Cloud-to-ground lightning observations over the eastern Amazon Region

Arthur C. Almeida
a,
, Brgida R. P. Rocha
a
, Jos Ricardo S. Souza
a
,
Jos Alberto S. S
a
, Jos A. Pissolato Filho
b
a
Federal University of Para (UFPA) Av. Augusto Correa no. 1, Belem, PA, 66075-110, Brazil
b
State University of Campinas (UNICAMP) C. U. Zeferino Vaz, Campinas, SP, 13083-970, Brazil
a r t i c l e i n f o a b s t r a c t
Article history:
Received 30 January 2011
Received in revised form 8 July 2011
Accepted 29 August 2011
Cloud-to-ground (CG) lightning variables, space and time distributions near the equator, over
Eastern Amazon, were analyzed in this work. Over seven million lightning flash events of this
type, detected by a VAISALA lightning detection network were processed. This data set was
collected by the Brazilian Amazon Protection System between 2006 and 2008. The results
showed the sub-areas of higher occurrence densities, polarity types, current peak intensity
intervals, as well as, the hourly distributions of CG lightning, for this region. A considerable
percentage (7%) of discharges with stroke current peaks above 100 kA, was observed. An at-
tempt was made to compare the statistical results of this work with observations made else-
where in Brazil and other regions of the world. The differences found suggest that, some
regional peculiarities of the CG lightning parameter values should be taken into consider-
ation, for electric and telecommunication systems protection against atmospheric electric
discharges, in the Amazon Region.
2011 Elsevier B.V. All rights reserved.
Keywords:
Amazon Region
Atmospheric discharges
CG lightning
Electric systems protection
1. Introduction
Considering its large area, humid tropical climate and
frequent cumulonimbus cloud cover, the Brazilian Amazon
Region is supposed to be the target of most of the lightning
discharge occurrences in Brazil. The performance of the
electric distribution systems in Amazonia is significantly
affected by cloud-to-ground (CG) lightning events (Santos
et al., 2011).
This fact has to be considered since several hydroelectric
power generating dams are being built now, and are sup-
posed to supply electric energy to the center and southern
regions of Brazil in the next decade, through very long trans-
mission lines.
The electric energy and telecommunications sectors have
often suffered disruptions attributed to lightning, which
resulted in heavy economic losses to the most populated
states in southern Brazil. Therefore, the determination of
lightning occurrence characteristics in the Amazon Region is
of great importance for economic interests in this country.
Except for some pioneering studies made, in the past
decades (Souza et al., 1999; Rocha et al., 1996), lightning
observations with modern equipment remained scarce in
that region The demand for characteristics of lightning in
the Amazon began to be supplied more recently after the
installationof regional lightning locationsystems. It is expected
that this study will contribute to establish parameters to opti-
mize the choice of appropriate facilities and equipments, capa-
ble of withstanding a large percentage of CG lightning direct
strikes on the local transmission lines.
The lightning location systems (LLS) represent important
remote data sources on lightning variables and cover several
regions of the Earth. There are more than 60 LLS networks
worldwide (Cummins and Murphy, 2009). As examples of
large LLS networks, one may mention: the European net-
works (EUCLID and LINET) (Diendorfer, 2002; Betz et al.,
2008), the Austrian Lightning Locating System (ALLS)
(Diendorfer et al., 1998; Schulz et al., 2005), the Canadian
Lightning Detection Network (CLDN) (Burrows et al., 2002),
Atmospheric Research 117 (2012) 8690
Corresponding author.
E-mail addresses: arthur@ufpa.br (A.C. Almeida), brigida@ufpa.br
(B.R.P. Rocha), jricardo@ufpa.br (J.R.S. Souza), jalbertosa@ufpa.br (J.A.S. S),
pisso@dsce.fee.unicamp.br (J.A.P. Filho).
0169-8095/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.atmosres.2011.08.015
Contents lists available at SciVerse ScienceDirect
Atmospheric Research
j our nal homepage: www. el sevi er . com/ l ocat e/ at mos
the U.S. National Lightning Detection Network (NLDN) (Cum-
mins et al., 1998; Biagi et al., 2006; Zajac and Rutledge,
2001; Orville et al., 2002), and others in Australia (Kuleshov
et al., 2005), China (Chen et al., 2004), and Japan (Shindo and
Yokoyama, 1998).
In Brazil, there is the Brazilian Lightning Detection Network
(BrasilDAT) which is composed of 47 sensors as a result of
the integration of three regional networks: RINDAT, SIDDEM
in southern Brazil, and SIPAM, in eastern Amazonia (Pinto Jr.
et al., 2007; Pinto Jr. and Pinto, 2008; Naccarato and Pinto Jr.,
2009; Pinto Jr. et al., 1996).
It is known that lightning occurrence space and time
distributions, depend on the regional climatology (Ribeiro
et al., 2011a), prevailing meteorological systems (Mattos
and Machado, 2011), land surface cover and uses (Ribeiro
et al., 2011b), topography (Bourscheidt et al., 2009) and
other factors. Previous analyses using data collected by the
SIPAM-LDN, gave preliminary indication that the regional
events, seemed to present higher spatial density of occur-
rences, and produced higher average intensity of the first
stroke peak currents, than the worldwide average values
(Almeida et al., 2010), adopted in electrical discharge pro-
tection projects. It has been observed also that the regional
lightning occurrence densities show a significant seasonal
increment during the months between December and May,
which correspond to the regional rainy season (Ribeiro et al.,
2011b).
This paper pursues this line of study, given the relevance
of these variables to the adequate protection of people, build-
ings, telecommunications and electric systems in the Amazon
Region.
2. Methodology
2.1. The SIPAM Lightning Detection Network
During the year 2005, the Brazilian Amazonian Protection
System Agency (SIPAM) began to operate its VAISALA, 12
LPATS IV sensor network centralized at a CP 8000 model sta-
tion located in Belm, in the northern portion of Brazil. Fig. 1
shows the geographical locations of the SIPAM LDN sensors.
The data collected by this system are stored in a continu-
ous sequence at the central station and recovered as text files
in UALF format, with twenty five attributes associated to each
stroke event.
A data set of flash events observed between October 2006
and December 2008 was selected for analysis. These data
correspond to a good maintenance and improved perfor-
mance period of the system. The data processing deliber-
ately excluded lightning flashes identified as cloud-to-cloud
(CC) or intra cloud (IC) events, as well as, preliminarily those
with estimated peak currents above 250 kA.
The intra-cloud flashes were excluded considering their
small significance to the subject of protection of electric
Fig. 1. SIPAM-LDN sensor locations.
87 A.C. Almeida et al. / Atmospheric Research 117 (2012) 8690
systems against lightning discharges. About nine thousand
flashes (0.13% of the overall events selected) with negative
or positive peak currents exceeding 250 kA were disregarded
for the main study, but analyzed later on, as a separate data
set of high peak current values, in the range of the so called
super strokes. The reason for this separation was that peak cur-
rent values are derived from a theoretical model and those
values may have been overestimated (Cooray, 2009), and the
VAISALA LPATS IV User's Guide Manual suggests that these re-
cords may be unreliable. Besides these exclusions, the analysis
was restricted to data collected within a rectangle defined by
geographic coordinates between longitudes 41 to 55 W and
latitudes between 1 to 15 S. This space coverage limitation
for this LDNdata was made, considering that this LDN presents
a 50 to 80% detection efficiency (DE) in the area considered
and shows an estimated abrupt decay of the lightning DE
outside of its borders, as indicated from studies by Pereira
et al. (2008) and Naccarato and Pinto Jr. (2009).
3. Results and discussion
The data input, collected within a grid of 0.1 mesh side,
after the processing yielded the flash density map displayed
in Fig. 2. This map reveals three main areas of higher CG light-
ning occurrence densities. Two of them show a more or less
northsouth distribution from2 to 10 S, somewhat separated
by a lower density area between 5 and 7 S, approximate-
ly. The first is centered around 4 S, 48 W, which is the
best efficiency point of the system, in the eastern part of
the State of Par. The second is around 8 S, 48 W, over the
northern part of the State of Tocantins, where a mountainous
region exists. The third stripe of higher flash density has a
northwest to southeast configuration over the continent, but
not too far from the Atlantic Ocean, following the coastal
areas of Maranho and Piau States. Its center was found to be
at the gulf of So Luis, around2 S, 44 W. This area is frequently
covered by squall line cloudiness produced by sea breezes.
It is evident that, the predominant feature conditioning
the appearance of these three areas was the sensor
distribution and detection efficiency of this LDN. Neverthe-
less, the meteorological and physiographic influences over
the CG lightning event distribution, were found also to be
very important for these regional CG lightning density
distributions.
A significant part of the area analyzed presented CG
lightning densities of up to 11 flashes/km
2
/year. Taking into
account the previously mentioned estimates of the detection
efficiency of this system, that density value could actually
double, which would place such regional values at a higher
level than elsewhere in the world, for this type of lightning.
It would also be comparable to the 17 events/km
2
/year
over the urban area of the city of So Paulo, reported for
all types of lightning by Pinto Jr. (2008).
The complete data set of CG lightning was processed to
separate the flashes by hourly intervals of occurrences and
four classes of their first stroke peak current intensities. The
current value intervals arbitrarily established were: 030;
3150; 51100; and 101250 kA, and resulted in Fig. 3.
The hourly evolution of the lightning number of events
segmented by the current peak intervals shows that the CG
lightning occurrence rate peaks at 15 h local time, therefore
nearly 1 h before the local hourly precipitation maximum.
The electrical activity remains high until 19 h (LT), dropping
to another level, before dawn. After dawn, this lightning
activity decreases again reaching its smallest daily value at
about 10 h (LT). A similar hourly distribution of the number
of lightning events was observed by Souza et al. (2009) for
southern Brazil, from TRMM-LIS satellite data.
These results show that the risk of damage to installa-
tions and electricity transmission lines increases about 15
times between 10 and 15 h LT, over this region. Conse-
quently, external routine operations and maintenance, should
be done in the morning and avoid the afternoon hours, in
order to minimize risks to technical personnel on the field.
It is evident from Fig. 3 that, the high peak current strokes
which exceed 100 kA may occur at any hour of the day. Their
occurrence rate seems to follow the occurrence rate incre-
ments of all other stroke current distribution intervals.
Fig. 2. Lightning spatial distribution (CG lightning events/km
2
/year).
88 A.C. Almeida et al. / Atmospheric Research 117 (2012) 8690
Fig. 4 displays the hourly CG lightning distributions, accord-
ing to their positive or negative first stroke polarity. The
number of events was presented as a percentage of each CG
lightning polarity, with relation to the total data set consid-
ered. Although the percentages of negative first stroke events
exceeded those of positive polarity at all hours, the ratio of
negative to positive strokes was considerably higher between
11 and 19 h LT. On the other hand the percentage of positive
first strokes was much higher than the 5 to 15% of the overall
CG events observed in other regions of the Earth. It has yet to
be determined by further studies, whether this is a regional
feature or may be attributed to occasional misinterpretation
of small current intra cloud events as positive CG light-
ning, by this VAISALA system. In order to avoid such even-
tual contamination of the positive CG events, Naccarato
and Pinto Jr. (2009), disregarded small stroke data for
positive CG analysis of data collected by BrasilDAT in
southeast Brazil. However, all current values were consid-
ered for the polarity analysis in this work.
Fig. 5 shows the log-normal distribution of peak current
in 10 kA bins. One can observe that the most frequent value
belongs to the class interval of 1020 kA. These current
peak values are a little over the corresponding values found
by Jerauld et al. (2005) from the US NLDN data. The estimat-
ed median for our data was of 42 kA and their average value
was 49 kA, with a standard deviation of 35 kA. These are
values similar to those measured at a lightning monitoring
tower in southeast Brazil, as reported by Pinto Jr. (2008).
Although the VAISALA LPATS IV User's Guide Manual does
not recommend the use of the so called super stroke data, of
current that registers above 250 kA, just for the sake of addi-
tional analysis, they were considered below.
Analysis done on a separate data set containing lightning
events with peak currents above 250 kA exhibited the distribu-
tion in Fig. 6. For these events, it was observed that the negative
lightning flashes were more frequent and intense. The maxima
positive and negative recorded peak currents were considered
outliers and were not shown on the histogram. The stroke
number for both polarities decayed towards larger current in-
tensities, as expected, and maintained the same characteristics
observed in Fig. 5, drawn for smaller peak current values.
4. Conclusions
A data set of about seven million CG lightning flashes
recorded by the SIPAM LDN between October 2006 and
December 2008, over an area of the eastern Amazon Region,
was analyzed.
The stroke polarities and peak current intensities, as well
as, the space and time distributions of these events were
drawn, seeking subsidies for the adequate protection and
operation of electric systems, against these atmospheric
electric discharges.
Fig. 3. Lightning distribution per hour and peak current intensity class.
Fig. 4. Positive and negative lightning overall percentage per hour.
Fig. 5. Peak current log-normal distribution.
Fig. 6. Super discharges polarity and peak current (N250 kA) distribution.
89 A.C. Almeida et al. / Atmospheric Research 117 (2012) 8690
The spatial distribution of flashes showed three areas of high
density of occurrences, as follows: just southward of the city
of Belm, the northern part of the State of Tocantins and in-
land from the beaches of So Luis Bay. Preliminary inter-
pretations for those concentrations of the CG lightning
incidence were given.
CG lightning occurrence densities of up to 11 events/km
2
/
year were found on large areas of the geographical rectangle
studied. This value may have been underestimated with regard
to its actual value, since the detection efficiency of the system
was computed to be between 50 and 80%, as a function of its
sensors space configuration.
The hourly event distribution indicated a maximum level
of occurrences between 15 and 20 h LT. During these hours
of the day, external maintenance of electric system should
be avoided, if one wishes to minimize lightning hazards.
Ninety three (93%) percent of the analyzed strokes were
found to be below 100 kA. The remaining seven (7%) percent
was determined to be between 100 and 250 kA, which is a
high fraction compared with statistics in other places in the
world. About nine thousand strokes exceeded 250 kA which
represented about 0.13% of the data set analyzed. Since
these strokes were detected by five or six sensors of the sys-
tem they were set apart for an extra processing. The median
value of the first stroke currents of the regional CG lightning
was determined as 42 kA, and their average was 49 kA. These
values are considered to be large in the literature, but are
compatible with currents measured at an instrumented
tower in southeastern Brazil, by other authors.
This work gives useful information about the characteris-
tics of CG lightning parameters over the Amazon Region,
which should be taken into consideration while projecting
systems to protect life and property in that region.
Acknowledgments
The authors wish to acknowledge the support received
through grants from the Brazilian National Research Council
(CNPq) and the Financier of Studies and Projects (FINEP)
REMAM 2 Project. They also thank the Amazonian Protection
System (SIPAM), for the lightning data analyzed in this work.
References
Almeida, A.C., Rocha, B.R.P., Souza, J.R.S., Monteiro, J.H.A., S, J.A.S., 2010. Cloud-
to-ground lightning observations over the eastern Amazon Region:
subsidies for the protection of electric systems. Proceedings of 30th Inter-
national Conference Lightning Protection-ICLP 2010. Cagliari, Italy.
Betz, H.-D., Schmidt, K., Oettinger, W.P., 2008. LINET an international VLF/
LF lightning detection network in Europe. In: Betz, H.-D., Schumann, U.,
Laroche, P. (Eds.), Lightning Principles, Instruments and Applications.
Springer-Verlag, Dordrecht, NL. ch. 5.
Biagi, C.J., Cummins, K.L., Kehoe, K.E., Krider, E.P., 2006. National Lightning
Detection Network (NLDN) performance in southern Arizona. J. Geo-
phys. Res. 112, D05208. doi:10.1029/2006JD007341.
Bourscheidt, V., Pinto Jr., O., Naccarato, K.P., Pinto, I.R.C.A., 2009. The influ-
ence of topography on the cloud-to-ground lightning density in South
Brazil. Atmos. Res. 91, 508513.
Burrows, W.R., King, P., Lewis, P.J., Kochtubajda, B., Snyder, B., Turcotte, V., 2002.
Lightning occurrence patterns over Canada and adjacent United States
from lightning detection network observations. Atmos. Ocean 40, 5981.
Chen, S.M., Du, Y., Fan, L.M., 2004. Lightning data observed with lightning
location system in Guang-Dong Province, China. IEEE Trans. Power
Deliv. 19 (3), 11481153.
Cooray, V., 2009. On the upper limit on peak current in return strokes of
lightning flashes. Proceedings of X SIPDA, pp. 167172. Curitiba, Brazil.
Cummins, K.L., Murphy, M.J., 2009. An overview of lightning locating sys-
tems: history, techniques, and data uses, with an in-depth look at the
U.S. NLDN. IEEE Trans. Electromagn. Compat. 51 (3), 499518.
Cummins, K.L., Krider, E.P., Malone, M.D., 1998. The U.S. national lightning de-
tection network and applications of cloud-to-ground lightning data by
electric power utilities. IEEE Trans. Electromagn. Compat. 40 (4), 465480.
Diendorfer, G., 2002. EUCLID technical structure and performance of the
European wide lightning location system. Proceedings of Int. Conf.
Grounding Earthing 3rd Brazilian Workshop Atmos. Electr, pp. 3942.
Rio de Janeiro, Brazil.
Diendorfer, G., Schulz, W., Rakov, V.A., 1998. Lightning characteristics based
on data from the Austrian lightning locating system. IEEE Trans. Electro-
magn. Compat. 40 (4), 452464.
Jerauld, J., Rakov, V.A., Uman, M.A., Rambo, K.J., Jordan, D.M., 2005. An
evaluation of the performance characteristics of the U.S. National
Lightning Detection Network in Florida using rocket-triggered light-
ning. J. Geophys. Res. 110, D19106. doi:10.1029/2005JD005924.
Kuleshov, Y., Mackerras, D., Darveniza, M., 2005. Spatial distribution and
frequency of lightning activity and lightning flash density maps for
Australia. J. Geophys. Res. 111, D19105. doi:10.1029/2005JD006982.
Mattos, E.V., Machado, L.A.T., 2011. Cloud-to-ground lightning and meso-
scale convective systems. Atmos. Res. 99, 377390.
Naccarato, K.P., Pinto Jr., O., 2009. Improvements in the detection efficiency
model for the Brazilian lightning detection network (BrasilDAT).
Atmos. Res. 91, 546563.
Orville, R.L., Huffines, G.R., Burrows, W.R., Holle, R.L., Cummins, K.L., 2002.
The North American Lightning Detection Network (NALDN) first
results: 19982000. Mon. Weather. Rev. 130, 20982109.
Pereira, C.S., Aleixo, J.G., Rocha, E.P., Simes Neto, B., 2008. A statistical
approach to the determination of the lightning detection efficiency of
the SIPAM's Lightning Detection Network. Proceedings of the Interna-
tional Conference on Grounding and Earthing & 3rd International
Conference on Lightning Physics and Effects, Florianopolis, Brazil.
Pinto Jr., O., 2008. An overview of cloud-to-ground lightning research in
Brazil in the last two decades. Proceedings of 20th International Light-
ning Detection Conference & 2nd International Lightning Meteorology
Conference. Tucson, Arizona, USA.
Pinto Jr., O., Pinto, I.R.C.A., 2008. On the sensitivity of cloud-to-ground lightning
activity to surface air temperature changes at different timescales in Sao
Paulo, Brazil. J. Geophys. Res. 113, D20123. doi:10.1029/2008JD009841.
Pinto Jr., O., Pinto, I.R.C.A., Naccarato, K.P., 2007. Maximum cloud-to-ground
lightning flash densities observed by lightning location systems in the
tropical region: a review. Atmos. Res. 84, 189200.
Pinto Jr., O., Gin, R.B.B., Pinto, I.R.C.A., Mendes Jr., O., Diniz, J.H., Carvalho,
A.M., 1996. Cloud-to-ground lightning flash characteristics in south-
eastern Brazil, for the 19921993 summer season. J. Geophys. Res.
101 (D23), 2962729635.
Ribeiro, W.M.N., Souza, J.R.S., Lopes, M.N.G., Lopes, J.L.M., Teixeira, V.A., Paca,
2011a. V.H.M. lightning and precipitation produced by severe weather
systems, over Belem, Brazil. Proceedings of XIV International Conference
on Atmospheric Electricity-ICAE 2011. Rio de Janeiro, Brazil.
Ribeiro, W.M.N., Souza, J.R.S., Almeida, A.C., Santos, F.A.A., Rocha, B.R.P.,
2011b. Lightning and precipitation distributions over different surface
types, in Eastern Amazonia. Proceedings of XIV International Conference
on Atmospheric Electricity-ICAE 2011. Rio de Janeiro, Brazil.
Rocha, B.R.P., Souza, J.R.S., Costa, J.M., 1996. Electric and magnetic fields from
lightning in Belem. Proceedings of II Brazilian Congress of Electromagne-
tism, pp. 5558. Ouro Preto, Brazil.
Santos, A.P.P., Souza, J.R.S., Souza, E.B., Carmo, A.M.C., Ribeiro, W.M.N.,
Oliveira, R.A.J., 2011. The climatic context of lightning storms, asso-
ciated to electric systems power outages in eastern Amazonia. Pro-
ceedings of XIV International Conference on Atmospheric Electricity-ICAE
2011. Rio de Janeiro, Brazil.
Schulz, W., Cummins, K., Diendorfer, G., Dorninger, M., 2005. Cloud-to-
ground lightning in Austria: a 10-year study using data from a lightning
location system. J. Geophys. Res. 110, D09101. doi:10.1029/2004JD005332.
Shindo, T., Yokoyama, S., 1998. Lightning occurrence data observed with
lightning location systems in Japan: 19921995. IEEE Trans. Power
Deliv. 13 (4), 14681474.
Souza, P.E., Pinto Jr., O., Pinto, I.R.C.A., Ferreira, N.J., Santos, A.F., 2009. The
intracloud/cloud-to-ground lightning ratio in Southeastern Brazil.
Atmos. Res. 91, 491499.
Souza, J.R.S., Rocha, B.R.P., Carrera, G.T., 1999. CG-lightning observations
(and applications) around Belm during the 19951998 period. Pro-
ceedings of the V International Symposium on Lightning Protection,
vol. 1, pp. 1721. So Paulo, Brazil.
Zajac, B.A., Rutledge, S.A., 2001. Cloud-to-ground lightning activity in the
contiguous United States from 1995 to 1999. Mon. Weather Rev.Am.
Meteorol. Soc. 129 (5), 9991019.
90 A.C. Almeida et al. / Atmospheric Research 117 (2012) 8690