RESULTS FROM GLOBAL ATMOSPHERICS’ LIGHTNING DETECTION AND RANGING (LDAR II)
RESEARCH NETWORK AT DALLAS-FORT WORTH, TEXAS, USA
Ronald L. Holle, Nicholas W. S. Demetriades, and Martin J. Murphy
Vaisala-GAI (formerly Global Atmospherics Inc.)
Tucson, Arizona, USA
Telephone: +1 520 806-7362
Fax: +1 520 741-2848
E-mail: ron.holle@vaisala.com
1.
Introduction
Global Atmospherics Inc. (GAI) installed a Lightning
Detection and Ranging (LDAR II) network for research in
the vicinity of the Dallas-Fort Worth International Airport
(DFW) in Texas, USA. The network began operation on
1 March 2001. LDAR II builds upon the VHF detection
technology first developed at NASA Kennedy Space
Center (Maier et al., 1995) called the Lightning Detection
and Ranging (LDAR) network. It was modified at the
New Mexico Institute of Mining and Technology (Rison
et al., 1999) into the Lightning Mapping Array (LMA).
LDAR II lightning networks detect all types of cloud
lightning and cloud-to-ground (CG) lightning with an
expected flash detection efficiency of greater than 95%.
The ability of LDAR II to map these lightning flashes in
three dimensions, coupled with its high flash detection
efficiency, allows a complete three-dimensional
reconstruction of a thunderstorm’s lightning production.
LDAR II data show promise for a wide range of
applications, especially in meteorology, aviation, and
atmospheric electricity. Data from this network have
been compared with CG lightning flashes from the U.S.
National Lightning Detection Network (NLDN®) and
WSR-88D base reflectivity images.
This paper summarizes recent LDAR II research in
the DFW area. Lightning channels have been detected
that initiate within convective cores and propagate up to
190 km in length.
Such flashes have obvious
implications for air and ground safety.
LDAR II
signatures have also been found within tornadic
supercells and radar-detected gust fronts.
2.
Figure 1. Map of DFW LDAR II network sensor sites. Each site
is represented by a triangle and its identifier. Texas
county boundaries and names are also shown.
in the network interior, and better than 2 km to a range
of 150 km from the center of the network. A more
complete description of LDAR II is in Cummins et al.
(2000).
The NLDN detects CG lightning with a flash
detection efficiency of approximately 90% and a median
location accuracy of 500 m (Cummins et al., 1998). The
radar data used in this study consist of Fort Worth WSR88D base reflectivity images.
Data
The DFW LDAR II network has 7 sensors on a
baseline of 20 to 30 km (Figure 1). These sensors
detect pulses of radiation produced by the electrical
breakdown processes of lightning in 5 MHz VHF bands
that have center frequencies between 61 and 64 MHz.
These pulses of radiation are used to reconstruct the
path of individual cloud and CG lightning flashes in three
dimensions. The DFW LDAR II network maps lightning
flashes in three dimensions within approximately 150 km
of the center of the network. LDAR II flash detection
efficiency is expected to be greater than 95% within the
interior of the network (a range of 30 km from DFW),
and greater than 90% out to 120 km from the airport.
LDAR II location accuracy for individual pulses of
radiation is expected to be between 100 and 200 m with-
3.
Spider flash on 17 August 2001
Figure 2 shows the intricate three-dimensional
structure of a lightning flash detected by LDAR II on 17
August 2001. This lightning flash initiated about 40 km
east-southeast of DFW airport, and propagated in a
westward arc. It terminated 25 km to the south of the
airport, as shown in the plan view in the largest panel of
Figure 2. A total of 337 LDAR II radiation sources were
detected along its path of about 100 km. The two
vertical cross-sections of Figure 2 show that this flash
mainly propagated in the middle to lower troposphere
(below 10 km). According to the NLDN, this flash
produced four isolated CG lightning flashes along its
path. The final positive CG flash injured a ground
worker at the DFW airport.
Figure 2. Spider flash detected by DFW LDAR II network during one second on 17 August 2001 starting at
1515:38 UTC. Dots are LDAR II sources that are shaded lightest at the start of the time period to dark at the
end. CG flashes from the NLDN are positive or negative signs, depending on peak current polarity. Top
panel: Altitude versus time. Lower left: Latitude versus longitude. Lower right: Latitude versus altitude.
Second panel from top: Longitude versus altitude. Smallest panel: Altitude versus source frequency.
4.
LDAR II compared with WSR-88D and NLDN
during intense squall line
An intense squall line moved through the DFW
area on 15 June 2001. Figure 3 shows the LDAR II
sources detected over a five-minute period. Baseballsized hail and straight-line winds over 113 km/hr were
reported as the squall line moved through north Texas.
LDAR II detected 105,212 sources during this 5-minute
interval in a southwest to northeast line. A smaller line
of thunderstorms was merging into the southwest end of
the squall line. In addition, smaller cells were merging
into the squall line northeast of Dallas.
The Fort Worth WSR-88D base reflectivity image is
shown by Figure 4 at 0054 UTC, in the middle of the
five-minute time period of Figure 3. The leading
convective portion of this squall line is well defined by
reflectivity over 45 dBZ. The trailing stratiform region
associated with the squall line extends on radar to near
Ardmore, Oklahoma. But this radar area is smaller than
the area covered by LDAR II sources in Figure 3. In
this case, base reflectivity did not detect the
thunderstorm hazard region as fully as did LDAR II.
The WSR-88D base reflectivity also shows the smaller
line merging into the southwest end of the squall line,
and the smaller cells merging into the line just east of
Dallas.
Figure 3. Same as Figure 2, except for all flashes from
0051:30 to 0056:30 UTC 15 June 2001.
Figure 4. Fort Worth WSR-88D base reflectivity image from
0054 UTC 15 June 2001. Grayscale key for reflectivity is
on the left. (Image from The Weather Underground,
2001, www.wunderground.com).
NLDN flashes in Figure 5 clearly delineate the
southwest-northeast squall line and other features
shown by LDAR II during the same five-minute period
(Figure 3). As with radar, LDAR II gives a more detailed
picture than the NLDN of the true two-dimensional
spatial extent of lightning activity associated with this
squall line, especially in the trailing stratiform region.
The NLDN detected about 30 CG flashes between the
leading convective line and the Oklahoma border,
whereas LDAR II detected thousands of sources. The
major reason for this difference is the ability of LDAR II
to detect a large number of radiation sources within
both cloud and CG flashes.
Figure 6. Altitude versus time plot of the two-minute LDAR II
95th percentile altitudes from a tornadic supercell
northwest of Fort Worth from 2320 UTC 12 October 2001
to 0130 UTC 13 October 2001. F2 tornado times are
shown by triangles at 0050 and 0100 UTC.
5.
A supercell thunderstorm produced two F2
tornadoes within 100 km of the center of the DFW
LDAR II network on 12-13 October 2001. An altitude
analysis was performed on this supercell as it
propagated to the northwest of Fort Worth between
2320 UTC 12 October and 0130 UTC 13 October.
The analysis involved tracking the 95 th percentile
altitude of LDAR II sources (near the top of the lightning
activity) during 2-minute intervals throughout this time
period (Figure 6). This measure of the altitude of the
top of the lightning activity remained quite constant
between 14 and 15 km during the first 80 minutes.
However, at 0040 UTC 13 October, a substantial
decrease occurred 10 minutes prior to the first report of
an F2 tornado with this storm at 0050 UTC. The
descent continued until 0056 UTC when the altitude of
the lightning top reached a minimum of 12 km. Four
minutes later at 0100 UTC, while the 95th percentile
altitude was still at 12 km, a second strong F2 tornado
was reported.
6.
Figure 5. CG flashes detected by the NLDN within 200 km of
the center of the LDAR II network from 0051:30 to
0056:30 UTC 15 June 2001. CG flashes are indicated by
black diamonds. LDAR II sites as in Figure 1.
LDAR II altitude changes in a tornadic supercell
LDAR II sources relative to gust fronts
The DFW LDAR II network has detected a
previously undocumented lightning signature in several
squall lines that passed through the DFW area. The
signature consists of flashes that are generated in the
leading convective portion of a squall line, and
propagate 10 to 45 km ahead of the line. Some flashes
coincide with a portion of the squall line’s gust front,
others terminate before reaching the gust front, and still
other flashes reach ahead of the gust front.
Figure 7 shows an example of this phenomenon
between 0205 and` 0207 UTC on 13 October 2001.
LDAR II sources associated with several flashes
originated in the highest reflectivity portion (not shown)
of the squall line. The flashes then propagated out of
the strongest convection, and formed an arc-shaped
line that is apparent in Figure 7 to the east and
northeast of the main line of storms.
Figure 7. DFW LDAR II lightning radiation sources detected between 0205 and 0207 UTC 13 October 2001. LDAR II sources
are shaded with cold colors at the start of the time period to warm at the end.
8.
Conclusions and future work
The ability to view practically all lightning within a
given thunderstorm, in the vertical as well as in the
horizontal, has lead to new insights about lightning
production in severe thunderstorms. Vaisala-GAI is
continuing LDAR II meteorological applications research
in several areas including the following:

A large emphasis is being placed on the
relationships of LDAR II source altitude trends and
cloud lightning flash rates to both severe and
nonsevere thunderstorms.

Spider lightning - flashes that extend tens of km in
length with many branches - will continue to be
investigated. Rates detected by LDAR II may
exceed 5 flashes per minute over extended periods
of time within strong thunderstorms and squall
lines. In the majority of cases, the spatial extent of
lightning activity defined by LDAR II is much larger
than the area defined by high radar reflectivity (>30
dBZ) and CG lightning data. LDAR II data have
been observed as far as 40 km outside the area
defined by CGs.

The utility of spider lightning flashes as a tool to
detect and forecast thunderstorm maturity will also
be investigated.

LDAR II lightning sources will be explored to help
define which radar reflectivity thin line features
represent gust fronts and which do not.
9.
References
Cummins, K.L., M.J. Murphy, E.A. Bardo, W.L. Hiscox,
and R.B. Pyle, 1998: A combined TOA/MDF
technology upgrade of the U.S. National Lightning
Detection Network, J. Geophys. Res., 103, 90359044.
—, M.J. Murphy and J.V. Tuel, 2000: Lightning detection
methods and meteorological applications. IV
International Symposium on Military Meteorology,
Malbork, Poland, September 25-28, 85-100.
Maier, L., C. Lennon, T. Britt and S. Schaefer, 1995:
LDAR system performance and analysis, in
Proceedings of the International Conference on
Cloud Physics, Amer. Meteor. Soc., Boston, Mass.,
Dallas, Tex.
Rison, W., R.J. Thomas, P.R. Krehbiel, T. Hamlin, and
J. Harlin, 1999: A GPS-Based three-dimensional
lightning mapping system: Initial observations in
central New Mexico, Geophys. Res. Lett., 26,
3573-3576.