paths.

3.3. Leader Characteristics

[13] The horizontal leader that initiated the IC discharge
in Case 1 was much more branched and bright than the
initial IC leader in Case 2. This suggests that the IC leader
in Case 1 was of opposite polarity to the leader in Case 2
(i.e., negative in Case 1 and positive in Case 2). The
durations of the leader propagation were 99 ms in Case 1
and 147 ms in Case 2. No significant increase in speed was
observed during the horizontal propagation in both cases.
Table 2 presents the duration and average speed of each IC
leader and the corresponding CG leader for each case.
[14] The speeds of both positive leaders increased as they

approached the ground (more details are given by Saba et

Figure 2. Selected negative images for the flash recorded
in Brazil (Case 1).

Table 2. Distance Traveled and Average Speed of the Leaders

Leader

IC + CG

Case 1
Negative
leader

Case 2
Positive
leader

Case 1
Positive
leader

Case 2
Positive
leader

Distance traveled (km) 8.9 8.3 2.5 2.7
Average Speed (ms�1) 0.9 � 105 0.6 � 105 1.6 � 105 1.6 � 105

Figure 3. Timeline and schematic representation of the
sequence of events that were recorded on video for Case 2.
The +CG struck 6.4 km from the camera and had an
estimated peak current of +23 kA.

L12811 SABA ET AL.: POSITIVE CG FLASHES FROM IC LIGHTNING L12811

3 of 5



al. [2008]). More measurements of positive and negative
leaders are found in the works by Saba et al. [2008, also
presented paper, 2007] and Kong et al. [2008], and there is a
good agreement between the values presented in those
papers and the values presented here.

4. Conclusion and Discussion

[15] Although both flashes produced positive CG strokes
and were initiated by leaders emanating from IC discharges
that were at or close to the cloud base, in Case 1 the IC flash

was initiated by a negative leader and in Case 2 it was
initiated by a positive leader. In addition to the leader
characteristics mentioned in section 3.3, the polarities of
the IC leaders were inferred from the facts listed below:
[16] 1. Case 1 the IC leader began in the same region of

the cloud that produced negative CG strokes. The channel
of the positive stroke was connected to the ending region of
the IC leader, which indicates that the IC leader was a
negative process propagating toward a positively charged
region.

Figure 4. Selected negative images for the flash recorded in Arizona (Case 2).

L12811 SABA ET AL.: POSITIVE CG FLASHES FROM IC LIGHTNING L12811

4 of 5



[17] 2. Case 2 the IC leader produced several RLs during
its development. This behavior was observed by Saba et al.
[2008] during the propagation of positive leaders towards
ground, and it has never been observed during the devel-
opment of negative leaders. Also, the channel of the positive
stroke was connected to the starting region of the IC leader
which indicates that this was a positively charged region.
[18] In Case 1, both the IC and the +CG channels

coexisted for a long time (approximately 300 ms), whereas
in the Case 2 they did not coexist at all. This may be
because in Case 1 a very intense CL was present in the IC
channel before the initiation of the +CG, whereas in Case 2
there was only an extremely faint luminosity in the IC
channel. The intense CL in Case 1 indicates that there was
likely intense ionization and a large charge transfer, and this
may have helped the IC and +CG channels to coexist after
the +CG stroke.
[19] At present, our knowledge about the physics of

positive lightning is not as good as that for negative
lightning, and many questions remain about the genesis of
positive discharges. Here, we have documented that +CG
flashes can be initiated by IC leaders of either polarity, and
this behavior may help us to understand why extensive
intracloud lightning commonly precedes positive CG
flashes.
[20] The association of extensive intracloud discharges

with positive flashes suggests that positive discharges
cannot always be modeled as the neutralization of simple,
vertically stacked monopoles [Rakov, 2003]. Moreover, this
complex structure can be helpful to studies on sprite
morphology [Mika and Haldoupis, 2008] and to studies
on why sprites are more common above some particular
kind of storms [Lyons et al., 2008].
[21] On the account of the large charge transfers in-

volved, positive CG strokes are of primary importance in
protection against lightning. Positive flashes are also a
major concern for lightning detection and location systems
because their complex waveforms are sometimes rejected or
attributed to intracloud flashes. A better knowledge of the
positive flash initiation and development may help in
identifying the electric and magnetic field waveforms that
are produced by positive CG flashes.

[22] Acknowledgments. We would like to thank M. G. Ballarotti,
N. J. Schuch, K. L. Cummins, C. D. Weidman, and S. A. Fleenor and for
their help in the data acquisition during the Brazil and Tucson campaigns.
This research has been supported by CNPq and FAPESP through the
projects 102356/2005-9, 475299/2003-5 and 03/08655-4, and the NASA
Kennedy Space Center, grant NNK06EB55G.

References
Boccippio, D. J., E. R. Williams, S. J. Heckman, W. A. Lyons, I. T. Baker,
and R. Boldi (1995), Sprites, ELF transients, and positive strokes,
Science, 269, 1088–1091.

Campos, L. Z. S., M. M. F. Saba, O. Pinto Jr., and M. G. Ballarotti (2007),
Waveshapes of continuing currents and properties of M-components in
natural negative cloud-to-ground lightning from high-speed video obser-
vations, Atmos. Res., 84, 302–310, doi:10.1016/j.atmosres.2006.09.002.

Campos, L. Z. S., M. M. F. Saba, O. Pinto Jr., and M. G. Ballaroti (2009),
Waveshapes of continuing currents and properties of M-components in
natural positive cloud-to-ground lightning, Atmos. Res., 91, 416–424,
doi:10.1016/j.atmosres.2008.02.020.

Cummins, K. L., M. J. Murphy, E. A. Bardo, W. L. Hiscox, R. B. Pyle, and
A. E. Pifer (1998), A combined TOA/MDF technology upgrade of the
U.S. National Lightning Detection Network, J. Geophys. Res., 103,
9035–9044.

Fuquay, M. D. (1982), Positive cloud-to-ground lightning in summer thun-
derstorms, J. Geophys. Res., 87, 7131–7140.

Kong, X., X. Qie, and Y. Zhao (2008), Characteristics of downward leader
in a positive cloud-to-ground lightning flash observed by high-speed
video camera and electric field changes, Geophys. Res. Lett., 35,
L05816, doi:10.1029/2007GL032764.

Lyons, W. A., S. A. Cummer, M. A. Stanley, G. R. Huffines, K. C. Wiens,
and T. E. Nelson (2008), Supercells and sprites, Bull. Am. Meteorol. Soc.,
89, 1165–1174, doi:10.1175/2008BAMS2439.1.

Mika, Á., and C. Haldoupis (2008), VLF studies during TLE occurrences in
Europe: A summary of new findings, Space Sci. Rev., 137, 489–510,
doi:10.1007/s11214-008-9382-8.

Moudry, D., H. Stenbaek-Nielsen, D. Sentman, and E. Wescott (2003),
Imaging of elves, halos and sprite initiation at 1 ms time resolution,
J. Atmos. Sol. Terr. Phys., 65, 509 – 518, doi:10.1016/S1364-
6826(00)00323-1.

Pasko, V. P. (2007), Red sprite discharges in the atmosphere at high altitude:
The molecular physics and the similarity with laboratory discharges,
Plasma Sources Sci. Technol., 16, S13–S29, doi:10.1088/0963-0252/
16/1/S02.

Rakov, V. A. (2003), A Review of positive and bipolar lightning discharges,
Bull. Am. Meteorol. Soc., 84, 767–776, doi:10.1175/BAMS-84-6-767.

Rakov, V. A., and M. A. Uman (2003), Lightning: Physics and Effects, 687
pp., Cambridge Univ. Press, New York.

Rodger, C. J. (1999), Red sprites, upward lightning, and VLF perturbations,
Rev. Geophys., 37, 317–336.

Rust, W. D., D. R. MacGorman, and R. T. Arnold (1981), Positive cloud to
ground lightning flashes in severe storms, Geophys. Res. Lett., 8,
791–794.

Saba, M. M. F., M. G. Ballarotti, and O. Pinto Jr. (2006), Negative cloud-to-
ground lightning properties from high-speed video observations,
J. Geophys. Res., 111, D03101, doi:10.1029/2005JD006415.

Saba, M. M. F., K. L. Cummins, T. A. Warner, E. P. Krider, L. Z. S.
Campos, M. G. Ballarotti, O. Pinto Jr., and S. A. Fleenor (2008), Positive
leader characteristics from high-speed video observations, Geophys. Res.
Lett., 35, L07802, doi:10.1029/2007GL033000.

Valdivia, J. A., G. Milikh, and K. Papadopoulos (1997), Red sprites:
Lightning as a fractal antenna, Geophys. Res. Lett., 24, 3169–3172.

Yashunin, S. A., E. A. Mareev, and V. A. Rakov (2007), Are lightning M
components capable of initiating sprites and sprite halos?, J. Geophys.
Res., 112, D10109, doi:10.1029/2006JD007631.

�����������������������
L. Z. S. Campos, O. Pinto Jr., and M. M. F. Saba, National Institute for

Space Research, P.O. Box 515, Av. dos Astronautas, 1758, São José dos
Campos, SP 12201-970, Brazil. (msaba@dge.inpe.br)
E. P. Krider, Institute of Atmospheric Physics, P.O. Box 210081, 1118 E.

4th Street, University of Arizona, Tucson, AZ 85721-0081, USA.

L12811 SABA ET AL.: POSITIVE CG FLASHES FROM IC LIGHTNING L12811

5 of 5