Upward leaders from instrumented lightning rods competing to connect a
downward leader during a lightning attachment process
Marcelo M. F. Saba1, Paola B.
Lauria1, Carina Schumann2, José
Claudio de O. Silva3, Felipe de L.
Mantovani1
1INPE – National Institute for Space Research - São
José dos Campos, Brazil.
2JLRL University of the Witwatersrand – Johannesburg,
South Africa.
3APTEMC - São José dos Campos, Brazil.
Corresponding author: Marcelo M. F. Saba
(marcelo.saba@inpe.br)
Key Points:
- Parameters of upward leaders from images, electric-field and current
measurements
- Competing upward leaders alternate their progression during initial
propagation
- Current pulses of upward leaders increase intensity and synchronize
right before attachment
Abstract
In this paper we analyze electric-field and current measurements of
competing upward leaders induced by a downward negative lightning flash
that struck a residential building. The attachment process was recorded
by two high-speed cameras running at 37,800 and 70,000 images per second
and the current measured in two lightning rods. Differently from
previous works, here we show, for the first time, the behavior of
multiple upward leaders that after initiation compete to connect the
negative downward moving leader. At the beginning of the propagation of
the leaders that initiate on the instrumented lightning rods, current
pulses appear superimposed to a steadily increasing DC current. The
upward leader current pulses increase with the approach of the downward
leader and are not synchronized but present an alternating pattern. All
leader speeds are constant. The upward leaders are slower than the
downward leader speed. The average time interval between current pulses
in upward leaders is close to the interstep time interval found by
optical or electric field sensors for negative cloud-to-ground stepped
leaders. The upward leaders respond to different downward propagating
branches and, as the branches alternate in propagation and intensity, so
do the leaders accordingly. Right before the attachment process the
alternating pattern of the leaders ceases, all downward leader branches
intensify, and consequently upward leaders synchronize and pulse
together. The average linear densities for upward leaders (49 and 82
µC/m) were obtained for the first time for natural lightning.
Plain Language Summary
The effectiveness of a lightning protection system depends on its
efficiency to intercept the down coming leader of a cloud-to-ground
lightning flash. The interception is usually done by an upward
connecting leader that initiates from grounded structures, humans, or
living beings that protrude from nearby ground. The understanding of the
upward connecting leader and of the attachment process with the downward
leader plays an important role in the determination of the zone of
protection and therefore in the improvement of a lightning protection
system. Unconnected upward leaders, i.e., upward leaders that fail to
connect the downward leader, are also of great importance in lightning
protection. They can be large enough to cause damage to equipment
vulnerable to sparks or induced currents, and enough to injure people
from who it initiates. In this paper we analyze electric-field and
current measurements of competing upward leaders induced by a downward
negative lightning flash that struck a residential building. The
attachment process was simultaneously recorded by two high-speed
cameras, an electric-field sensor, and current sensors installed on two
lightning rods. Differently from previous works, here we show, for the
first time, the behavior of multiple upward leaders that compete to
connect the negative downward moving leader.
1 Introduction
Previously, we have reported high-speed video images of attachment
process of three negative downward cloud-to-ground flashes to an
ordinary residential building (Saba et al., 2017). As mentioned in the
cited paper, the effectiveness of a lightning protection system (LPS)
depends on its efficiency to intercept the down coming lightning leader
which is related to its efficiency to emit upward connecting leaders
(UCL). The understanding of the characteristics of an UCL and of the
attachment process with the downward leader plays an important role in
the determination of the volume or zone of protection of a LPS and in
the improvement of LPS designs. Unconnected upward leaders (UUL), i.e.,
those events that initiate an upward leader but fail to make contact
with the downward leader, are also of great importance in lightning
protection. They can be large enough to cause damage to equipment
vulnerable to sparks or induced currents, and enough to injure people.
Although a few current measurements of upward leaders have been reported
from tall towers higher than 60 m (e.g. Saba et al., 2015; Visacro et
al., 2017; Arcanjo et al., 2019; Nag et al., 2021 for towers over
mountains), from buildings (Saba et al., 2017), and from small
structures (Schoene et al., 2008, vertical conductor of 7 m height), no
current measurements of upward connecting leaders from common
residential buildings have been reported in the literature. Moreover, no
study has ever been done on upward leaders competing to connect a
downward leader. Besides, some of these past studies do not have
electric-field and current measurements together with high-speed video
observations which is crucial to visualize what is happening with the
upward and downward leaders involved in the attachment process.
This study presents observational data of several positive upward
leaders competing to connect a negative leader of a downward
cloud-to-ground flash that strikes an instrumented lightning rod of a
residential building. It is the first to report current measurements of
two upward leaders induced by the same downward leader. The use of
high-speed video images and electric field measurements reveal the
nature of the physical process that is generating the currents measured
on the vertical lightning rods on the top of buildings.
2 Instrumentation
The lightning attachment to the building was observed by two high-speed
video cameras Vision Research Phantom v12 and v711 operating at 70,000
and 37,800 frames per second with exposure times of 13.55 µs and 25.85
µs and time intervals of 14.29 µs and 26.46 µs respectively (videos
available in Supplementary Information). Image spatial resolution used
for the flashes herein was 128 × 360 pixels and 368 × 416 pixels,
respectively. They were positioned at 220 m from a pair of identical
14-story apartment buildings, named P1 and P2, located in São Paulo City
(23.483°S, 46.728°W), Brazil (Figure 1a). Their steel reinforced
concrete structures are used as natural LPS. Each building has a
vertical lightning rod, and their tips are at a height of 52 m
respective to ground level. All reported distances and speeds given by
the analysis of the images from the high-speed videos were measured in
2D and therefore underestimated.
The electric field changes caused by the attachment process was measured
by a flat plate antenna with an integrator and amplifier. The antenna
was located on top of building P2 only 4 m away from the lightning rod
that was struck by a cloud-to-ground lightning flash (see Figure 1a). A
fiber-optic link was used to transmit the signal from the
integrator/amplifier to the digitizer. The bandwidth of the system
ranged from 306 Hz to 1.5 MHz. The physics sign convention is used when
referring to the electric field and its change. The approach of a nearby
negative leader produces positive electric field change, and a negative
CG return stroke produces a negative field change.
One current transformer sensor (Pearson model 301-X) was installed on
the lightning rod of each building. This current sensor is capable of
recoding current up to 50,000 A with a useable rise time of 200
nanoseconds, a low frequency 3 dB cut-off of approximately 5 Hz and a
high frequency 3 dB cut-off of approximately 2 MHz. The output of the
sensor is split in two channels (20 dB and 50 dB attenuation over 50 Ω)
and sent to a data acquisition system through a pair of fiber optic
links. Before installation, both sensors were tested and calibrated in a
high voltage facility. The electric field and current were continuously
recorded at a sampling rate of 5 MS/s. GPS antennas were used to
synchronize all measurements and video images.
Data from lightning location systems (LLS) were used to obtain the
polarity, the time, and an estimate of the peak current of the return
stroke. A complete study on the accuracy of peak current estimation
given by the LLS has not been performed yet. However, for one recent
event of a cloud-to-ground flash that struck one of the buildings, the
error was within 20% for the strokes that were correctly classified as
cloud-to-ground. In that event, four strokes were detected by the LLS
and they were directly measured by the current sensor installed in the
vertical lightning rod to where the attachment occurred. Further
information about these systems and their performance are given in
Naccarato et al. (2012 and 2017).
More information about the cameras, the locations of the two buildings
and the topography of the terrain can be found in the previous work
(Saba et al., 2017).
3 Data
On 1 February 2017 a single-stroke negative cloud-to-ground lightning
discharge struck the tip of the lightning rod B on building P2.
According to the LLS, its peak return stroke current was -73 kA and
occurred at 19:01:10.689307 (UT). During the approach of the stepped
leader, a positive UCL was launched from the tip of the lightning rod B
on building P2 together with five positive UULs from the vertical
air-termination rod A of the other building (P1) and other nearby
structures and corners (named C, D, E, F), as shown in Fig. 1. The first
upward leader to start a continuous propagation was the UCL leader. It
started propagation when the downward leader closest tip was 102 m away
from the tip of the P2 lightning rod where it started. The leaders had
their origin at different distances from the electric field sensor and
at different times (t = 0 s correspond to the attachment and beginning
of return stroke in all tables and graphs). The leader types, 2D lengths
(measured one frame before the occurrence of the return stroke), their
horizontal distances from the electric field sensor, and inception times
are shown in Table 1.