Detect and Locate Lightning Events from
Geostationary Satellite Observations
Report Part I
Review of existing lightning location systems
EUM/CO/02/1016/SAT
U. Finke and O. Kreyer
Institute für Meteorologie und Klimatologie
Universität Hannover
September 2002
1
2
Contents
1
Introduction
2
Lightning Location Basics
2.1 Lightning Process and Emissions . . . . . . . . . .
2.2 Atmospheric Influences on EM-Wave Propagation .
2.3 Lightning Detection . . . . . . . . . . . . . . . . .
2.4 Lightning Location . . . . . . . . . . . . . . . . .
2.5 Noise . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3
3 5 5 6 8
Ground Based Operational Networks
3.1 LF Sensors and Networks: Vaisala-GAI . . . . .
3.2 VHF Sensors and Networks: Vaisala-Dimensions
3.3 British VLF ATD . . . . . . . . . . . . . . . . .
3.4 Operational Networks Worldwide . . . . . . . .
3.5 Long Range Lightning Location . . . . . . . . .
3.6 Data Characteristics . . . . . . . . . . . . . . . .
3.7 Lightning Products and Services . . . . . . . . .
3.8 Network Summary . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
8
8 11 12
13
14 15 17 17 Satellite Based Lightning Location
4.1 Optical Location: OTD and LIS . . .
4.2 Combined Radio-Optical Observation:
4.3 VHF-Sensor: ORAGES . . . . . . . .
4.4 Satellite Summary . . . . . . . . . . .
. . . . .
FORTE
. . . . .
. . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
18
18 21
22
23
Use of Lightning Location Data
5.1 Comparative Studies . . . . . . . .
5.2 Climatologies . . . . . . . . . . . .
5.3 Weather Forecast and Nowcasting .
5.4 Total Lightning . . . . . . . . . . .
5.5 Atmospheric Chemistry . . . . . . .
5.6 Non – Weather Service Applications
5.7 Applications Summary . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
23
23 25
25
26
27
27
29
3
4
5
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
6
Summary and Conclusion
30
7
Acknowledgements
31
A Abbreviations
42
B Related Internet Links
43
1 INTRODUCTION
3
1 Introduction
Lightning is a dangerous and destructive atmospheric phenomenon. The strong currents associated
with the lightning discharge cause severe damages in the stroked object but also in its close vicinity
by electromagnetic induction. Lightning can kill and injure people and animals. Enormous economic
losses are caused by ignition of fire or damaging structures and facilities. Lightning is the main natural
origin of forest fires. With the constantly increasing use of electronic devices in the industrial and the
consuming sphere the vulnerability to lightning damages is growing.
Lightning is an indicator for convective activity in climatology and on the cloud scale. The lightning
rate of a thundercloud is correlated with other storm related phenomena like precipitation, hail and
gust. Lightning is an important natural source of nitrogen oxides contributing about 10% to the global
budget.
All these implications keep the high interest in lightning location data. This study reviews the current
state of operational lightning location systems (LLS). Special emphasis is given to the meteorological
applications of lightning data.
A comprehensive review of LLS was given by Holle and Lopez in 1993. Another overview is con­
tained in MacGorman and Rust (1998).
This study starts in section 2 with a description of the basic concepts of lightning location. A review
of the existing operational ground based networks is given in section 3, while section 4 describes the
space based location techniques. Finally, section 5 discusses the applications of the lightning location
data.
2 Lightning Location Basics
2.1 Lightning Process and Emissions
Lightning is a spark-like electrical discharge produced in thunderstorms. The various forms of light­
ning are discriminated by their location and by polarity of the transported charge into Cloud-to-Ground
lightning (CG) of positive or negative polarity and intracloud lightning (IC).
A lightning flash builds up in successive phases (Uman, 1987). In the following the phases of a typi­
cal negative cloud-to-ground flash are described. The initiation of lightning (preliminary breakdown)
takes place in cloud regions with sufficient strong electrical field. A barely visible discharge (stepped
leader) moves stepwise towards the earth. When the leader approaches the ground, an upward leader
is released from elevated objects near the ground and closes the lightning channel. This initiates the
return stroke, which transports charge between ground and cloud in currents of several 10 kA. Due to
Ohmic heating the temperature in the core reaches 30,000 K. This produces the visible lightning flash
and the thunder.
The first return stroke may be followed by subsequent return strokes each introduced by a ’dart leader’
using generally the same channel. This series of return strokes composes the lightning flash. The num­
ber of subsequent return strokes gives the ’multiplicity’ of the flash. Total flash duration is normally
less 1 s, the time intervals between the return strokes are in the order of few 100 ms.
Intracloud discharges take place between space charges inside the cloud. The leader process is fol­
lowed by several recoil streamers which are normally of less amplitude than CG return strokes.
2 LIGHTNING LOCATION BASICS
4
Lightning detection and location bases on the detection of the radiation emitted by lightning. During
the lightning process electromagnetic and acoustic radiation is generated in various forms:
radio by accelerated charges during the fast changing current steps. Radio emissions occur in form
of short pulses, the corresponding broad band frequency spectrum is roughly ∼ 1/f (Uman,
1987). Low frequency components are generated by the return stroke. The spectrum of the
other processes is concentrated in the higher frequency ranges. The time series of the magnetic
and electric components of radiation field exhibit a characteristic shape. The receiving systems
identify lightning by this wave form and distinguish between CG return strokes and IC lightning.
optical by thermal radiation of the hot (up to 30,000 K) lightning channel. Contributions to this
continuous spectrum come from the excited ionized and dissociated gases and compounds of
the atmosphere. The luminous period lasts a few milliseconds.
acoustical thunder, generated by the shock wave of the expanding hot air.
V H F
V H F
V H F
V H F
V H F
L F
L F
Figure 1: Radiation from lightning (schematically). The return stroke of Cloud-to-Ground lightning
emits strong LF radiation. Leader processes and recoil streamer in intracloud lightning generate pre­
dominately VHF radiation. All discharges emit optical radiation.
For the purpose of lightning location the electromagnetic radiation in the optical and radio range is
used. Most of the ground based operational lightning location systems operate in the radio frequency
ranges (Sec. 3) while the optical signal is employed in space based lightning sensors (Sec. 4).
Strong lightning flashes excite the ’Schumann’ resonances of the ground-ionospheric cavity (Schu­
mann, 1952; Volland, 1995). These resonances are of extreme low frequencies at 7.8, 14, 20, 26, 33,
39 and 45 Hertz with variations of 0.5 Hz. The analysis of the signals at only few globally distributed
points permits a lightning location with accuracy of a few 100 km.
2 LIGHTNING LOCATION BASICS
5
2.2 Atmospheric Influences on EM-Wave Propagation
The propagation of the electromagnetic waves in the atmospheric medium is affected by a number of
factors. They are listed below without detailed discussion. See e.g. Volland (1995) for the theory of
electro-magnetic wave propagation in the atmosphere.
Attentuation decrease of the wave’s amplitude due to increasing distance R from the source. The
amplitude decreases approximately with 1/R.
Ground reflections (not really an atmospheric effect) depending on the properties of the soil and
water bounding the propagation space. The propagation is also affected by large mountain
ridges, which delay the radiation signal (e.g. Rubinstein et al., 1994). Large obstacles may a
obscure a part of the sky and thus limit the detection sector. Additionally, metallic objects close
to the antenna can distort the magnetic field.
Ionospheric reflections Electromagnetic waves interact with the upper atmospheric layer of high
conduction (Ionosphere). For waves with frequencies f smaller than the ionospheric cutoff
frequency (fc ≈ 5 MHz) this results in reflexion back towards the ground. Thus long waves
propagate in a wave guide formed by the ground and the ionosphere and can travel over long
distances with low energy loss. Waves with higher frequency can penetrate the ionosphere.
Dispersion, scattering, refraction, absorption cause distortion of the wave front and change the
frequency spectrum.
For frequencies higher than 80 MHz the waves propagate in good approximation along the line of
sight. If the wave’s frequency is lower than 5 MHz, then the energy is trapped in the atmospheric
wave guide and long range propagation is observed.
The main propagation effect is the distance attenuation, which mainly reduces the amplitude. The
other propagation effects depend on frequency and lead to changes in the frequency spectrum and the
waveform of the signal. Particularly wave pulses become less steep.
2.3 Lightning Detection
Lightning detection sensors receive the electric and/or magnetic components of the radiation field
in certain frequency ranges. Low frequency (LF, 30 kHz – 300 kHz) and very high frequency (VHF,
30 MHz – 300 MHz) bands are employed, avoiding thereby the communication and radio transmission
bands. The used detection frequency (center and bandwidth) determines the detection range and
accuracy. Also the size of the antenna and the sensitivity to the various propagation effects depend on
the detection frequency.
2.3.1
Detectable Discharge Processes
Low frequency (LF) systems detect the signal from the return stroke of Cloud-to-Ground (CG) light­
ning. The strongest radio emission is generated, when the downward leader meets the upward going
streamers approximately 100 m above the ground.
2 LIGHTNING LOCATION BASICS
6
High frequency (VHF) antennas detect single discharge processes from stepped leader, recoil streamer,
and other processes which comprise the intra-cloud discharges (IC). Hence a more complete picture
of the discharge process is provided. On the other hand the ground striking point has to be determined
by data processing.
2.3.2
Detection Efficiency
Detection efficiency (DE) is defined as the fraction of lightning flashes (or return strokes) which is
detected and located by the sensor or network. The ability to detect a flash depends on the amplitude
of the signal, which must be within the dynamic range of the receiving system. The amplitude of the
lightning signal must be high enough to be discernable from the internal and external noise. On the
other hand receivers may distort signals with too high amplitude.
Since the signal’s amplitude drops with increasing distance from the source the DE is a function of
both location relative to the station and lightning amplitude. With typical frequency distribution of
lightning amplitudes an effective detection range can be calculated for each station. The DE differs
for different lightning types (CG, IC) and polarity.
For a network of receiving stations the DE is also a function of antenna spacing. Taking advantage of
the ambiguity due to multiple receiving stations a constant DE inside the network is aimed. Generally
it drops outside the network polygon with increasing distance. Most modern operational LLS have
DE of 70-90% (see below).
Signals arriving with high event rate (per time interval) can saturate a receiving station. This may
limit the DE during intense thunderstorms which are very close to the receiving station.
2.3.3
Detection Capacity
The location of lightning is limited by the detection capacity of single sensors and the total network.
It is quantified as the number of lightning events which can be detected and located per time unit. This
limit arises from sensor (e.g. re-arm time) or data transmission restrictions. In situations with heavy
storm activity a part of the lightning events cannot be processed by the system and is lost.
2.4 Lightning Location
For the location of lightning the signals detected at more than one station (network) are analysed.
Predominantly two location methods are applied: direction finding and ’time of arrival’.
Time of arrival: The exact time of the signals arrival at the antenna stations is compared for the dif­
ferent locations. The position of the signal source is given by the intersection of the hyperbolas
for equal time differences.
Direction finding: The direction of the incident wave is determined at the antenna stations. The
intersection of these directions gives the source location.
In modern networks a combination of both location methods is used, thus minimizing the geometric
uncertainty in location accuracy.
2 LIGHTNING LOCATION BASICS
7
Figure 2: Location principles: TOA – Time of Arrival (left), DF – Direction Finding (right). (From
MacGorman and Rust (1998))
2.4.1
Time of Arrival
This method is used for 2D and 3D location. More sophisticated systems take into account propagation
effects. A high time synchronisation of the senors is required. This was realized in the past using
highly stabilized oscillators and the synchronisation signal of telecommunication satellites (Loran-C).
All the modern systems employ the GPS signal.
To determine the 3 unknown parameters (time, longitude, latitude) of the signal source, at least 3
independent measurements are necessary. Due to the ambiguity of the intersection of 2 hyperbolas in
the general case 4 measurements of arrival time are needed. The uncertainty in the determination of
the time is in the order of the time uncertainty of the single sensors (∆ti ). The location uncertainty
inside the network is in the order of ∆ti · c. Hence for location uncertainty lower than 1 km a time
resolution higher than 3 µs is necessary. Outside the network polygon the uncertainty becomes large
because the hyperbolas intersect at a small angle.
2.4.2
Direction Finding
The comparison of the phase lag of different oriented or dislocated antennas permits the determination
of direction of the wave. This is realized by various methods of antenna arrays, e.g. crossed loop
antennas in magnetic direction finders. Dipole arrays are used in interferometric antennas.
For 2D location only 2 receiving stations are necessary, for 3D location at least one antenna must
determine the elevation angle. The uncertainty in location determination depends on the angular
difference φi − φj and is largest for differences close to 0◦ or 180◦ . Operational networks consist
therefore of at least 3 stations.
2.4.3
Location Accuracy
The location accuracy is an important characteristic of LLS. It is determined by resolution and accu­
racy of single antennas, location principle and the network geometry, i.e. the spacing and distribution
3 GROUND BASED OPERATIONAL NETWORKS
8
of the single antenna stations. In modern LLS the location accuracy is better than 1 km.
2.5 Noise
Lightning emitted radio signals are superposed by noise originating from natural and anthropogenic
sources. Natural noise sources are remote thunderstorms which lightning signals can propagate over
extraordinary long distances under favourable ionospheric conditions. Another high frequent noise
source is radiation from corona discharges at objects close to the antenna.
Anthropogenic noise sources are communication (TV, radio, cell phone, etc) and control transmis­
sions. Usually these signals are modulated and confined to narrow frequency bands. Additionally
random broadband noise arises from various electric equipment, causing e.g. sparks. This type of
noise is eliminated by waveform and amplitude criteria. Signals not matching a certain wave form
which is typical for lightning are discarded. However, the application of this criterion may reject
some lightning signals as well. The amplitude criterion takes advantage of the fact, that the amplitude
from lightning is usually stronger than artificial noise. Again, the weaker lightning events are lost,
especially intracloud lightning. In strong noise conditions the limited detection capacity of the sensor
or network can cause data loss.
3 Ground Based Operational Networks
Lightning location systems (LLS) consist of 2 components: sensor and network. The sensor detects
the lightning by it’s signal and records relevant parameters (direction angle or arrival time). A com­
bination of several sensors – the network – determines the location using the transmitted parameters.
Larger networks, which have been grown over many years, consist usually of different sensor types
and use different location principles in combination, sometimes with redundancy.
The majority of today’s operational networks originate from the two worldwide leading manufactures
of LLS: 1. GAI (Global Atmospherics Inc., USA) with sensors operating in LF range and 2. Dimen­
sions (France) employing VHF sensors. Since 2002 both companies are incorporated in the Vaisala
Group.
3.1 LF Sensors and Networks: Vaisala-GAI
3.1.1
LF Sensors: ALDF, LPATS, IMPACT
ALDF - Advanced lightning direction finder: Direction finding with magnetic loop antennas.
These sensors were originally used in Lightning Location and Protection (LLP) networks (Krider
et al., 1976).
The sensor consists of 2 vertically oriented crossed magnetic loop antennas for direction determination
and a horizontal electric plate for resolving the ambiguity in direction. The detection bandwidth is
1 kHz to 1 MHz. Only signals with certain electric wave form are recorded, this discriminates electro­
magnetic noise from the lightning signals. The quality of detection of the magnetic signal is sensitive
to distortions of the magnetic field by obstacles and metallic objects. Hence the installation site must
be carefully chosen. The detection range for lightning signals is 200 km and more.
3 GROUND BASED OPERATIONAL NETWORKS
9
The sensor’s output is time, angle, signal amplitude and polarity of the first return stroke, and number
of return strokes in the flash. If two or more sensors report a flash within a time interval less than
20 ms (programmable) the flash location is computed. Due to the applied principle only 2 sensors
are necessary for location. However large errors occur for intersection angles close to 0◦ or 180◦ ,
therefore more sensors are used for optimized calculation.
LPATS - time of arrival: In the Lightning Position And Tracking System (LPATS) (Bent and
Lyons, 1984) the lightning location is determined by the TOA method using the electrical signal.
A simple electric whip antenna is used with a detection band of 2 to 500 kHz. Detection by this
method is less sensitive to environment conditions than the magnetic direction finder. The wave form
of the signal is recorded and the ’arrival time’ is fixed. This arrival time is compared between the
different sites and the impact point is calculated.
The sensor reports for each return stroke: time, location and peak amplitude estimation. Intra-cloud
lightning is also detected, but with much lower detection efficiency. The time between single detec­
tions is 15 ms, more than 50 strokes per second can be discriminated. The maximum sensor detection
range is 200–300 km. For lightning location at least 3, generally 4 sensors are necessary. The net­
work operation depends crucially on exact time synchronisation of the sensors. Since the 1990s this
is accomplished by GPS time stamps.
Latest sensor generation: LPATS IV and IMPACT-ESP (Enhanced Sensitivity & Performance).
IMPACT ESP combines magnetic and electric field measurement in the frequency range 0.4 kHz –
400 kHz. Magnetic field is measured with 2 crossed loop antennas for the determination of the waves
incident direction. A plate antenna detects the electric component of the wave field for the fixation of
the arrival time of the signal. For discrimination between CG and IC a wave form criterion is used.
A crucial element for the quality of the whole network still remains the thoroughly choice of the sensor
location place with respect to noise condition and orography (Diendorfer, personal communication).
3.1.2
NLDN
The world wide largest LLS is the North American Lightning Detection Network (NALDN) (Cum­
mins et al., 1999) created in 1998, which comprises the National Lightning Detection Network
(NLDN) in the United States and the Canadian Lightning Detection Network (CLDN). The NLDN
as of 1991 uses 47 Advanced Lightning Direction Finders (ALDFs) in combination with 59 LPATS
III electric field sensors (Fig. 3). The CLDN uses 26 IMPACT-ES sensors and 55 LPATS-IV sensors.
The total area covered by the NALDN is 20 Million km2 .
The network location range extends a few hundred kilometers into the adjacent Oceans and the Gulf of
Mexico. All sensor data are centrally processed in the Network Control Center (NCC) in Tucson. The
location calculation is optimized employing both direction and arrival time information (Cummins
et al., 1998). The ground stroke lightning data includes information on latitude and longitude, time,
polarity, and amplitude.
Data from the NLDN are used for a long time (Orville and Huffines, 1999). First results from the
combined network NALDN were given e.g. in Orville et al. (2002) or Burrows et al. (2002).
3 GROUND BASED OPERATIONAL NETWORKS
10
Figure 3: Location of the lightning sensors comprising the NLDN. The red triangles are impact sites;
blue circles are TOA sites.
3.1.3
European GAI-networks, EUCLID
EUCLID (EUropean Cooperation for LIghtning Detection) is a compound of national lightning de­
tecting networks with the aim to locate lightning all over the European area.
Since the end 1980’s many LLS were installed in the European countries. These LLS were national
networks and operated by weather services, power utilities, etc. The performance of the network gen­
erally decreases outside the bounding polygon. Since storms can have life times of a few hours, they
travel over several 100 km. It is evident, especially for smaller countries, that storms may develop
outside the country, i.e. the polygon of the national network. A simple patch of the national networks
data is unsatisfactory due to methodical problems with double detection, different location methods
and sensor types. To overcome these problems in 1999 the operators of the European LLS basing on
the GAI-technology created the EUCLID compound. The countries involved are Germany, Austria,
Hungary, Czech Republic, Slovenia, Holland, Belgium, Luxembourg, Italy, Poland, Slovakia, Nor­
way, Finland, Denmark, Sweden and France. The networks from Portugal and Greece are being joint
recently. In the near future Spain will be added (S. Thern, personal communication). Each network
retains its national independence and provides lightning data for individual user applications in their
own country as well. EUCLID covers almost the whole Europe (except Russia) and the adjacent
oceanic areas: North Sea, Mediterranean Sea, Biscaya. At the moment the complete network consists
of 75 sensors of 4 different types (LPATS3, IMPACT, LPATS-IV, IMPACT-ESP). Figure 4 shows the
countries covered by the EUCLID compound.
The raw data from every sensor station are transmitted to the data centres in Karlsruhe (Germany)
and Vienna (Austria) where the data are archived and processed for location calculation. This central
processing allows an optimized calculation of the location points.
The detection efficiency of the network is above 90% for CG-lightning with amplitudes >5 kA. Lo­
cation accuracy is < 1 km. IC-lightning is detected only partially with a low detection efficiency.
With EUCLID a significant progress has been made towards a uniform homogeneous LLS. This is
achieved by the central processing of the raw data and also by a step-by-step addition of new sensors at
crucial locations which improve the performance of the total network. The benefit from this compound
is evident also inside the national networks: the lightning location at the country’s border is supported
by the adjacent sensors. Hence the detection efficiency and location accuracy remain more uniform.
3 GROUND BASED OPERATIONAL NETWORKS
11
Figure 4: Sensor locations of the EUCLID compound as of 2000. Open symbols represent IMPACT
sensors, darker symbols represent LPATS sensors.
There are several other LLS compounds in Europe, mostly between neighbouring countries (BLIDS,
CELDN, BLDN) which are partly included in EUCLID. E.g. a Mediterranean network involves the
French, Italian, and Spanish networks employing GAI sensors (Montariol, 2000). This is advanta­
geous for storm observation in the coastal areas of these countries and even more over Mediterranean
Sea, which is poorly covered by each single network.
3.2 VHF Sensors and Networks: Vaisala-Dimensions
3.2.1
SAFIR Sensor
The SAFIR (Surveillance et Alerte Foudre par Interférométrie Radioélectrique) lightning location sys­
tem was originally developed by a research group of the ONERA, the French space research institution
(Richard and Auffray, 1985).
The SAFIR detection station operates at a detection frequency in VHF in the worldwide protected
frequency band at 110 – 118 MHz. This high frequency allows to detect signals from the single steps
of the leader process and from recoil streamers of cloud discharges. These processes emit presumably
in the high frequency range. Another consequence of the VHF range is the line-of-sight propagation
of the lightning signals. Thus the earth curvature limits the detection range of the antennas. A ray
arriving at a detection station from 200 km distance was emitted at least in an altitude of 3 km. The
antenna represents an array of 5 electric dipoles comprising an interferometer.
Additionally to the VHF registration a second antenna at LF (300 Hz – 3 MHz) is used for the detection
of return stroke signals. The combination of both antennas enables the discrimination between ICand CG-lightning. All detected and located VHF events, which meet certain distance-time criteria are
combined into an IC or CG flash.
3 GROUND BASED OPERATIONAL NETWORKS
12
SAFIR networks locate lightning by direction finding. The mean distance of antennas in SAFIR
networks is at about 150-180 km. With the angular resolution of 0.25◦ this yields a location accuracy
better than 1 km. The detection efficiency is about 90% for IC and CG for distances smaller than about
150 km.
The detection capacity is limited currently to 100 events per second. Since SAFIR detects many
signals from a single lightning channel this limit can be reached easily on intense storms. This low
value was chosen to meet the slow data transmission lines. It is expected that this limit will be moved
in sensor upgrades (P. Richard, personal communication).
The SAFIR LLS is less affected to atmospheric propagation conditions. On the other hand any obsta­
cle in the line of sight prevents signal detection from the corresponding sector. This imposes additional
installation efforts for SAFIR sensors in mountainous areas.
3.2.2
SAFIR Networks
SAFIR lightning location networks are installed now in 7 European countries (Netherlands, Belgium,
France, Germany, Hungary, Slovakia, Poland) as schematic displayed on Fig. 5. Several countries with
SAFIR networks are cooperating in exchange of raw data for central data processing. Other SAFIR
systems operate in Asia: e.g. in Singapore and in Japan, the latter consisting of over 30 stations. The
launch place of the European Space Agency (ESA) in Kourou, French Guyana is equipped with a
SAFIR network (Molinié and Pontikis, 1995).
Figure 5: Area covered by SAFIR networks in Europe (left) and the sensor locations of the British
long range network ATD (right).
3.3 British VLF ATD
The UK Met. Office operates a LLS which detects lightning in the VLF range and utilizes an ’Arrival
Time Difference’ (ATD) method for lightning impact location (Lee, 1986). It was recently upgraded
significantly as described by Nash et al. (2000). Currently it consists of 7 stations: 5 in the range of
3 GROUND BASED OPERATIONAL NETWORKS
13
the British Islands, completed by stations in Cyprus and Gibraltar (Fig. 5). The detection frequency
is 10 kHz. The resulting long detection range enables the system to detect and locate lightning surely
over the whole Europe, the Northern and the Mediterranean Sea. As the developers report lightning
is also detected from thunderstorms at large distances in South America, Africa and eastern Russia.
The location accuracy is best inside UK with 1-2 km, in the bounds of Europe it reaches 5 km. The lo­
cation accuracy reduces with increasing distance from the network. For South American lightning the
location error may reach 100 km. The strong variations in propagation parameters such as ionospheric
height and ocean and soil conduction properties cause strong errors in location for the distant flashes.
An improved wave propagation model will be utilized to reduce these sources of uncertainty. A further
consequence of the changes in propagation conditions is a large variation in DE with daytime.
The upgrade of the system also increased the detection capacity of the sensors from former 400 per
hour to present 4000 per hour and is expected to increase up to 10,000 after the upgrade finished.
Additional lightning parameters such as flash strength, polarity, type (cloud to ground strike,or cloud
to cloud) and multiplicity (number of strokes per flash) will be provided.
Compared with the LPATS/IMPACT type systems the ATD detects only about 5% (Finke, 2000), i.e.
only the return strokes with the strongest amplitudes.
3.4 Operational Networks Worldwide
Lightning location systems are in operation worldwide in most of the industrialized countries. De­
pending on the local situation the ownership of the network and the distribution of the data varies from
country to country. Typically, the networks are operated by national meteorological services, power
utility companies, environmental services, flight security agencies, aerospace, or military.
Table 1 summarizes the basic parameters for the operational network types.
Sensor type
SAFIR
GAI
ATD
location frequency
location method
detection range
detection efficiency
location accuracy in Europe
detected lightning type
VHF
interferometry, DF
150-200km
90%
< 1 km
total lightning (CG+IC)
LF
DF and TOA
200-300km
90%
< 1 km
CG lightning
VLF
TOA
500-1000km
10%
< 5 km
CG lightning
Table 1: Parameters of operational lightning location networks.
The worldwide coverage with LLS is confined however mostly to the industrialized countries. Partic­
ularly in the tropics where most of the thunderstorms take place lightning location data are missing.
Also the oceanic areas are not covered by LLS except of the first few hundred km from the coasts. Re­
cently there are activities by GAI to create a Global Lightning Detection Network which integrates the
worldwide operating LLS and employs the long range detection capabilities (Cramer and Cummins,
1999) to cover the Oceans.
It follows an incomplete list of the countries with operational LLS. As mentioned earlier (Sec. 3.1),
the majority of LLS is equipped with either VHF SAFIR sensors or LF sensors from GAI. In some
countries networks of both types work in the same area.
3 GROUND BASED OPERATIONAL NETWORKS
Europe:
Africa:
North America:
South America:
Asia:
Australia:
14
France, Germany, Austria, Switzerland, Great Britain, Sweden, Norway, Den­
mark, Finland, Iceland, Poland, Hungary, Slovakia, Czechia, Belgium, Nether­
lands, Luxembourg, Spain, Portugal, Italy, Slovenia, Croatia.
South Africa, Tunisia, Morocco
USA (NLDN) and Canada (CLDN)
Brazil (Beneti and Sato, 2000), Colombia (Torres et al., 1996), Costa Rica,
Venezuela
Japan (Ishii et al., 2000), South Korea (Woo et al., 2000), Indonesia (Laksmi­
wati, 1998), Thailand, Singapore, Hong Kong
Australia Sharp (1999), New Zealand (Pannett, 2000)
3.5 Long Range Lightning Location
Recently attempts are made to extend the detection range of LF location systems into the oceans areas
by a method which employs the signal from waves reflected at the ionospheric rather than the ground
wave in normal processing. A test configuration with the NLDN sensors was used by Cramer and
Cummins (1999) to estimate the data quality parameters of such a long range detection. The authors
found a detection efficiency of 10-25% during the night and a few percent during the day. For lightning
with amplitude > 45 kA the DE is about 50%. This strong variation with the time of the day is caused
by the diurnal changes in the ionosphere structure. To reduce these variations it is proposed to place
more sensors at larger latitudinal distances. The location accuracy is approximately 5-10 km.
Several methods for long range detection and location are currently investigated and operate in several
research organisations.
3.5.1
VLF Arrays
The Los Alamos Sferics Array (LASA) was installed in 1998 in support of the FORTE satellite mis­
sion (sec. 4.2). It represents an array of now 11 electric field changes sensors located in 3 sub-arrays
in stations in New Mexico, Texas, Florida, and Colorado (Massey et al., 1999; Heavner et al., 2000).
The sensors detect in the frequency range 300 Hz to 500 kHz with high time resolution and GPS
time synchronization. The network detects the waveform and locates CG and also IC using a time of
arrival method with a location accuracy at about 2km within 130km of the New Mexico and Florida
sub-array.
A VLF-array operates in the South of Germany (Betz et al., 1996). This network consists of 3 stations
and is used for storm nowcasting (Oettinger et al., 2000).
3.5.2
ELF and Schumann-Resonance Lightning Location
Strong lightning flashes excite ’Schumann resonances’ (SR) in the earth-ionospheric cavity. SR de­
tection stations are located around the world mostly at research institutions (e.g. Lyons et al., 2000;
Boccippio et al., 1998). Using inversion calculations the global detection and location of lightning is
possible (Heckman et al., 1998; Burke and Jones, 1995) with accuracy of a few 100 km (Füllekrug
et al., 1996). The detection efficiency is only a few percent compared with the operational LLS. In
3 GROUND BASED OPERATIONAL NETWORKS
15
view of the low DE and the large location uncertainty the use for operational purposes is limited. How­
ever the SR location method may be a very useful instrument for global and climatological studies
(Williams, 1992).
3.6 Data Characteristics
Lightning data are available in near real-time with delays less than a few minutes. This is sufficient
with regard to typical evolution times of storm cells. The data contain time, spatial location in geo­
graphical coordinates, peak current amplitude, polarity and flash multiplicity (number of single return
strokes per flash).
Lightning location data are usually represented as points on geographical maps with colour-coded
time information (Fig. 6). Averaging and integrating produces lightning densities for certain time
intervals and time series of lightning totals for some areas. The point structure of the data makes it
easy to overlay lightning data on other geographically distributed data, e.g. from radar and satellites.
The IC lightning data detected by the SAFIR networks contain for every lightning flash many single
point locations. These are assembled by data processing algorithms into complete lightning paths.
3.6.1
Quality Parameters
Important quality parameters of the networks are: detection efficiency, false alarm rate; the quality
of location determination and amplitude estimation. Another characterising network property is the
ability to detect and to discriminate between different lightning types, i.e. CG lightning or total
lightning (CG and IC).
Detection The detection efficiency in modern networks is 90% but drops down with increasing
distance from the network. For instance, at a 200 km distance outside of the NLDN the DE is less
than 40% (e.g. Nierow and Showalter, 2000).
Consistency tests and comparison with cloud data show, that the rate of false alarms is very low
(less than 10 per day (Diendorfer, personal communication)). In some special constellations bogus
mapping of flashes to far locations is possible. These cases are however easily recognized by the
forecaster and can be eliminated by special filters.
Accuracy Comparisons with ground truth observations (see below) showed that in thoroughly in­
stalled networks the location accuracy is better than 1 km.
CG and IC lightning For distinction between CG and IC a wave form criteria is used in LF systems.
In the past this method classified some of the IC flashes wrongly as positive flashes (Cummins, 1996
personal communication). In SAFIR systems the signal from the additional LF antenna is used for
discrimination of the ground flashes.
3 GROUND BASED OPERATIONAL NETWORKS
16
Figure 6: Example plot of lightning location data over Europe 04.08.2001 (EUCLID)
Flash multiplicity For organising the single detected return strokes into flashes, usually a distance
and time criterion is applied. Return strokes which follow a short time (1 s) after the first stroke
and are located closer than 10 km are assigned to one flash. Finke (1999) suggested from a correlation
analysis of LPATS lightning data distance interval of 5 km and a time interval of 1 s for the subsequent
return strokes.
3.6.2
Validation Studies
Much validation efforts have been made by operators and users of the ground based LLS (Diendorfer
et al., 1994). Validation studies use for ground truth impacts of lightning in towers (Hussein et al.,
1995; Heidler et al., 2001) or optical video observations of lightning (Idone et al., 1998a,b).
Recently Diendorfer et al. (2002) presented a validation study comparing lightning detection of
the Austrian ALDIS with tower impacts. Additionally high-speed video observations were made.
Analysing 463 strokes with peak amplitudes > 2 kA the authors derived a detection efficiency of 94%
3 GROUND BASED OPERATIONAL NETWORKS
17
and a average location accuracy of about 500 m. Comparing the peak current estimate with direct
tower measurements a linear regression (IALDIS = 0.95 · IT ower ) was established with a correlation
coefficient of 0.95. These findings demonstrate the high quality of today’s ground based LLS.
3.6.3
Network Intercomparisons
For areas which are covered by 2 networks the intercomparison of the data is possible. The data are
compared with respect to the lightning distribution statistics and coincident events detected by both
system. In Europe (Ruffieux and Rast, 2001) compared the data from 2 LLS (BLIDS and Meteorage)
covering the area of Switzerland with lightning counter detections. Tzanos and Senesi (1998) pre­
sented a comparison of the Meteorage-network (IMPACT sensors) and a SAFIR-system around Paris
in France.
For an 8-week period during the field experiment EULINOX Finke (2000) compared the data from
LPATS and ATD over the range of Germany. He established detection efficiency for the ATD of only
about 5-10%. A i.e. only the return strokes with the strongest amplitudes were detected. A high
percentage of coincident events was found, 84% of all ATD flashes could be identified in the LPATS
data. Thery (2000) compared the LPATS data for the same time period against observations with a
3D SAFIR type system. She found a slight overestimation of CG flashes by LPATS due to wrongly
identified cloud flashes.
3.7 Lightning Products and Services
Various products and services are offered by the network operators: full access to lightning location
data, lightning distribution maps, climatologies for sub-areas. A broad spectrum of warning services
are available via different communication lines.
Online lightning information with reduced resolution in time and space is provided free of charge
from the various LLS operators:
EUCLID compound
BLIDS Germany+Switzerland
Meteorage in France
Vaisala GAI
http://www.euclid.org
http://onweb.blids.de
http://www.meteorage.fr
http://www.lightningstorm.com
3.8 Network Summary
Ground based LLS operate in the most high industrialized countries now. Europe, North-America and
Australia are continuously covered by LLS. In many countries operate more than one LLS. E.g. in
Europe and Japan operate SAFIR and GAI-networks simultaneously. Additionally, Europe is covered
by the long range British ATD system. Currently the national networks are combined into large
compounds with central data processing. Outside Europe and North-America the LLS operate mostly
isolated in the national bounds.
Newly installed and upgraded networks working with the latest sensor generation are able to detect
approximately 90% of all CG flashes with amplitudes larger than 5 kA. The location accuracy in these
networks is normally better than 1 km. The LF networks (from Vaisala-GAI) are designed to locate
4 SATELLITE BASED LIGHTNING LOCATION
18
only the CG flashes, while the total lightning information is provided by the SAFIR systems (VaisalaDimensions).
Lightning information from ground based LLS is generally missing over the oceanic areas far from
the continental coasts. Large continental areas without LLS are met in Africa, many Asiatic countries
and also in Russia. A continuous coverage of these regions with the existing technology networks
cannot be expected for the near future due to the missing technical and financial preconditions in
these countries.
Long range detection methods are being currently developed e.g. by special processing of the LF sen­
sor data. This may extend the detection range to over 1000 km, however with low detection efficiency,
a location accuracy of about 5 km and strong variations with daytime.
A further improvement in location accuracy and amplitude estimation can be expected from the im­
plementation of wave propagation models which take into account orography and different ground
conductivity.
In contrast to the majority of other meteorological information, LLS are owned and operated in many
cases by private, non-governmental companies. This implies great differences in the data distribution
policy. The complete lightning location data are restricted in access and have to be paid. Reduced
information however, is made available to the public via internet.
4 Satellite Based Lightning Location
Satellite based lightning detection and location methods are not operationally implemented, but a few
platforms are in work now since the middle of the 1990s. These include the NASA development
of optical lightning sensor arrays (OTD, LIS) and the multisensor (RF and optical) satellite FORTE
launched by the Los Alamos National Laboratory.
4.1 Optical Location: OTD and LIS
First satellite based lightning location data were derived from the Operational Linescan System (OLS)
on board of the sun-synchronous U.S. Air Force DMSP (Defense Meteorological Satellite Program)
(Goodman and Christian, 1993). Observation was limited to night time, the geographical and spatial
coverage was very sparse. The data were processed by NASA MSFC for 13 years between 1973
and 1997. The first dedicated observation of lightning from space in the optical range was started by
NASA with the OTD (1995) and the LIS (1997) sensors (Christian, 1999). It is planned to install a
geostationary lightning mapper (LMS) on GOES (Weber et al., 1998).
4.1.1
Emission and Propagation of Optical Radiation
Lightning emits optical radiation during its hot phase. This thermal radiation is described by Planck’s
law and the total intensity is proportional to the 4th power of the channel’s surface temperature.
The optical radiation is scattered in the atmosphere by gases, aerosols and cloud particles. Most
significant is the multiple scattering by cloud particles. According to Christian et al. (1989) the energy
loss by this process due to absorption is low. Most of the optical energy is conserved and the part
4 SATELLITE BASED LIGHTNING LOCATION
19
scattered into the upper hemisphere can be detected from a satellite based sensor. More recently Light
et al. (2001) simulated the optical signal in satellite based lightning detection.
The optical arrays are sensitive to high energetic particles (e.g. protons) which cause non-lightning
detection and can saturate the sensor. This hampers the lightning detection in regions like the South
Atlantic Anomaly (SAA) with unusually high proton levels caused by a minimum of the geomagnetic
field.
4.1.2
OTD Sensor
The Optical Transients Detector (OTD) was developed by NASA and was launched into space in
1995. It is carried by the Microlab satellite, which is on a 70◦ sun-synchronous orbit at an altitude of
740 km above the ground.
The optical radiation of lightning is detected with a 128 × 128 pixel CCD matrix. The field of view
projected onto the earth surface is 1300 km × 1300 km. The radiation is passed through a narrow
interference filter which is centred at the wavelength of ionized Oxygen at 777.4 nm.
Special data processing (the Real Time Event Processor) was developed in order to address the specific
problems such as short time of the flash, daylight conditions, optical noise form many other sources.
An instrument description is given e.g. by Christian et al. (1996). Figur 7 shows an example plot of
lightning detected by OTD.
Figure 7: OTD lightning observation on 21 July 1998 during all ascending passes. (From NASA
webpage)
4 SATELLITE BASED LIGHTNING LOCATION
4.1.3
20
LIS Sensor
The Lightning Imaging Sensor (LIS) was launched on board of the Tropical Rainfall Measuring Mis­
sion (TRMM) satellite in 1997 (Kummerow et al., 1998). According to the objectives of the TRMM
the satellite’s orbit limited the observation area to the latitudes ±35◦ . The construction of the LIS in­
strument resembles the OTD with several improvements in sensitivity, design and electronics (Chris­
tian, 1999).
4.1.4
Data Characteristics
The data sets from both sensors along with the data analysis software library are freely available for
the scientific community. The data include for every detected event location data and radiation energy.
This is completed by ancillary information about the orbit, observation conditions, noise, instrument
status and also the background images.
The orbital parameters of both sensors are listed in table 2.
Sensor
OTD
LIS
Orbit
Field of View
Spatial resolution
Observation time for a fixed
ground location during overpass
70◦ ,
35◦ , 350km
550 km × 550 km
5 km
90 sec
740km
1300 km × 1300 km
10 km
1-270 sec
Table 2: Orbital parameters of OTD and LIS sensors
The sophisticated primary data processing performs a hierarchical clustering of data into events,
groups, flashes and areas.
OTD and LIS detect both CG and IC flashes, i.e. total lightning. However, basing on the optical data
alone, it is difficult to discriminate between both types, i.e. to discern the ground strikes.
The big advantage of the OTD and LIS data is the global homogeneity of the observation. Some
external influences (sun glints, SAA) depend however on satellite position or viewing angle.
The short observation time for fixed locations at the ground during the overpasses limit presently the
use of these data in forecast and nowcasting. The main applications are in climatology, comparing
studies and single case studies on storm cloud scale. It was a major aim of NASA to perform inten­
sive tests and validations on these sensors data to develop various algorithms and methods for future
sensors in space.
Data from OTD and LIS have been investigated by many scientific groups worldwide. Comparative
analyses have been performed with ground based data from operational LLS (NLDN) and total light­
ning systems (e.g. LDAR). The detection accuracy was found to be within 10 km (OTD) and 3-6 km
(LIS). Due to non-perfect geolocation of the satellite the location data may be biased by few km. The
detection efficiency for CG lightning is about 45-70%, slightly higher for IC lightning (Boccippio
et al., 2000).
4 SATELLITE BASED LIGHTNING LOCATION
4.1.5
21
Scientific Results and Applications
The observations with the both sensors made it first possible to observe lightning continuously on the
whole globe. Since OTD operated almost continuously from 1995 until 1999 the data set is suited
for a 5-year global climatology. A lot of scientific facts about global lightning distribution could
be established firstly or have been proved (Boccippio et al., 2001). This includes the land-ocean
distribution, latitudinal variations, and the global lightning flash rate (Christian et al., 1999).
The global OTD data set found use in other application, e.g. for the parameterisation of lightning
produced NOx in Chemical Transport Models (Allen and Pickering, 2002).
The LIS being an integral part of TRMM made it possible to interrelate the electric discharge activity
in clouds to their particle content as observed by the radiometers and radars. This also enables the
investigation of cloud electrification. Figur 8 displays a 3 month lightning summary from LIS.
Figure 8: LIS lightning observation summary for 3 months: Dec 1997 - Feb 1998. (From NASA
webpage)
Basing on the experience gathered with the both sensors NASA is prepared to the launch of the
Lightning Mapping Sensor (LMS), which will be on a geostationary orbit.
4.2 Combined Radio-Optical Observation: FORTE
4.2.1
FORTE Sensors
The Fast On-Orbit Recording of Transient Events (FORTE) satellite is a joint Los Alamos National
Laboratory (LANL) and Sandia National Laboratories project. The satellite was launched in 1997 in
a circular, 825-km-altitude orbit with 70◦ inclination. It was designed to test technologies to remotely
monitor compliance with arms control treaties (nonproliferation) and for lightning and ionospheric
research.
4 SATELLITE BASED LIGHTNING LOCATION
22
The satellite carries 3 sensors:
Broadband VHF receivers with a total frequency range of 26 – 300 MHz. Two independent re­
ceivers both with 22 MHz bandwidth can be used. For lightning detection usually a lower band
(26–48MHz) and upper band (118–140 MHz) is employed. The receivers record at high sam­
pling rate (50 MHz). The used 35 ft antenna has an aperture of 80◦ .
Broadband photodiode (PDD) in the visible – near infrared wavelength interval (400–1100 nm) for
the record of the optical waveform during 2 ms with 15 µs resolution. The viewing angle is
approximately 80◦ , which corresponds to an area of 1200 km diameter on earth surface.
Narrow band CCD-array of 128×128 pixels – the Lightning Location System (LLS) with an ob­
serving wavelength of 777.6 nm. The pixel size projection gives a 10 km×10 km footprint on
the earth.
The payload was described e.g. Massey et al. (1998). The optical lightning sensor (OLS) is based
on the design of NASA’s detectors OTD and LIS described above with electronics developed by the
Sandia National Laboratories.
FORTE’s instruments detect, record, and analyze transient radio frequency and optical signals that
arise from near the Earth’s surface. It can operate in various modes, RF ranges and trigger conditions.
The geometrical orbit parameters are nearly the same as for the OTD sensor on the microlab satellite.
4.2.2
Scientific Results and Applications
The FORTE lightning data have been compared against various ground based LLS: with the LASA
VLF-array (Massey et al., 1999), the NLDN (Jacobson et al., 2000) and the ATD (Jacobson and
Williams, 2001).
New results were gained in lightning research investigating strong IC discharges, which are detected
from space such as trans-ionic pulse pairs (TIPP) (Holden et al., 1995). It was established that these
events originate from narrow-bipolar events within the clouds: the compact intracloud discharges
(CID) which emit fast and isolated bipolar electric field change pulses (Smith et al., 1999).
FORTE has provided a better understanding of the relationship between optical and RF lightning
events. There exists a strong correlation of optical and RF power for CG events, almost proportional­
ity. The combined analysis of the RF and the optical signal enables the identification and distinguish­
ing between CG and IC lightning and also leader processes (Suszcynsky et al., 2000).
It is proposed to provide global lightning and severe storm monitoring information which would be
continuously downlinked from the GPS Block IIF satellites (Suszcynsky et al., 2000).
4.3 VHF-Sensor: ORAGES
ORAGES is a project of a space-born VHF lightning locator which will localize intra-cloud and cloud­
to-ground lightning flashes. Basing on the SAFIR antenna technology lightning will be localized by
interferometry at a receiving frequency of about 200 MHz (Bondiou-Clergerie et al., 1999). The
sensor is planned to be launched in 2004 onto a low orbit (700 km altitude) with an inclination of
13◦ , which allows an intense observation of the tropics. The antenna system will be experimented
on a stratospheric balloon flight. The project is directed by the French aerospace research institution
ONERA in scientific collaboration with the Laboratoire d’Aerologie (Toulouse).
5 USE OF LIGHTNING LOCATION DATA
23
4.4 Satellite Summary
Detection and location of lightning from satellite based sensors have been developed to high quality
during the last 5 years. Optical location sensors (OTD, LIS and OLS) detect total lightning (CG and
IC) with a location accuracy of approximately 5-10 km. With these sensors for the first time lightning
can be observed globally. The data were validated against ground based LLS, mainly for the USA,
but also outside the US.
The additional VHF broadband sensor on the multi-sensor platform FORTE opens new possibilities
in lightning observation. A combined analysis of VHF and optical data permits the discrimination
between the lightning types.
The data from both systems are available for the scientific community and are intensively explored
currently.
The platforms are intermediate steps towards continuous global observation of lightning from geo­
stationary platform (NASA-MSFC) or a family of low orbit satellites (LANL). A lightning mapper
(LMS) is scheduled to be launched on GOES-East and GOES-West for continuous lightning observa­
tion over America.
5 Use of Lightning Location Data
Lightning location data are widely used now in research and in various application fields. A part of
applications make operational use of the real-time data for all kind of warning purposes (weather ser­
vices, power plants, aviation), while other applications use the archived data (insurances, climatolo­
gies, research). This section gives an overview of typical applications of lightning data in atmospheric
research, meteorological services and other applications.
5.1 Comparative Studies
Lightning data from all the described LLS have been analysed in various fields of thunderstorm re­
search on the various scales in combination with other cloud data from radar and satellites.
5.1.1
Radar and Satellite
Identification and characterisation of thunderstorms is accomplished by cloud satellites and radars.
Cloud observing (IR) satellites provide images of the cloud top temperature and height with spatial
resolution of 5 km and repetition rate of currently 30 min. Precipitation and cloud radars use the
reflectivity information for determination of rain rate or in terms of particle density and size. Spatial
resolution ranges form 100 m – 2 km, the sampling rate for operational radars is about 10-20 min.
Lightning data add value to this information by indicating the most active parts of the storm clouds.
Thus lightning data help to narrow the potentially hazardous regions. Another advantage is the real
time availability of lightning information. Thereby lightning can bridge the time between successive
scans and fill the gaps caused by failures of the scanning systems.
5 USE OF LIGHTNING LOCATION DATA
24
This is valuable in any stage of the storm development. In the growing phase it helps to recognize the
fastest growing cells. During the mature phase the storm cloud is large and not homogeneous. Light­
ning marks the most active part, when satellite images show largely extended anvil clouds. Light­
ning data help to distinguish active cells from dissipating regions in clouds with uniform high radar
reflectivity. Furthermore, the development tendency - intensification or ceasing - goes along with
characteristic changes in the total lightning activity. Lightning also identifies embedded convection in
stratiform clouds, which is sometimes difficult to accomplish with radar alone.
Various quantitative studies have been performed which relate lightning rate and density to cloud
parameters. We give here a short summary after MacGorman and Rust (1998):
• CG lightning occurs when the cloud’s reflectivity at a height corresponding to −10◦ C exceeds
35-40 dBZ and the cloud top extends at least to 9 km (corresponding cloud top temperature of
-40◦ C).
• Lightning is located closely to the reflectivity maximum, but more frequent at the border of this
area.
• High IC lightning activity is correlated to the volume of large reflectivities cloud regions.
5.1.2
Lightning and Precipitation Data
Lightning activity and precipitation amount of a storm cloud is in close relation. There is a large num­
ber of publications on this topic (Goodman and Raghavan, 1993; Richard and Lojou, 1996; Cheze and
Sauvageot, 1997; Areitio et al., 1999; Anderson, 2000; Soriano et al., 2001; Soula and Chauzy, 2001).
Petersen and Rutledge (1998) compared the data for different latitudes and geographical locations.
Generally a good correlation is found between lightning density and rain amount. Some authors
present empirical regressions functions (mostly potential law): the rain-yield per flash. Remarkable
is the large scatter in these relations for different storms and locations, which limits the practical use
of these regressions. This suggests to include in the analysis additional atmospheric data such as
moisture content, precipitable water. The storm type must be considered as well as synoptic scale
lifting and orographic influences.
A high lightning rate is always a reliable indicator for heavy precipitation. In the absence of radar or
in mountainous regions, where radar coverage is obstructed, lightning data can be a useful indicator
for flash floods.
5.1.3
Lightning and Other Storm Characteristics
The relation of lightning activity and other storm related phenomena have been studied. This includes
hail (MacGorman and Burgess, 1994; Hohl and Schiesser, 2001), tornados (Buechler et al., 2000;
McCaul et al., 2002), gust and microburst (Laroche et al., 1991) and flush floods (Petersen et al.,
1999). These case studies established characteristic signatures in the flash rate time series preceding
the occurrence of e.g. hail or tornados.
Positive lightning is less frequent (<10%) in the average but occurs to high percentage in winter
storms (Shimura et al., 1999) and at the end of storms (MacGorman and Rust, 1998).
5 USE OF LIGHTNING LOCATION DATA
25
Lightning activity also varies with the dynamic organisation of the storm. Supercell storms were
observed with an unusual large percentage of IC lightning (Ray et al., 1987). High lightning densities
are observed in squall-lines and in Mesoscale Convective Systems (MacGorman and Morgenstern,
1998).
5.2 Climatologies
Various statistical analyses of lightning data have been published mostly for the national networks.
(e.g. Orville and Huffines, 1999; Finke and Hauf, 1996; Tuomi, 1996; Hidayat and Ishii, 1998). Light­
ning climatologies show the geographical distribution of the lightning density, time series of lightning
totals for various regions, such as annual and diurnal cycles, frequency distribution of number of
lightning impacts and other parameters.
A common feature of these climatologies particularly for the middle latitudes is the high variability of
the annual mean values. The reason is the concentration of most of the lightning events on few days in
certain thunderstorms. This ’fractal’ property of the lightning distribution requires a long observation
time in order to get statistically significant mean values.
The network’s DE is not uniform in space and changes with time (with various improvements of
sensors and network). Statistics from different networks can hardly be compared due to the many dif­
ferences between them. With the tendency towards network integration a more homogeneous spatial
DE will be achieved and climatologies of larger regions will become available.
The relation between lightning density and the synoptic observations of thunderstorm days has been
investigated (Changnon, 1989, 1993; Finke and Hauf, 1996). A good correlation has been found, in
general with deviations due to varying acoustic conditions at the synoptic stations.
Another application is the use of lightning data for derivation of storm parameters. Hagen et al. (1999)
applied a semi-automatic method for a statistical analysis of storm motion parameters. An automated
tracking of lightning and radar data was presented by Steinacker et al. (2000).
Global distribution of lightning can now be calculated from the satellite (OTD) data (Christian et al.,
1999). It turned out, that most of the lightning activity is concentrated over the tropical land masses
with an expressed diurnal cycle, while lightning over the oceans is more constant in time. The mean
global flash rate was established to amount 37 flashes per second.
Lightning statistics are the base for lightning risk assessment. Changes in lightning distribution may
indicate regional and global manifestations of global climate change (Price and Rind, 1994). This
suggests the use of lightning statistics for climate monitoring.
5.3 Weather Forecast and Nowcasting
The value of lightning data for nowcasting of severe convection is evident and methods for integration
of lightning data in various algorithms have been developed (e.g. Goodman, 1991; Morel and Senesi,
2000).
Lightning data report uniquely the electric activity in clouds and lightning strike hazard. Combined
with other cloud observations from radar and satellite it adds value by indicating the most active parts
in the cloud.
5 USE OF LIGHTNING LOCATION DATA
26
Lightning data are easy to display and to interpret. The sequence of lightning location points gives
the forecaster a fast overview of the development of convective activity in the whole region. The data
arrive practically without delay, hence lightning information fills the gaps between sampling times of
other instruments.
The synergistic effects of combining lightning, radar and satellite data support the early detection of
convection, the detection of intensification or ceasing, and a qualified tracking and motion forecasting.
Forecast of convective storms can benefit from additional information provided by the IC lightning
data. Total lightning gives a more complete picture of the electrical activity. It requires however more
interpretation effort and the development of proper display techniques which reduce the amount of
data to useful information.
In the COST Action 78 (COST = European Co-operation in the Field of Scientific and Technical
Research) a special task was dedicated to the use lightning data for early detection, nowcasting and
classification of thunderstorms (COST-78, 2001).
An algorithm for tracking and nowcasting of motion and evolution of storms on various scale was
recently developed by Morel and Senesi (2000). The authors use satellite (Meteosat) data together
with lightning data from a ground based LLS for identification of storms. Each identified storm is
treated as an ’object’ with the properties: size parameters, motion vectors, cloud top temperature
structure and lightning content.
In Jacobs and Raatz (2001) numerical model output is combined with observed data from radar, satel­
lite and lightning detection system for a pixel-wise nowcasting for 3 hours in advance. Knüpffer
(2001) presented a method for predicting lightning with a MOS (Model Output Statistics)-schema.
The input data are observed lightning data and forecast from the numerical model GMS of the Ger­
man Weather Service.
5.4 Total Lightning
Total lightning includes the Cloud-to-Ground flashes and the intracloud lightning. Generally the IC
discharges precede the CG lightning in the life cycle of convective clouds. It was shown in compar­
ative studies that the time difference between the first detection of IC lightning and the first CG flash
varies strongly (Tzanos and Senesi, 1998). For fast developing clouds this time may amount only a
few minutes (Richard and Kononov, 2001), while other clouds exist long time without producing any
CG flash. Among the currently employed operational LLS only SAFIR detects IC lightning.
For many applications the total lightning information is of significance, since total lightning gives
a more complete information about the electric activity in the storm. A cloud with IC lightning is
potentially dangerous, because lightning can be triggered by objects close enough. Furthermore, the
cloud may intensify and produce after some time CG lightning. Residual clouds left from convective
systems may be electrified and represent a hazard for aviation.
For practical applications it is important to reduce the IC lightning information into proper products
which can be easily used by the end users (Richard and Kononov, 2001). The usefulness of total
lightning information for aviation was demonstrated with the LISDAD product (see sec. 5.6.1)
5 USE OF LIGHTNING LOCATION DATA
27
5.5 Atmospheric Chemistry
In the hot lightning channel at temperatures above 3000-4000 K nitrogen oxide NO is created in
high concentration (Chameides et al., 1977). Due to the fast cooling of the lightning channel this
high concentration remains ’frozen’ after the return to ambient temperatures (Zeldovich and Raizer,
1966; Hill, 1979). The created NO is in fast establishing photochemical equilibrium with NO2 under
tropospheric conditions. The sum of both gases is named NOx. This NOx is partially transported,
together with NOx from other sources, by the strong updrafts in thunderstorms up to the tropopause
level (Huntrieser et al., 1998). The contribution of lightning produced NOx is currently investigated
both on the cloud scale of single thunderstorms (Ridley et al., 1996) as well as on regional and global
scale (Lee et al., 1997; Price et al., 1997).
Of crucial importance for these estimations is the knowledge of the distribution of CG and IC light­
ning. Lightning detection systems on the various scales contribute these data. On the cloud scale the
3D lightning path relative to the up- and downdrafts is reconstructed from local VHF system observa­
tions (Defer et al., 2000; Thery, 2001). For regional analyses the operational LLS can be used (Höller
et al., 1999; Finke, 2000).
The absence of lightning data in most parts of the globe, especially over the tropical regions disabled
global studies on lightning NOx production. Only recently, with the OTD and LIS sensors, lightning
data became available for these areas even though discontinuously in time. Nesbitt et al. (2000) used
OTD data for calculation of seasonal global lightning NOx production. In global chemical transport
models the OTD data provide parameterizations for lightning NOx (Allen and Pickering, 2002), while
total NOx columns can be gathered with other satellites (e.g GOME).
5.6 Non – Weather Service Applications
The lightning based forecast and climatology products provided by weather services and LLS opera­
tors are used in many applications. The following section describes the use of lightning location data
in a few application fields where lightning data are of special interest.
5.6.1
Aviation and Flight Security
Lightning is a hazard for airplanes and can cause severe damage of the construction, fuel ignition and
damages of the electronic equipment or harm the crew members. Halsey and Patton (1999) reported
on the lightning hazard for helicopter operations over the Northern Sea. Military flight operations are
coordinated taking into account lightning hazard.
Airplanes normally avoid thunderstorms due to the prevailing strong air motions and icing hazard.
However the lightning hazard in non-thunderstorm clouds is often underestimated and lightning strikes
are triggered by the aircraft itself. Hence for aviation applications the information about total lightning
(IC and CG lightning) is of importance.
In the LISDAD (Lightning Imaging Sensor Data Demonstration) project total lightning information
from NASA’s (LDAR) is merged with CG lightning data from the NLDN and with WSR-88D radar
data. The US National Weather Service Office at Melbourne (Florida) uses LISDAD for storm nowcasting and warning products for the Florida airports (Boldi et al., 1998).
5 USE OF LIGHTNING LOCATION DATA
28
Airport management can benefit from real time lightning data in decision finding on closing and re­
opening the flight operations. Autones et al. (1999) presented an algorithm for support of the routing
of air traffic in case of thunderstorms in the approach area which merges lightning and radar data in
the surrounding of the Paris airport.
Nierow et al. (1999) investigated the benefit from the implementation of total lightning data for con­
vective weather forecast. The Federal Aviation Administration uses lightning data for improved rout­
ing of airplanes (Nierow and Showalter, 2000). Due to the lack of lightning location data over the
Oceans the experimental long range lightning detection network (Cramer and Cummins, 1999) was
employed for this purpose. The (Aviation Weather Center) AWC is using these oceanic lightning
data as input to their operational oceanic weather hazard alerts referred to as International SIGMETs
(which depict significant meteorology).
5.6.2
Aerospace
Apollo-12 and Atlas/Centaur-67 were strucked by triggered lightning during launch (Christian et al.,
1989). To avoid these incidents lightning forecasts and flight rules are given for scheduled missile
and space shuttle launches (Garner and Oram, 2000). Primary inputs are cloud observations with
radar and satellite as well as lightning observation from the NLDN and the NASA operated local high
resolution LLS (LDAR, (Maier et al., 1996)).
The European Space Agency (ESA) operates a SAFIR LLS at the European Space Center CSG in
Kourou, French Guyana.
5.6.3
Power Utilities
Lightning is a major cause of transmission and distribution faults and outages. Power utility operators
use the real time lightning data for the network management to avoid electricity disruption and for
dispatching the repair crews. Lightning data also aid in locating and identifying the cause of faults
and damages of transmission lines (Fister et al., 1992; Kappenman et al., 2001). Another application
is the estimation of outage statistics from lightning distribution data (Diendorfer, 2001).
5.6.4
Fire Weather Warning
Lightning is the main natural cause of forest fires and consequently, one of the first applications of
LLS served the forest fire detection (Krider et al., 1980). LLS can help to identify lightning producing
storms which might cause wildfires. Especially in large unhabitated areas the lightning location data
can narrow the areas to be observed by aircraft.
Bothwell (2000) reported on the use of data from the NLDN in combination with computer models
and remote automatic weather stations for the prediction lightning from dry thunderstorms, that create
50 percent of all wildfires in the U.S. The special role of positive lightning in fire ignition in Canada
was examined by Kochtubajda et al. (2002).
5 USE OF LIGHTNING LOCATION DATA
5.6.5
29
Explosives
Substantial losses result from lightning caused ignition of fuels and explosives (e.g. Zoro et al., 1997).
Well known is the Picatinny Arsenal incident in July 1926, which killed 14 people and cost 70 million
Dollar. More recently in June 2001 at Buryatia, Russia 17 people were killed, 3000 people were
evacuated when lightning hit a munition depot. The munitions losses exceeded 20 million rubles
(Kithil, 2002). Lightning detection systems help to recognize the lightning hazard for warning and
activation of protection measures.
5.6.6
Recreation and Public Event Management
Most lightning casualties happen outdoor and during recreation activities (Holle et al., 2001). Hence,
lightning data can successfully be used for early warning at places with intensive recreation activity
e.g. at lakes, mountains or golf courses. A prominent application was the use of NLDN data in
combination with several other data for short-term forecast for the whole area of the 1996 centennial
Olympic games in Atlanta (Johnson et al., 2000).
5.6.7
Insurance Companies
Lightning damages represent approximately 9% of all insurance claims amounting to several 100
million USD per year (Kithil, 1999, 2000). Insurance companies use lightning data for verification of
the lightning damage claims. Insurances make use of lightning climatologies for lightning and storm
damage risk assessment (Dinnes, 1999). There is an ongoing research in re-insurance companies in
implications of global climate change. An increase in global lightning frequency is expected due
to increase of the global temperature (Price and Rind, 1994). Mills et al. (2001) presented strong
correlation between average temperature and lightning related insurance claims between 1990-95 in
the United States.
5.7 Applications Summary
Lightning data are widely used in real time application for forecast and warning purposes. The inten­
sive research on lightning data resulted in the development of many algorithms for identification and
nowcasting of convective storms motion. Lightning data are a valuable addition to the other cloud
observation tools like radar and satellite. Lightning is a useful proxy for other storm related hazards.
For certain applications (e.g. weather forecast at airports) a higher location accuracy than provided by
the present operational systems is desirable. A higher spatial resolution and accuracy can be achieved
with closer spacing of the receivers. Best suited for lightning location with accuracy in the order of
100 m are VHF systems, which are in use in few local LLS operated by research institutions for cloud
and lightning research.
Many users of lightning data are interested in lightning impacts to the ground. This information is
provided by the LF (GAI) networks with high accuracy and detection efficiency. For applications in
weather forecast (nowcasting) and aviation, as well as for cloud research and atmospheric chemistry
the information about total lightning which includes IC lightning is important. Among the commer­
cially available systems only SAFIR provides total lightning data.
6 SUMMARY AND CONCLUSION
30
Global lightning data could be a useful information for atmospheric research, climate monitoring, and
applications such as long range air traffic and communication services.
Satellite based lightning location detects uniformly CG and IC lightning over any region. The existing
prototype sensors on low orbits have been used in many research applications. Due to the short
observation times caused by the low orbits, their use for most of the real time applications is limited.
6 Summary and Conclusion
Lightning location by ground based networks is well established now. The detection and location
principles are in use for more then 10 years and were continuously improved. The reliability and
quality of the data (detection efficiency 90%, location accuracy 0.5 – 1 km) is close to the physical
limit and is sufficient for the most user requirements.
The main difference in the existing systems is in the used detection frequency band (LF or VHF).
One implication is the difference between total lightning detection (VHF, SAFIR) and return stroke
detection (LF, GAI).
The concentration of the production of the LLS in one company now (Vaisala Group) should favour
the convergence of the still existing differences of the systems as well as the exchange of knowledge
and methodology.
The majority of the ground based LLS is owned by private companies. This may limit the access to
data for the general public.
The existing ground based LLS operate in the industrialized countries. For large parts of the tropics
(Africa, Asia) and other remote areas, as well as for the oceans no lightning data are provided. The
necessary infrastructure for installation and operation (communication lines, technical support, trained
operators) of the LLS is not available or affordable everywhere. The networks operating in the LF or
VHF range do not reach presently far into the oceanic areas.
The existing satellite based lightning detection sensors give reliable data. Much experience have
been gathered in processing, validation and application of the data. The value of optical location
princip from space has been proven. The instruments detect both IC and CG lightning, the detection
efficiency for CG lightning is comparable to ground based LLS. The location accuracy is in the order
of 10 km and is limited by scattering in the clouds. Advantageous is the spatial homogeneity and
the possibility of global observation from satellite. The low orbit of the presently used platforms has
limiting implications such as short observation time and low repetition rate for a fixed point on earth.
To achieve a permanent lightning detection over the American continent a geostationary platform is
prepared by NASA.
The lightning data, where available, are used for warning of lightning hazard and the prevention of
lightning induced damages. The requirements of the most applications with respect to data availability
and quality can be met by the existing ground based LLS taking into account the improvements of
these systems.
Weather forecast and nowcasting of thunderstorms use lightning data successfully but could get ad­
ditional benefit from information about the intracloud lightning. This concerns particularly air traffic
applications and nowcasting. Global lightning studies e.g. on NOx production or climate monitoring
depend on global lightning information, which is currently not continuously available.
7 ACKNOWLEDGEMENTS
31
7 Acknowledgements
We acknowledge gracefully the provision of information and helpfully discussion which were given
in interviews and email communication by S. Senesi (Meteo-France), H. Christian (NASA), S. Th­
ern (Siemens AG, BLIDS), P. Richard (Vaisala), P. Laroche (ONERA), G. Diendorfer (ALDIS), F.
Montariol (Meteo-France), T. Kratzsch (DWD).
REFERENCES
32
References
Allen, D. J. and K. E. Pickering (2002). Evaluation of lightning flash rate parameterization for use in
global CTM. In AGU Spring Meeting, May, 28-31 Washington, D.C – USA. A51F-10.
Anderson, K. (2000). Predicting convective rainfall amounts from lightning flash density. In 20th
Conference on Severe Local Storms, September 11-15, Orlando, Florida, pp. 166–168. American
Meteorological Society, Boston.
Areitio, J., I. Herrero, and A. Ezcurra (1999). Lightning characteristics in the 1992-1996 period in
the Basque country area: Lightning-precipitation relationships. In 11th International Conference
on Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 392–395.
Autones, F., J.-M. Carriere, S. Girres, S. Senesi, and P. Thomas (1999). ASPOC - a French project
for a thunderstorm product designed for air traffic control. In 8th Conference on Aviation, Range
and Aerospace Meteorology, January 10-15, Dallas, Texas, pp. 235–239. American Meteorological
Society, Boston.
Beneti, C. and F. Sato (2000). Integration of weather radar and lightning detection systems in Brazil.
In International Lightning Detection Conference, November 7-8, Tucson, Arizona. Global Atmo­
spherics,Inc.
Bent, R. B. and W. A. Lyons (1984). Theoretical evaluations and initial operational experiences of
LPATS (lightning position and tracking system) to monitor lightning ground strikes using a timeof-arrival (TOA) technique. In VII International Conference on Atmospheric Electricity, June 3-8,
Albany, N.Y., pp. 317–328.
Betz, H.-D., B. Eisert, A. Gerl, R. Kulzer, and W. P. Oettinger (1996). Measurement of the true
magnetic and electric waveforms of VLF-atmospherics. In Proc. 25th General Assembly URSI, pp.
221.
Boccippio, D., W. Koshak, R. Blakeslee, K. Driscoll, D. Mach, W. Boeck, H. Christian, and S. Good­
man (2000). The Optical Transient Detector (OTD): Instrument characteristics and cross-sensor
validation. J. Atmos. Oceanic Technol. 17, 441–458.
Boccippio, D. J., K. L. Cummins, H. J. Christian, and S. J. Goodman (2001). Combined satelliteand surface-based estimation of the intracloud-cloud-to-ground lightning ratio over the continental
United States. Mon. Weather Rev. 129, 108–122.
Boccippio, D. J., C. Wong, and S. J. Goodman (1998). Global validation of single-station Schumann
resonance lightning location. J. Atmos. Terr. Phys. 60, 701.
Boldi, R., E. Williams, A. Matlin, M. Weber, S. Hodanish, D. Sharp, S. Goodman, and R. Raghavan
(1998). The design and evaluation of the lightning imaging sensor data applications display (LIS­
DAD). In 19th Conference on Severe Local Storms, September 14 - 18, Minneapolis, Minnesota,
pp. 631–634.
Bondiou-Clergerie, A., P. Blanchet, C. Thery, A. Delannoy, J. Y. Lojou, A. Soulage, P. Richard,
F. Roux, and S. Chauzy (1999). ORAGES: a project for space-born detection of lightning flashes
using interferometer in the VHF-UHF band. In 11th International Conference on Atmospheric
Electricity, June 7-11, Guntersville, Alabama, pp. 184–187.
REFERENCES
33
Bothwell, P. (2000). Forecasting dry thunderstorms as a part of fire weather outlooks at the Storm
Prediction Center. In 3rd Symposium on Fire and Forest Meteorology, January 9-14, Long Beach,
California, pp. 76–79.
Buechler, D. E., K. T. Driscoll, S. J. Goodman, and H. J. Christian (2000). Lightning activity within
a tornadic thunderstorm observed by the Optical Transient Detector (otd). Geophys. Res. Lett. 27,
2253–2256.
Burke, C. P. and D. L. Jones (1995). Global radiolocation in the lower ELF frequency band. J.
Geophys. Res. 100, 26,263.
Burrows, W. R., P. King, P. J. Lewis, B. Kochtubajda, B. Snyder, and V. Turcotte (2002). Light­
ning occurence patterns over canada and adjacent united states from lightning detection network
observations. Atmosphere-Ocean 40, 59–80.
Chameides, W. L., D. H. Stedman, R. R.Dickerson, D. W. Rusch, and R. J. Cicerone (1977). NOx
production in lightning. J. Atmos. Sci. 34, 143–149.
Changnon, S. A. (1989). Relations of thunderstorms and cloud-to-ground lightning frequencies. J.
Climate 2, 897–921.
Changnon, S. A. (1993). Relationships between thunderstorms and cloud-to-ground lightning in
United States. J. Appl. Meteor. 32, 88–105.
Cheze, J.-L. and H. Sauvageot (1997). Area-average rainfall and lightning activity. J. Geophys.
Res. 102, 1707–1715.
Christian, H. C., V. Mazur, B. D. Fisher, L. H. Runke, K. Crouch, and R. P. Perala (1989). The
Atlas/Centaur lightning strike event. J. Geophys. Res. 94, 169–177.
Christian, H. J. (1999). Optical detection of lightning from space. In 11th International Conference
on Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 715–718.
Christian, H. J., R. J. Blakeslee, D. J. Bocippio, W. L. Boeck, D. E. Buechler, K. T. Driscoll, S. J.
Goodman, J. M. Hall, W. J. Koshak, D. A. Mach, and M. F. Stewart (1999). Global frequency and
distribution of lightning as observed by the Optical Transient Detector OTD. In 11th International
Conference on Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 726–729.
Christian, H. J., R. J. Blakeslee, and S. J. Goodman (1989). The detection of lightning from geosta­
tionary orbit. J. Geophys. Res. 94, 13,329–13,337.
Christian, H. J., K. Driscoll, S. Goodman, R. Blakeslee, D. Mach, and D. Buechler (1996). The optical
transient detector (OTD). In 10th International Conference on Atmospheric Electricity, June 10-14,
Osaka - Japan, pp. 368–371.
COST-78 (2001). COST Action 78 – Improvement of nowcasting techniques, Final Report (Lagou­
vardos, K. and E. Liljas and B. Conway and J. Sunde ed.). European Commission.
Cramer, J. A. and K. L. Cummins (1999). Long-range and trans-oceanic lightning detection. In
11th International Conference on Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp.
250–253.
REFERENCES
34
Cummins, K., 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.
Cummins, K. L., R. B. Pyle, and G. Fournier (1999). An integrated Nort American lightning detection
network. In 11th International Conference on Atmospheric Electricity, June 7-11, Guntersville,
Alabama, pp. 218–221.
Defer, E., J. Dye, M. Venticinque, P. Blanchet, C. Thery, P. Laroche, and K. Cummins (2000). Simul­
taneous observations of CG activity from NLDN and ONERA-ITF interferometric mapper during
the STERAO-A experiment. In International Lightning Detection Conference, November 7-8, Tuc­
son, Arizona. Global Atmospherics,Inc.
Diendorfer, G. (2001). Correlation of Power Line Failures and Lightning Location Data. In 5th
International Workshop on Physics of Lightning, September 10-13, Nagoya, Japan, pp. 89–90.
Diendorfer, G., W. Hadrian, F. Hofbauer, M. Mair, and W. Schulz (2002). Evaluation of lightning
location data employing measurements of direct strikes to a radio tower. In CIGRE Session 2002,
Aug, 26-30, Paris – France.
Diendorfer, G., W. Schulz, F. Hofbauer, and A. Stimmer (1994). Results of a performance analysis of
the Austrian Lightning Location Network ALDIS. In 22nd International Conference on Lightning
Protection, September 19-23, Budapest - Hungary.
Dinnes, D. (1999). Blitzgefährdung in Deutschland. Master’s thesis, Universität München. (in Ger­
man).
Finke, U. (1999). Correlation functions of lightning distributions. Mon. Weather Rev. 127, 1850–1861.
Finke, U. (2000). Lightning on the European scale. In EULINOX - The European lightning nitrogen
oxides project, DLR - Forschungsbericht, 2000-28, pp. 85–99.
Finke, U. and T. Hauf (1996). The characteristics of lightning ocurrence in southern Germany. Con­
trib. Atmos. Phys. 69, 361–374.
Fister, V., R. Gampenrieder, E. Hoffmann, and H. von Rheinbaben (1992). Lightning positioning in
electric power supply systems. In 21nd International Conference on Lightning Protection, Septem­
ber 21-25, Berlin - Germany.
Füllekrug, M., S. Reising, and W. A. Lyons (1996). On the accuracy of arrival azimuth determination
of sprite-associated lightning flashes by Earth-ionosphere cavity resonances. Geophys. Res. Lett. 23,
3691.
Garner, T. and T. D. Oram (2000). Natural and triggered lightning forecasts for space shuttle land­
ings. In International Lightning Detection Conference, November 7-8, Tucson, Arizona. Global
Atmospherics,Inc.
Goodman, S. J. (1991). Sensor fusion technique for predicting thunderstorm evolution using lightning
and radar networks. In 25th International Conference on Radar Meteorology, Paris - France, pp.
97–102.
REFERENCES
35
Goodman, S. J. and H. J. Christian (1993). Global observations of lightning. In R. J. Gurney, J. L.
Foster, and C. L. Parkinson (Eds.), Atlas of satellite observations related to global change, pp.
191–219. Cambridge University Press.
Goodman, S. J. and R. Raghavan (1993). Investigating the relation between precipitation and lightning
using polarimetric radar observations. In 26th International Conference on Radar Meteorology,
May 24-28, Norman, Oklahoma.
Hagen, M., B. Bartenschlager, and U. Finke (1999). Motion characteristics of thunderstorms in south­
ern Germany. Meteorol. Appl. 6, 227–239.
Halsey, N. G. J. and R. Patton (1999). Investigation into lightning strikes to helicopters operating over
the North Sea. In 8th Conference on Aviation, Range and Aerospace Meteorology, January 10-15,
Dallas, Texas, pp. 259–263. American Meteorological Society, Boston.
Heavner, M. J., D. A. Smith, and J. Harlin (2000). Current Los Alamos sferic array studies. In Interna­
tional Lightning Detection Conference, November 7-8, Tucson, Arizona. Global Atmospherics,Inc.
Heckman, S., E. Williams, and R. Boldi (1998). Total global lightning inferred from Schumann
resonance measurements. J. Geophys. Res. 103, 31,775 – 31,779.
Heidler, F., C. Hopf, T. Zundl, E. Montandon, M. Rubinstein, , and H. Steinbigler (2001). Lightning
location and direct current measurement. In 23rd International Conference on Lightning Protection,
September 23-27, Florence, Italy, pp. 166–169.
Hidayat, S. and M. Ishii (1998). Spatial and temporal distribution of lightning activity around Java. J.
Geophys. Res. 103, 14,001–14009.
Hill, R. D. (1979). On the production of nitric oxide by lightning. Geophys. Res. Lett. 6, 945–947.
Hohl, R. and H.-H. Schiesser (2001). Cloud-to-ground lightning activity in relation to the radarderived hail kinetic energy in Switzerland. Atmos. Res. 56, 375–396.
Holden, D. N., C. P. Munson, and J. C. Devenport (1995). Satellite observations of transionospheric
pulse pairs. Geophys. Res. Lett. 22, 889 – 892.
Holle, R. L. and R. E. Lopez (1993). Overview of real-time lightning detection systems for use by
meteorological services. Report 51, WMO. 71 pp.
Holle, R. L., R. E. Lopez, and B. C. Navarro (2001). U.S. Lightning Deaths, Injuries and Damages in
the 1890s Compared to 1990s. NOAA Technical Memorandum OAR NSSL-106.
Höller, H., U. Finke, M. Hagen, and H. Huntrieser (1999). Lightning produced NOx (LINOX) –
experimental design and first results. J. Geophys. Res. 104, 13,911–13,922.
Huntrieser, H., H. Schlager, C. Feigl, and H. Höller (1998). Transport and production of NOx in
electrified thunderstorms: Survey of previous studies and new observations at midlatitudes. J.
Geophys. Res. 103, 28,247.
Hussein, A., W. Janischewskyj, J. Chang, V. Shostak, W. Chisholm, P. Dzurevych, and Z. Kawasaki
(1995). Simultaneous measurement of lightning parameters for strokes to the Toronto Canadian
Tower. J. Geophys. Res. 100, 8853–8861.
REFERENCES
36
Idone, V. P., D. A. Davis, P. K. Moore, Y. Wang, R. W. Henderson, M. Ries, and P. F. Jamason (1998a).
Performance evaluation of the U.S. National Lightning Detection Network in Eastern New York;
part i: Detection efficiency. J. Geophys. Res..
Idone, V. P., D. A. Davis, P. K. Moore, Y. Wang, R. W. Henderson, M. Ries, and P. F. Jamason (1998b).
Performance evaluation of the U.S. National Lightning Detection Network in Eastern New York;
part i: Location accuracy77. J. Geophys. Res..
Ishii, M., T. Shindo, N. Honma, and Y. Miyake (2000). Lightning location systems in Japan. In 25th
International Conference on Lightning Protection, September 18-22,Rhodes,Greece, pp. 161–165.
Jacobs, W. and W. E. Raatz (2001). Automatic analysis and prognosis of nowcasting relevant weather
phenomena based upon pixel attributes. In 5th European Conference on Application of Meteorology
– ECAM 2001, September 24 - 28, Budapest – Hungary.
Jacobson, A. and E. Williams (2001). Joint observations of marine lightning over the Atlantic by
the FORTE satellite and the United Kingdom Meteorological Office sferics array. In AGU Fall
Meeting, Dec, 10-14 San Francisco – USA.
Jacobson, A. R., K. L. Cummins, M. Carter, P. Klingner, D. Roussel-Dupré, and S. O. Knox (2000).
FORTE radio-frequency observations of lightning strokes detected by the National Lightning De­
tection Network. J. Geophys. Res. 105, 15,653–15,662.
Johnson, J. T., M. D. Eilts, D. Ruth, W. Goodman, and L. Rothfusz (2000). Warning operations in sup­
port of the 1996 centennial Olympic Games. Bulletin of the American Meteorological Society 81,
543–554.
Kappenman, J. G., M. E. Gordon, and T. W. Guttormson (2001). High-Precision Location of
Lightning-Caused Distribution Faults. In 16th IEEE/PES Transmission and Distribution Confer­
ence and Exposition, Oktober 28 - November 2, Georgia, Atlanta.
Kithil, R. (1999). Results of investigations into annual USA lightning costs and losses. In 11th
International Conference on Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 204–
206.
Kithil, R. (2000). Results of investigations into annual USA lightning costs and losses. Technical
report, Lightning Safety Institute, Louisville Colorado USA.
Kithil, R. (2002). Annual usa lightning costs and losses. Technical report, Lightning Safety Institute,
Louisville Colorado USA.
Knüpffer, K. (2001). Lightning MOS – a new approach for thunderstorm prediction. In 5th European
Conference on Application of Meteorology – ECAM 2001, September 24 - 28, Budapest – Hungary.
Kochtubajda, B., R. Stewart, J. Gyakum, and M. Flannigan (2002). Summer convection and lightning
over the Mackenzie River Basin and their impacts during 1994 and 1995. Atmosphere-Ocean 40,
199–220.
Krider, E. P., R. C. Noggle, A. E. Pifer, and D. L. Vance (1980). Lightning direction-finding systems
for forest fire detection. Bull. Amer. Meteor. Soc. 61, 980–986.
REFERENCES
37
Krider, E. P., R. C. Noggle, and M. A. Uman (1976). A gated, wideband magnetic direction finder for
lightning return strokes. J. Appl. Meteor. 15, 301–306.
Kummerow, C., W. Barnes, T. Kozu, J. Shiue, and J. Simpson (1998). The Tropical Rainfall Measuring
Mission (TRMM) sensor package. J. Atmos. Oceanic Technol. 15, 809–817.
Laksmiwati, H. (1998). The Indonesia lightning detection and information system JADPEN con­
figuration, organization and first lightning data. In 24th International Conference on Lightning
protection, September 14-18, Birmingham, United Kingdom, Staffordshire University, pp. 216–
220.
Laroche, P., C. Malherbe, A. Bondiou, M. Weber, C. Engholm, and V. Coel (1991). Lightning activity
in microburst producing storm cells. In 25th International Conference on Radar Meteorology,
Paris-France, pp. 85–88.
Lee, A. C. L. (1986). An experimental study of the remote location of lightning flashes using a VLF
arrival time difference lightning technique. Q. J. R. Meteorol. Soc. 112, 203–229.
Lee, D. S., I. Köhler, E. Grobler, F. Rohrer, R. Sausen, L. Gollardo-Klenner, J. J. G. Olivier, and F. J.
Dentner (1997). Estimations of global NOx emissions and their uncertainties. Atmos. Environ.. in
press.
Light, T. E., D. M. Suszcynsky, M. W. Kirkland, and A. R. Jacobson (2001). Simulations of lightning
optical waveforms as seen through clouds by satellites. J. Geophys. Res. 106, 17,103–17,114.
Lyons, W. A., E. Nelson, J. Fossum, R. A. Armstrong, M. A. Stanley, E. R. Williams, and C. Price
(2000). Initial results from STEPS: characterizing convective systems and lightning which produce
sprites. In International Lightning Detection Conference, November 7-8, Tucson, Arizona. Global
Atmospherics,Inc.
MacGorman, D. and W. Rust (1998). The Electrical Nature of Storms. Oxford University Press, New
York. 442 pp.
MacGorman, D. R. and D. W. Burgess (1994). Positive cloud-to-ground lightning in tornadic storms
and hailstorms. Mon. Weather Rev. 122, 1671–1697.
MacGorman, D. R. and C. D. Morgenstern (1998). Some characteristics of cloud-to-ground lightning
in mesoscale convective systems. J. Geophys. Res. 103, 14,011–14,024.
Maier, L., C. Lennon, P. Krehbiel, and M. Maier (1996). Lightning as observed by a four-dimensional
lightning location system at Kennedy Space Center. In 10th International Conference on Atmo­
spheric Electricity, June 10-14, Osaka - Japan, pp. 280–283.
Massey, R. S., K. B. Eack, M. H. Eberle, X. M. Shao, D. A. Smith, and K. C. Wiens (1999). Operation
of an array of field-change detectors to provide ground truth for FORTE data. In 11th International
Conference on Atmospheric Electricity, June 7-11, Guntersville Al – USA, pp. 684–687.
Massey, R. S., S. O. Knox, R. C. Franz, D. N. Holden, and C. T. Rhodes (1998). Measurements of
transionospheric radio propagation parameters using the FORTE satellite. Radio Sci. 33, 1739 –
1753.
REFERENCES
38
McCaul, E. W., D. Buechler, S. Hodanish, and S. J. Goodman (2002). The Almena, Kansas, tornadic
storm of 3 june 1999: A long-lived supercell with very little cloud-to-ground lightning. Mon.
Weather Rev. 130.
Mills, E., E. Lecomte, and A. Peara (2001). U.S. Insurance Industry Perspectives on Global Climate
Change. Lawrence Berkeley National Laboratory MS 90-4000: Norman.
Molinié, J. and C. Pontikis (1995). A climatological study of tropical thunderstorm clouds and light­
ning frequencies on the French Guyana coast. Geophys. Res. Lett. 25, 1085–1088.
Montariol, F. (2000). Gestion, efficacite et qualite des reseaux francais de detection d’impacts de
foudre. In WMO Technical Conference on Meteorological and Environmental Instruments and
Methods of Observation (TECO-2000), 23-27 October, Beijing, China, pp. 254–257. Instruments
and Observing Methods Report No.74, WMO/TD No.1028 [CD-ROM].
Morel, C. and S. Senesi (2000). Rapid developing thunderstorms. Scientific Report on PGE11,
Météo-France/CNRM.
Nash, J., E. Hibbert, G. Calaghan, P. Taylor, N. Daly, R. Burness, and G. West (2000). The UK Met.
Office’s upgraded lightning location system. In WMO Technical Conference on Meteorological and
Environmental Instruments and Methods of Observation, October 23-27, 2000, Beijing, China, pp.
250–253. WMO report no. 74, WMO/TD-No. 1028.
Nesbitt, S., R. Zhang, and R. Orville (2000). Seasonal and global NOx production estimated from the
Optical Transient Detector (OTD). Tellus 52B, 1206–1215.
Nierow, A. and R. C. Showalter (2000). An evaluation of using lightning data to improve aviation
oceanic convective forecasting for the Gulf of Mexico. In 9th Conference on Aviation, Range,
and Aerospace Meteorology, Sep 11-15, Orlando, Florida, pp. 387–392. American Meteorology
Society,Boston.
Nierow, A., R. C. Showalter, and C. G. Souders (1999). Preliminary FAA investigations into using
total lightning to improve convective forecasting for aviation. In 11th International Conference on
Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 246–249.
Oettinger, W. P., B. Eisert, and H.-D. Betz (2000). Amplitude and event-time distribution of sferics.
In 25th International Conference on Lightning Protection, September 18-22, Rhodes, Greece, pp.
188–191.
Orville, R. E. and G. R. Huffines (1999). Lightning ground flash measurements over the contigous
United States. Mon. Weather Rev. 127.
Orville, R. E., G. R. Huffines, W. R. Burrows, R. L. Holle, and K. L. Cummins (2002). The
North American Lightning Detection Network (NALDN)-first results: 1998-2000. Mon. Weather
Rev. 130, 2098–2109.
Pannett, R. A. (2000). A lightning detection and location network for New Zealand. In WMO Tech­
nical Conference on Meteorological and Environmental Instruments and Methods of Observation
(TECO-2000), 23-27 October, Beijing, China, pp. 246–249. Instruments and Observing Methods
Report No.74, WMO/TD No.1028 [CD-ROM].
REFERENCES
39
Petersen, W. and S. Rutledge (1998). On the relationship between cloud-to-ground lightning and
convective rainfall. J. Geophys. Res. 103, 14025–14040.
Petersen, W. A., L. D. Carey, and S. A. Rutledge (1999). Polarimetric radar observations and cloud
modeling studies of low lightning producing convection in the Fort Collins flash flood. In 11th
International Conference on Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 480–
483.
Price, C., J. Penner, and M. Prather (1997). NOx from lightning, 1. Global distribution based on
lightning physics. J. Geophys. Res. 102(D5), 5929–5941.
Price, C. and D. Rind (1994). Possible implications of global climate change on global lightning
distribution and frequencies. J. Geophys. Res. 99, 10,823–10,831.
Ray, P. S., D. R. MacGorman, W. D. Rust, W. L. Taylor, and L. Rasmussen (1987). Lightning location
relative to storm structure in a supercell storm and a multicell storm. J. Geophys. Res. 92, 5713–
5724.
Richard, P. and G. Auffray (1985). VHF-UHF interferometric measurements, applications to lightning
discharge mapping. Radio Sci. 20, 171–192.
Richard, P. and I. Kononov (2001). Total lightning characteristics of thunderstorm, contribution
to storm nowcasting applications. In 5th European Conference on Applications of Meteorology,
September 24-28, Budapest, Hungary.
Richard, P. and J. Y. Lojou (1996). Assessment of application of storm cell electrical activity monitor­
ing to intense precipitation forecast. In 10th International Conference on Atmospheric Electricity,
June 10-14, Osaka - Japan, pp. 284–287.
Ridley, B. A., J. E. Dye, J. G. Walega, J. Zheng, F. E. Grahek, and W. Rison (1996). On the production
of active nitrogen by thunderstorms over New Mexico. J. Geophys. Res. 101, 20,985 – 21,005.
Rubinstein, M., E. Montandon, and M. Ianoz (1994). Analysis of multi-station cloud lightning elec­
tric field pulses recorded with the Swiss LPATS network. In 22nd International Conference on
Lightning Protection, Budapest, Hungary, pp. R1c–07.
Ruffieux, D. and J. Rast (2001). Comparison of three lightning detection systems in Switzerland.
In 5th European Conference on Application of Meteorology – ECAM 2001, September 24 - 28,
Budapest – Hungary.
Schumann, W. O. (1952). Über die strahlungslosen Eigenschwingungen einer leitenden Kugel, die
von einer Luftschicht und einer Ionosphärenhülle umgeben ist. Z. f. Naturforschung 7a, 149, 250.
Sharp, A. J. (1999). Operational LPATS network in Australia. In 11th International Conference on
Atmospheric Electricity, June 7-11, Guntersville, Alabama, pp. 234–237.
Shimura, T., R. Kobayashi, T. Shindo, A. Wada, and T. Sakai (1999). Characteristics of ’superbolt’
related winter thunderclouds over the Japan Sea. In 11th International Conference on Atmospheric
Electricity, June 7-11, Guntersville, Alabama, pp. 380–383.
Smith, D. A., X. M. Shao, D. N. Holden, C. T. Rhodes, M. Brook, P. R. Krehbiel, M. Stanley, W. Ri­
son, , and R. J. Thomas (1999). A distinct class of isolated intracloud lightning discharges and their
associated radio emissions. J. Geophys. Res. 104. 4189 – 4212.
REFERENCES
40
Soriano, L. R., F. de Pablo, and E. G. Diez (2001). Relationship between convective precipitation and
cloud-to-ground lightning in the Iberian peninsula. Mon. Weather Rev. 129, 2998–3003.
Soula, S. and S. Chauzy (2001). Some aspects of the correlation between lightning and rain activities
in thunderstorms. Atmospheric Research 56, 355–373.
Steinacker, R., M. Dorninger, F. Wölfelmaier, and T. Krennert (2000). Automatic tracking of convec­
tive cells and cell complexes from lightning and radar data. Meteor. Atmos. Phys. 72, 101–110.
Suszcynsky, D., A. Jacobson, J. Fizgerald, C. Rhodes, E. Tech, and D. Roussel-Dupré (2000).
Satellite-based global lightning and severe storm monitoring using VHF receivers. In AGU Fall
Meeting, Dec, 15-19 San Francisco – USA.
Suszcynsky, D. M., M. W. Kirkland, A. R. Jacobson, R. C. Franz, S. O. Knox, J. L. L. Guillen,
and J. L. Green (2000). FORTE observations of simultaneous VHF and optical emissions from
lightning: Basic phenomenology. J. Geophys. Res. 105. 2191– 2201.
Thery, C. (2000). Evaluation of LPATS data using VHF interferometric observations of lightning
flashes during the EULINOX experiment. In U. Schumann and H. Höller (Eds.), EULINOX - The
European lightning nitrogen oxides project, DLR - Forschungsbericht, 2000-28, pp. 77–83.
Thery, C. (2001). Evaluation of LPATS data using VHF interferometric observations of lightning
flashes during the EULINOX experiment. Atmospheric research 56, 397–409.
Torres, H., L. K. Ruales, L. Barreto, and F. Herrera (1996). Experience and first results of Colombian
lightning location network. In 23rd International Conference on Lightning Protection, September
23-27, Florence, Italy, pp. 186–190.
Tuomi, T. J. (1996). Lightning observations in Finland, 1996. Technical report, Finnish Meteorologi­
cal Institute, Helsinki.
Tzanos, D. and S. Senesi (1998). A comparison between a SAFIR network and the METEORAGE
network for detection of electrical activity. In 2nd Workshop on Experimentation and Instrumenta­
tion, December 1-2, Toulouse, France. INSU/METEO-France.
Uman, M. A. (1987). The Lightning Discharge, Volume 39 of International Geophysics Series. Aca­
demic Press, Inc. 377 pp.
Volland, H. (1995). Atmospheric Electrodynamics. Boca Raton: CRC Press.
Weber, M., E. Williams, M. Wolfson, and S. Goodman (1998). An assessment of the operational
utility of a GOES lightning mapping sensor. Project Report NOAA-18, MIT Lincoln Laboratory.
Williams, E. R. (1992). Schumann resonance: A global tropical thermometer. Science 256, 1184–
1187.
Woo, J. W., E. B. Shim, J. B. Kim, and G. W. Lee (2000). A study on the statistical analysis of
lightning parameters for pertinent insulation design by LPATS in Korea (’96-’99). In International
Lightning Detection Conference, November 7-8, Tucson, Arizona. Global Atmospherics,Inc.
Zeldovich, Y. B. and Y. P. Raizer (1966). Physics of Shock Waves and High-Temperature Hydrody­
namic Phenomena. New York: Academic Press. 445 pp.
REFERENCES
41
Zoro, R., S. Sudirham, and D. Sasongko (1997). Kerosene tank explosions due to lightning strikes
in an Indonesian refinery plant. In Lightning and Mountains, Jun, 1-5, Chamonix – France, pp.
178–182.
A ABBREVIATIONS
Appendices
A
Abbreviations
ALDF
ALDIS
ATD
DE
ELF
EUCLID
FORTE
GAI
GPS
IMPACT
LANL
LF
LIS
LLP
LLS
LPATS
MSG
NALDN
NASA
NLDN
ONERA
OTD
OLS
RF
SAA
SAFIR
TOA
VHF
VLF
Advanced Lightning Direction Finder
Austrian Lightning Detection and Information System
Arrival Time Difference
Detection Efficiency
Extremely Low Frequency (30 Hz – 300 Hz)
European cooperation on Lightning Detection
Fast On-orbit Recording of Transient Events
Global Atmospherics Inc. (now Vaisala-GAI)
Global Positioning System
IMProved Accuracy from Combined Technology
Los Alamos National Laboratory
Low Frequency (30 kHz – 300 kHz)
Lightning Imaging Sensor
Lightning Location and Protection
Lightning Location System
Lightning Positioning and Tracking System
Meteosat Second Generation
North American Lightning Detection Network
National Aeronautic and Space Administration
National Lightning Detection Network (USA)
Office National d’Etudes et de Recherches Aérospatiale
Optical Transient Detector
Optical Lightning Sensor (on FORTE)
Radio Frequency
Southern Atlantic Anomaly
Surveillance et Alerte Foudre par Interférométrie Radioélectrique
Time of Arrival
Very High Frequency (30 MHz – 300 MHz)
Very Low Frequency (3 kHz – 30 kHz)
42
B RELATED INTERNET LINKS
43
B Related Internet Links
http://www.lightningstorm.com/
http://www.vaisala.com/
http://www.euclid.org/
http://thunder.msfc.nasa.gov
http://forte.lanl.gov/
http://www.lightningsafety.com
Vaisala-GAI
Vaisala Group
EUCLID
NASA Marshall Space Flight Center, Huntsville AL
FORTE Science page
National Lightning Safety Institute