L
II
POSITIVE LIGHTNING AND
BIPOLAR LIGHTNING PATTERNS:
OBSERVATIONAL CHARACTERISTICS
by
Cynthia Dorothy Kidder Engholm
B.S., North Carolina State University
(1985)
Submitted to the department of
Earth, Atmospheric, and Planetary Sciences
in partial fulfillment of the requirements
for the degree of
Master of Science in Meteorology
at the
Massachusetts Institute of Technology
(October 1988)
(c) Massachusetts Institute of Technology
Signature of Author
Department of Eart1Atmospheric, andi lanetary Sciences
October 1988
Certified by
Earle R. Williams
Thesis Supervisor
Certified by
Randall M. Dole
Thesis Supervisor
Accepted by
Thomas H. Jordan
Chairman, Departmental Committee on Graduate Students
i-;Tta
ABSTRACT
(1) to
The primary objectives of this study are:
document the characteristics of positive (anomalous polarity)
cloud-to-ground lightning focusing on the differences between
(2) to determine the
the summer and winter environments;
extent to which current observations support or refute the
and inverted dipole hypotheses for the
dipole
tilted
production of positive lightning; and (3) to document the
life cycles of mesoscale bipolar patterns of lightning
distributions. In order to meet these objectives, advantage
is taken of the extensive data on lightning distributions now
available from lightning detection networks.
The primary data source for this study was the Genesis
of Atlantic Lows Experiment (GALE) dataset, which includes
sounding data, radar data from the MIT radar in North
Carolina, corona point records from the radar site, and
SUNYA
the
from
information
cloud-to-ground lightning
This data is taken as
network.
detection
lightning
winter
the
in
activity
lightning
representative of
compared
then
were
analyses
these
of
Results
environment.
from
taken
with summertime radar and electrical data
Alabama.
Huntsville,
The results of these analyses largely agree with results
Areas in which
from previous studies about lightning.
positive flash
(1)
are:
results disagree with past findings
and the
locations
flash
locations upshear from negative
currents
positive
small
region of strongest reflectivity; (2)
(3) positive lightning at shear levels
in summer storms;
(4) active charge
below a previously defined threshold;
which is also
region
separation in trailing stratiform cloud
and (5) the
flashes;
producing a few positive cloud-to-ground
to the
isotherm
relationship of the height of the -10*C
importance.
primary
of
not
is
percentage of positive flashes
(1) an expansion on the
Major new findings include:
relationship between positive flashes and colder climatic
conditions, including the result that the percentage of
positive flashes increases with increasing latitude along the
East coast; (2) an examination of the diurnal variation in
(3) an
no diurnal cycle;
showing
activity
lightning
and
Stream;
Gulf
the
by
activity
enhancement of electrical
The
patterns.
lightning
bipolar
(4) the life cycle of
level
upper
and
patterns
alignment between bipolar lightning
winds favors the tilted dipole hypothesis.
geostrophic
Evidence is given of occurrences of both tilted dipole and
inverted dipole mechanisms for positive lightning production.
- 2 -
Table of Contents
Abstract
. . . . . . . . . .
. . . . . . . . . . . . . . .2
List of Tables
. . . . . . . . . .
. . . . . . . . . . . . . . .4
List of Figures
. . . . . . . . . .
. . . . . . . . . . . . . . .5
Chapter 1
1.1
1.2
1.3
Chapter 2
2.1
2.2
2.3
2.4
2.5
Chapter 3
3.1
3.2
3.3
3.4
Chapter 4
4.1
4.2
4.3
4.4
4.5
Introduction
Objectives . . . . .
Review of Past Work
Data Sources .
. .
.
S . . . . . . . . . . . . . 16
S . . . . .
. .
. . 18
S . . . . . . . . . . . . . 31
Climatological Phenomena
Introduction . ................
Geographic Variability . .............
First Stroke Peak Currents . ...........
Diurnal Effects . ................
Summary . . . . . . . . . . . . . . . . . ..
. . 39
41
57
69
.. 73
Convective Scale Phenomena
Introduction
Summer Storms
Winter Storms
Comparison and
. . . . . . . . . . . . . . . . . . 75
. . 78
. ......
..........
. .................
98
Summary ..............
.117
Mesoscale Phenomena
Introduction
. ...............
.. 122
... .124
Environment of Anomalous Lightning Events
. . .. 128
Symmetric Instability ..........
Bipolar Lightning Patterns ............
.150
Summary . . . . . .. . . . . . . . . . . .. .. .187
. . . . . . . . . . . . . . . . . . .189
Chapter 5
Conclusions
Appendix A
Symmetric Instability
Appendix B
List of Upper Air Sounding Stations
Appendix C
Bipole and Environmental Characteristics . . . . .205
. . . . . . . . . . . . . .200
. . . . . . .204
Acknowledgements . . . . . . . . . . . . . . . .
. . . . . . . . .210
. . . . . . . . . . . . . . . .
. . . . . . . . .211
References
- 3 -
List of Tables
Chapter 2
2.1
Chapter 3
Climatological Phenomena
Number of flashes and average first stroke
peak currents for the Yearly, Winter and
. . . . . . 58
Summer observations . .........
Convective Scale Phenomena
3.1
vertical
Values used in calculating the
with the
associated
reflectivity
of
gradient
and
flash
negative
a
of
location
strike point
. . . . . 82
.
......
...
.
flash
of a positive
3.2
Cloud-to-ground flashes detected by the SUNYA
network associated with field changes in the
corona point record form Fort Fischer ......
3.3
3.4
Chapter 4
4.1
.104
Summary of mean reflectivities associated
with strike point locations for Huntsville
and GALE data . . . . . . . . .. . . . . . . .
. .118
and vertical
for the two
presented in
. . .. . . . .
. .119
Determination of magnitude of deviation of
bipole orientation that can be explained by
inability to accurately determine geographic
. . . . .. . . . .
centers . . . ........ .
. .160
Summary of cloud top height
gradient means
reflectivity
summer and two winter storms
.
chapter 3 . . . ........ .
Mesoscale Phenomena
-
4 -
LIST OF FIGURES
Section 1.1
Review of Past Work
1.1
Life-Cycle of a thunderstorm (Krehbiel, 1986)
1.2
Inverted Dipole
1.3
Tilted Dipole
1.4
Percent of lightning flashes to
ground
lowering
positive charge as a function of month (Orville, et al,
1987)
. . . . . . . . . . . . . . . . .
. . .
. . .. .
28
The variation inthe number of flashes (solid line) and
median peak radiation field (dashed line) by month for
negative flashes (Orville, et al, 1987)
. .......
29
As in 1.5a except for positive flashes . ........
29
1.5a
1.5b
Section 1.2
1.6
1.7
1.8
1.9
. ....
. ..................
. . . . . . . . . . .
26
. 27
. . . . . . . . . . 27
Data Sources
Inner and Regional GALE Area rawinsonde network.
Open
circles are surrounding NWS sites, closed circles are
NWS sites in the Regional Area. Stars are VIZ sites,
squares are GMD sites and triangles are CLASS sites.
The Inner Area extends from LHW-SRL in the southwest to
WA
in the northeast.
The Research Vessels Cape
Hatteras and Endeavor are also shown . .........
35
The GALE Doppler radar network. Circles indicate the
ground-based Doppler radars: NASA SPANDAR (SPA), NCAR
CP-3, NCAR CP-4, and MIT C-band (farthest south at Fort
Fischer).
Range circles are at 200, 140, 140, and
140km for the respective radars.
The University of
Washington TPA-11 is shown by a square . ........
36
Locations of the SUNYA lightning detectors and their
area
of
coverage during GALE
(Orville, personal
communication) ...................
..
37
Location of MIT C-band
radar
in
Huntsville
Alabama.
Circle indicates range of observation which was 56km . . 38
1.10
Location of corona point recorder sites for
the
experiment in Huntsville.
Site number 25 is located
near the radar ...................
. . 38
- 5 -
Section 2.2
2.1a
Geographic Variability
Spatial coverage of the SUNYA lightning detection
network in 1985 (Orville, et al, 1987) . ........
51
2.1b
Comparison of area of coverage by the SUNYA lightning
detection network for the Summer, Winter and Yearly
datasets . . . . . . . . . . . . . . . . . . . . . . . . 51
2.2a
Number of positive flashes observed within a one degree
latitude belt divided by the total number of flashes
observed within that same belt for the entire GALE
timeperiod . . . . . . . . . . . . . . . . . . . . . . . 52
2.2b
Total number of flashes of either polarity observed
within a one degree latitude belt for the entire GALE
timeperiod . . . . . . . . . . . . . . . . . . . . . . . 52
2.3
Sea Surface Temperature contours drawn for every degree
Celsius, with an outline of the area of negative flash
densities greater than 100 flashes/square kilometer
. . . . . . 53
during the GALE timeperiod . ........
2.4a
Negative Flash Densities per square kilometer for the
. . . . . . . 54
entire GALE timeperiod . .........
2.4b
Positive Flash Densities per square kilometer for the
. . . . . . . . . 54
entire GALE timeperiod . .......
2.5
Mean number of thunderstorm days in February for the
years from 1951 through 1975 (Kessler, 1986) . .....
55
2.6
Topographic map of the eastern United States, with
(U.S.
and 500 meters.
200
at
drawn
contours
Geological Survey) . . . . . . .. . . . . . . . . . . . 55
2.7
Geographic Distribution of the Percentage of Positive
. . . . . . . . . . . . . . . . .
Flashes . . . . .
Section 2.3
56
First Stroke Peak Currents
2.8a
Number of flashes of positive polarity observed within
the region of the lightning detection network in the
Yearly dataset (Orville, et al, 1987). . ........ . 64
2.8b
As in 2.8a except for negative flashes (Orville, et al,
. . . . . . . . . . . . . .
.........
1987) .
64
As in 2.8a except for the Winter dataset . .......
65
2.8c
- 6 -
2.8d
As in 2.8c except for negative flashes . ........
65
2.8e
As in 2.8a except for the Summer dataset . .......
66
2.8f
As in 2.8e except for negative flashes . ........
66
2.9a
Average first stroke peak current for negative flashes
in kiloamperes for the Winter dataset . ........
67
2.9b
As in 2.9a except for positive flashes . ........
67
2.10
Estimate of net current per unit area
. ........
68
2.11
Average first stroke peak current versus flashrate . .
Section 2.4
Diurnal Effects
72
. .........
2.12a
Percent Positive vs Hour of the Day
2.12b
Number of Flashes vs Hour of the Day .........
Section 3.2
. 68
. 72
Summer Storms
3.1a
Solid
500 millibar chart for June 22, 1987, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. . 87
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
temperature
3. 1b
Solid
Surface chart for June 22, 1987, 0 UTC.
are isobars, fronts are analyzed . ..........
lines
. 87
3.1c
NWS radar summary chart for June 21, 1987 20:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . . 88
3.1d
Sounding from Nashville Tennessee (BNA)
1987, 0 UTC.
for
June
22,
Temperature and dewpoint are plotted . .
. 88
3.2a-c
in
radar
Constant elevation scans from the MIT
Dates and times (in UTC) are labelled; a
Huntsville.
. 89
is earliest time, c is latest . ...........
3.3a-c
As in figure 3.2 but for constant azimuth
Contours are drawn for 15, 30, and 45 dBz.
indicate range and first stroke peak current in
of lightning flashes within 7.5km and 3 minutes
observation . . . . . . . . . . . . . . . . . .
3.4
Sample of reflectivity versus height for a negative
flash strike location (solid line) and a positive flash
-
7 -
scans.
Arrows
kamps
of the
. . . . 90
strike location (dashed line).
............ .
91
3.5
Stations run
Corona point records for June 21, 1987.
east (top) to west (bottom). See figure 1.10 for exact
locations . . . . . . . . . . . . . . . . . . . . . . . 92
3.6a
Solid
500 millibar chart for July 12, 1988, 0 UTC.
are
lines
dashed
contours,
height
are
lines
.
.
.
.
............... . . . . . .
temperature ..
3.6b
3.6c
3.6d
93
NWS radar summary chart for July 12, 1988 1:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined .............. . . . . . . . . . ...
1988, 0 UTC.
3.7a-d
lines
Solid
Surface chart for July 12, 1988, 0 UTC.
are isobars, fronts are analyzed . ...........
Sounding from Nashville Tennessee (BNA)
for
93
July
94
12,
Temperature and dewpoint are plotted . . . 94
in
radar
Constant elevation scans from the MIT
Dates and times (in UTC) are labelled; a
Huntsville.
.... ...........
is earliest time, d is latest
95
scans.
Arrows
kamps
of the
bright
. . . . 96
3.8a-b
As in figure 3.7 but for constant azimuth
Contours are drawn for 15, 30, and 45 dBz.
indicate range and first stroke peak current in
of lightning flashes within 7.5km and 3 minutes
observation. Dashed parallel lines indicate
band . . . . . . . . . . . . . . . . . . . . . .
3.9
Stations run
Corona point records for July 12, 1988.
for exact
1.10
figure
See
(bottom).
west
to
(top)
east
. 97
locations. . . . . . . . . . .. . . . . . . . . . . .
Section 3.3
Winter Storms
3.10a
Solid
500 millibar chart for January 30, 1986, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. .105
. . . .. . . . .
temperature . . . ..........
3.10b
Surface chart for January 30, 1986, 0 UTC. Solid lines
.105
are isobars, fronts are analyzed ............
3.10c
NWS radar summary chart for January 30, 1986 2:35 UTC.
VIP levels are contoured, maximum cell heights are
. .106
.. . . . .
underlined . . . . . . . . . . . . ...
- 8 -
3.10d
Sounding from Cape Hatteras, North Carolina (HAT) for
January 30, 1986, 0 UTC. Temperature and dewpoint are
plotted . . . . . . . . . . . . . . . . . . . . . .. .106
3.11a-c
Constant elevation scans from the MIT radar in Ft.
Dates and times (in UTC) are labelled; a is
Fischer.
earliest time, d is latest ..............
.107
3.12a-b
As in figure 3.11 but for constant azimuth scans .
3.13
Corona point record for January 30, 1986 from
.
.........
Fischer, North Carolina .......
.
.
.108
Ft.
.109
3.14a
500 millibar chart for March 14, 1986, 0 UTC.
Solid
lines
are
height
contours,
dashed
lines
are
temperature . . . . . . . . . . . . . . . . . . . ... .110
3.14b
Solid
Surface chart for March 14, 1986, 0 UTC.
are isobars, fronts are analyzed . ...........
lines
.110
3.14c
Solid
500 millibar chart for March 15, 1986, 0 UTC.
lines
are
height
contours,
dashed
lines
are
temperature . . . . . . . . . . . . . . . . . . . . . .111
3.14d
Solid
Surface chart for March 15, 1986, 0 UTC.
are isobars, fronts are analyzed ...........
3.14e
NWS radar summary chart for March 14, 1986 6:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined .......
. . . .
...
. ....... . .112
3.14f
Sounding from Wilmington, North Carolina (ILM) for
March 14, 1986, 0 UTC. Temperature and dewpoint are
plotted . . . . . . . . . . . . . . . . . . . . . . . .112
3.15a-d
Constant elevation scans from the MIT radar in Ft.
Fischer.
Dates and times (in UTC) are labelled; a is
.113
earliest time, d is latest ..............
3.16a-b
As in figure 3.15 but for constant azimuth scans .
.
.
3.16c-e
As in figure 3.15 but for constant azimuth scans .
.
. .115
3.17
Corona point record for March 14, 1986
Fischer, North Carolina ................
Section 4.2
4.1
from
lines
. . .111
.114
Ft.
.116
Environment of Anomalous Lightning Events
Probability of the production of a positive flash
related to magnitude of wind speed shear (Takeuti,
1984) . . . . ... . . . . . . . . . . . . ... . . . .. 127
- 9
4.2
Section 4.3
Probability of the production of a positive flash
related to the height of the -10C isotherm (Takeuti,
1984) . . . . . . . . . . . . . . . . . . . . . . . ...127
Symmetric Instability
4.3a
Location of cross-section and stations for symmetric
instability analysis for January 26, 1986, 0 UTC.
.134
Lightning strike points are also noted ........
4.3b
Location of cross-section and stations for symmetric
instability analysis for February 6, 1986, 0 UTC.
.134
Lightning strike points are also noted ........
4.3c
Location of cross-section and stations for symmetric
instability analysis for March 12, 1986, 12 UTC.
.135
Lightning strike points are also noted ........
4.3d
Location of cross-section and stations for symmetric
instability analysis for March 14, 1986, 12 UTC.
.135
Lightning strike points are also noted ........
4.4a
Solid
500 millibar chart for January 26, 1986, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. .136
.
.
.
.
.
.
.
.
.
.
.
.
.......
.
.
temperature
4.4b
Surface chart for January 26, 1986, 0 UTC. Solid lines
.136
are isobars, fronts are analyzed ............
4.4c
NWS radar summary chart for January 25, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
. .137
underlined . . ........ . . . . . . . . . . . .
4.5a
Solid
500 millibar chart for February 6, 1986, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. .138
.
.
.
.
.
.
.
.
.
.
.........
.
.
temperature
4.5b
Surface chart for February 6, 1986, 0 UTC. Solid lines
.138
are isobars, fronts are analyzed ............
4.5c
NWS radar summary chart for February 5, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
. .139
.. . . . .
underlined . . . . . . . . . . . . ...
4.6a
Solid
500 millibar chart for March 12, 1986, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. .140
.
.
.
.
.
.
. . .
temperature . . . ......... .
-
10 -
4.6b
Solid
Surface chart for March 12, 1986, 0 UTC.
are isobars, fronts are analyzed . ...........
lines
.140
4.6c
500 millibar chart for March 13, 1986, 0 UTC.
Solid
lines
are
height
contours,
dashed
lines
are
temperature . . . . . . . . . . . . . . . . . . . . . . 141
4.6d
Surface chart for March 13, 1986, 0 UTC.
Solid
are isobars, fronts are analyzed ............
lines
.141
4.6e
NWS radar summary chart for March 11, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . .. .142
4.6f
NWS radar summary chart for March 12, 1986 12:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . .143
4.6g
NWS radar summary chart for March 12, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . .143
4.7a
Solid
500 millibar chart for March 14, 1986, 0 UTC.
lines
are
lines
are
height
contours,
dashed
temperature . . . . . . . . . . . . . . . . . . . . . .144
4.7b
Solid lines
Surface chart for March 14, 1986, 0 UTC.
are isobars, fronts are analyzed . . . . . . . . . . . .144
4.7c
500 millibar chart for March 15, 1986, 0 UTC.
Solid
lines
are
lines
are
height
contours,
dashed
temperature . . . . . . . . . . . . . . . . . . . . . .145
4.7d
Solid lines
Surface chart for March 15, 1986, 0 UTC.
are isobars, fronts are analyzed . . . . . . . . . . . .145
4.7e
NWS radar summary chart for March 13, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . .146
4.7f
NWS radar summary chart for March 14, 1986 12:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . .147
4.7g
NWS radar summary chart for March 14, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined .............. . . . . . . . . . ....... . .147
4.8a
Symmetric instability cross-section for January 26,
1986, 0 UTC. Origin of x at 35.0 north latitude, 84.7
west longitude, aligned with wind from 344 degrees . . .148
-
11 -
4.8b
Symmetric instability cross-section for February 6,
1986, 0 UTC. Origin of x at 34.0 north latitude, 87.2
west longitude, aligned with wind from 335 degrees . . .148
4.8c
Symmetric instability cross-section for March 12, 1986,
12 UTC. Origin of x at DAY, aligned with wind from 322
. .149
. . . ... . . . .
degrees . . . .......... .
4.8d
Symmetric instability cross-section for March 14, 1986,
12 UTC. Origin of x at 36.4 north latitude, 85.0 west
longitude, aligned with wind from 317 degrees. ... . . .149
Section 4.4
Bipolar Lightning Patterns
.164
4.9
Example of a bipolar lightning pattern ........
4.10
Locations of the six bipoles that will be examined in
Bipoles are labelled by the date on
this chapter.
which they started. Arrows point from the negative
geographic center to the positive geographic center
. .165
4.11a
Solid
500 millibar chart for January 24, 1986, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. .166
.
.
.
.
.
.
.
.
.
.
.........
temperature . .
4.11b
Surface chart for January 24, 1986, 0 UTC. Solid lines
.166
are isobars, fronts are analyzed ............
4.11c
NWS radar summary chart for January 23, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
. .167
. . . .. . . . .
underlined . . . . .......... .
4.12a
Solid
500 millibar chart for January 30, 1986, 0 UTC.
are
lines
dashed
contours,
height
are
lines
. .168
.
.
.
.
.
.
.
. . . . .
temperature . . ...... .
4.12b
Surface chart for January 30, 1986, 0 UTC.
Solid lines
are isobars, fronts are analyzed ............
.168
4.12c
NWS radar summary chart for January 29, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
. .169
. . .. . . . .
underlined . . . . . . . . . ....
4.13a
500 millibar chart for February 21, 1986, 0 UTC. Solid
are
lines
dashed
contours,
height
are
lines
. .170
.
.
.
.
.
.
.
.
.
...
.
.
.
.
.
temperature . . .
4.13b
Solid
Surface chart for February 21, 1986, 0 UTC.
.170
........
lines are isobars, fronts are analyzed
-
12 -
4.13c
500 millibar chart for February 22, 1986, 0 UTC. Solid
lines
are
height
contours,
dashed
lines
are
temperature . . . . . . . . . . . . . . . . . . . . . . 171
4.13d
Surface chart for February 22,
1986, 0 UTC.
Solid
lines are isobars, fronts are analyzed ........
.171
4.13e
NWS radar summary chart for February 20, 1986 23:35
UTC.
VIP levels are contoured, maximum cell heights
are underlined ........
.
............. . . .172
4.13f
NWS radar summary chart for February 21, 1986 12:35
UTC.
VIP levels are contoured, maximum cell heights
are underlined ...................
...172
4.13g
NWS radar summary chart for February 21, 1986 23:35
UTC.
VIP levels are contoured, maximum cell heights
are underlined ...................
...173
4.14a
500 millibar chart for March 7, 1986, 0 UTC.
Solid
lines
are
height
contours,
dashed
lines
are
temperature . . . . . . . . . . . . . . . . . . . . . . 174
4.14b
Solid
Surface chart for March 7, 1986, 0 UTC.
are isobars, fronts are analyzed ............
lines
.174
4.14c
NWS radar summary chart for March 6, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . .175
4.15a
500 millibar chart for March 11, 1986, 0
lines
are
height
contours,
dashed
temperature ...................
4.15b
Solid lines
Surface chart for March 11, 1986, 0 UTC.
.. 176
are isobars, fronts are analyzed . . ........
4.15c
NWS radar summary chart for March 10, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are
underlined . . . . . . . . . . . . . . . . . . . . . . .177
4.16
Comparison of geostophic wind speed and direction
the levels of 850, 700, 500, and 250 millibars ...
4.17
Comparison of orientations of mean geostrophic wind,
See
text
for further
cloudmass.
bipole,
and
. . . . . .. ..... .179
explanation. ...........
4.18
Fictitous bipole that was created to illustrate the
. .180
mean relative orientations for the six bipole cases
-
13 -
UTC.
Solid
lines
are
.. .176
for
. .178
4.19
Comparison of the wind speed shear in the layers of 850
-
700,
700 - 500, and 500 - 250 millibars for the six
bipole cases . . . . . . . . . . . . . . . . . . . . . . 181
4.20a
Each dashed line
Movement of the January 30 bipole.
represents the size and orientation of the bipole for a
given hour, which is labelled
at
the
negative
geographic center.
Positive geographic centers are
marked with a "+". ..
..
......
.........
.
.182
4.20b
As in 4.20a, but for the February 21 bipole. . ... . . .182
4.21a
Distance between the geographic centers as a function
of time for the January 30 bipole. Error bars were
estimated using the standard deviation of the mean
latitude and longitude for each geographic center. . .
.183
4.21b
As in 4.21a, but for the February 21 bipole. . ... . . .183
4.22a
Storm-wide flashrate for the January 30 bipole as a
function of time.
Negative flashrate is shown as a
solid line with boxes, positive flashrate is shown as a
dashed line with triangles ..............
.184
4.22b
As in 4.22a, but for the February 21 bipole. . ... . . .184
4.23a
Maximum cloud top height of the cloudmass as reported
by NWS radar summary charts for the January 30 bipole. .185
4.23b
As in 4.23a, but for the February 21 bipole. . ... . . .185
4.24
Example of the westward tilt of the cold air aloft as
represented by a trough in the height contours at a
standard pressure level. Each line represents a single
time spaced twelve hours apart. Hatched area indicates
time during which a bipole existed ..........
.186
4.25
As in 4.24, but averaged over the six bipole cases . . .186
Chapter 5
Conclusions
5.1a
Sketch of a single tilted dipole.
Cloud-to-ground
lightning flashes are noted, with "x" for negative
strike point and "+" for positive strike point ... . .197
5.1b
Top view of a group of tilted dipoles. Symbols as in
figure 5.1a. . .................... .. . .197
5.2a
Row of identical tilted dipoles. Symbols as in figure
5.1a . . . . . . . . . . . . . . . .
. . . . . . ... ..198
-
14 -
5.2b
Row of tilted dipoles illustrating the effect of
temperature. Symbols as in figure 5.1a. . ....... .198
5.3
Sketch of one way of creating a bipole with
a
combination of upright and inverted dipoles. Symbols
as in figure 5.1a. . ................... . .199
5.4
Sketch of a second way of creating a bipole
combination of upright and inverted dipoles.
as in figure 5.1a. . ..................
Appendix A
A.1
with a
Symbols
.199
Symmetric Instability
Sample symmetric instability plots ..........
-
15 -
.203
CHAPTER 1
Introduction
1.1
Objectives
With the implementation of lightning detection networks has
come
the ability to study cloud-to-ground lightning from a new perspective.
of
Lightning studies are no longer restricted to special observations
individual
storms
over
lightning
flashes that occur within large
Since
areas.
limited
all cloud-to-ground
detection
areas,
network
as within the SUNYA network, are recorded, it is now possible to
such
determine certain
properties
climatological
of
lightning
such
as
seasonal and geographical differences in cloud-to-ground lightning.
The focus of
lightning
only
study
cloud-to-ground
(i.e.
will
events.
lightning
cloud-to-ground
refer to
this
In
this thesis "lightning" will
events.
Normal
polarity
lowering negative charge to ground) will be referred
to as "negative", and the reverse "anomalous" polarity lightning
be
called
"positive".
electrical
will
"Peak current" will always refer to the peak
current of the first stroke of the flash.
gross
of
on
lightning
the
characteristics
be
structure
of
"Dipole" will indicate
the
the cloud, and in particular the two
-
16 -
Introduction
The word "bipole" will be used to
main charge regions.
describe
the
of strike points of cloud-to-ground flashes (see chapter
organization
4, section 4).
A
objective
basic
relationships
between
occurrences
temperature,
vertical
wind
shear,
of
positive
and
cloud
The
will
be
geographic
cloud-to-ground flashes known as a "bipole" will also be
of
An examination will be made of
documented.
of
production
the
the
and
lightning
structure
Evolution of the mesoscale organization of
examined.
document
to
of positive lightning in the winter environment.
characteristics
location
is
study
this
of
positive
hypotheses
two
for
the
the "inverted dipole" hypothesis
lightning:
and the "tilted dipole" hypothesis.
This study will concentrate
on
observed
winter,
since
indicated that the
highest
percentages
observed
(Orville,
et
positive
this is when previous studies have
the
in
lightning
of
features
of
positive
lightning
are
General characteristics of winter
al, 1987).
lightning will be identified, with a major goal being to identify
key
differences in the characteristics of positive and negative lightning.
Convective scale
compared
with
lightning
patterns
will
results of previous work.
will be presented and the characteristics
described.
The
also
be
documented
and
Six case studies of bipoles
and
evolution
of
bipoles
results will then be used as a basis for identifying
the strengths and weaknesses of the inverted dipole and tilted
hypotheses.
-
17 -
dipole
Introduction
1.2
Review Of Past Work
been
study
under
than
more
for
fifty
Figure 1.1 shows a
years.
when negative lightning will have its
is
associated
usually
heavy
with
(Simpson and Scrase, 1937; Kuettner, 1950; Jacobson and
precipitation
Following the active phase
Krider, 1975).
in
increase
relative
phase
This
highest flashrate.
the
production
of
toward
tendency
a
is
Although
1983).
the
picture
shows
occurring between the ground and a positive
into
a
positive lightning in the
mature and late stages (figure 1.lf) (Rust, et al, 1981a; Orville,
al,
is
Note that there
typical thunderstorm life-cycle (Krehbiel, 1986).
an "active" phase (figure 1.1d),
has
thunderstorms
of
The life-cycle of the electrical activity
the
et
late stage lightning
charge
extending
region
the anvil, positive lightning has been documented as originating
from anywhere in the cloud (Simpson and Scrase,
1937;
Rust,
al,
et
1981b).
The tripolar structure of the thunderstorm's electrical field has
been
established,
well
the idea that the
location
(Simpson and
related
and there is substantial evidence supporting
of
Scrase,
the
centers
charge
1937;
Simpson
and
Kuettner, 1950; Reynolds, 1954; Pierce, 1955b; Magono,
et al, 1983; Takahashi, 1978).
Christian,
above.
isotherm.
be
The
temperature
Robinson,
1941;
al,
1983;
et
The negative charge center
is found somewhere around -5 to -20oC with the
center
is
main
positive
charge
0
lower positive charge center occurs near the 0 C
This charge center - temperature dependence is believed
related
to
to
the ice-physics involved in the charging mechanism of
- 18 -
Introduction
the thunderstorm (Jayaratne, et al,
will tend to become negatively charged
at
particles
and
temperatures
lower
through
fall,
will
and a positive
are
particles
charge
ice
with
collisions
charged at higher
positively
graupel
The
temperatures, for typical values of cloud water content.
particles
-100 C
approximately
At
reversal temperature such that graupel particles
charge
a
is
there
1983).
creating a negative charge center around -100C
center
carried
below.
the
by
The
ice
charged
positively
updraft into the top of the cloud or
anvil, creating an upper positive charge center (Takahashi, 1978).
Observations indicate that the
but
spherical
has
more
of
a
negative
charge
particles
the
(Takahashi, 1978).
horizontal
et
al,
1983;
The lower positive charge is usually localized and
1986).
relatively weak, possibly because some of the charge is
by
not
layer structure, with the horizontal
extent much larger than the vertical depth (Christian,
Krehbiel,
is
center
when
carried
away
they precipitate out the bottom of the cloud
The upper positive charge center may cover a large
area, extending into the cirrus anvil of the thunderstorm.
The result is a tripolar structure, with the main dipole consisting of
a
concentrated
negative
layer
charge
in
the
region of the -100C
isotherm, an extensive region of positive charge above, and
localized
positive
region
OOC,
near
dipole.
-
19 -
a
lower,
which is not part of the main
Introduction
Two hypotheses have been suggested to account for
of
positive lightning.
the
existence
One is the "inverted dipole" hypothesis.
The
inverted dipole hypothesis suggests that the main dipole consists of a
negative
charge
region
above,
possibly
located
positive charge region below (see figure 1.2).
near -10 C, and a
Lightning then
between the lower (positive) charge region and the earth.
dipole could occur in a moderately
were
updraft
not
strong
enough
convective
charge
earlier but now
structure
region.
with
now
is
a
above
environment.
If
the
not
be
significant
a
upper
This would result in a tripole as described
the
charge
main
a
inverted.
regions
dipole
positive charge below, with
charge
An inverted
to attain heights above the charge
reversal temperature, then there would
positive
occurs
The
tripolar
of negative charge near -100C and
relatively
weak
region
of
positive
the negative charge region that is not part of the main
dipole.
In this case, the inverted dipole would be
shallow
clouds.
characterized
by
Since convective clouds during winter are typically
shallower than during summer, this hypothesis would suggest a seasonal
preference
for
positive
lightning.
It also suggests an association
between positive flashes and shallow clouds of low radar reflectivity.
The tilted dipole hypothesis proposes that vertical wind shear is
responsible
for
slanting the updraft and thus creating a dipole that
is tilted with respect to the vertical.
region
upper
positive
charge
is then located downshear from the main negative charge region
as shown in figure 1.3.
al,
The
1982),
Based on observational
evidence
(Brook,
et
a threshold value of 1.5 m/s/km has been suggested as the
- 20 -
Introduction
minimum wind shear necessary for the production of positive lightning.
In
winter,
the
meridional
temperature gradient in mid-latitudes is
typically much greater than in summer.
predicts
greater
wind
temperature gradient.
Greater wind shear increases the probability of
updrafts,
winter.
This mechanism may
region
be
making
closer
the
tilted
also
dipole more likely to form in
require
that
that
the
the
upper
positive
to the ground than is found in the typical
summertime thunderstorm in order to produce
suggests
relationship
shear in the vertical for a larger horizontal
slanted
charge
The thermal wind
positive
lightning,
and
probability of producing a positive flash will be
directly related to the closeness of the positive charge region to the
ground (Takeuti, 1985).
Since the main dipole is tied to temperature,
the winter's lower surface
dipole,
increasing
temperatures
would
yield
a
lower
main
the likelihood of producing positive lightning in
this season.
The tilted dipole hypothesis predicts a
lightning
for
relationship
positive
and
the
winter
between
the
the
environment.
likelihood
magnitude
of
the
of
preference
It
a
charge region.
positive
also suggests a direct
lightning
event
being
wind shear, and a relationship
between the probability of positive lightning and the
positive
of
height
of
the
Since lower temperatures are associated with
both lower charge regions and shallower clouds, it
is
expected
that
shallow clouds will have higher percentages of positive lightning than
deep clouds.
This does not necessarily mean
relationship
between
that
there
will
be
a
flash polarity and cloud height within the same
-
21 -
Introduction
group of thunderstorm cells.
Because of the prediction that
positive
flashes may be associated with the anvil cloud part, and this part may
be nearly as tall as the tallest cloud part,
direct
relationship
between
flash
may
there
polarity
and
not
cloud
be
the
top height
predicted by the inverted dipole hypothesis.
Many early observational studies of positive lightning considered
storms
had
that
only
a small number of positive flashes per storm.
These flashes were most often seen in the late stages
and
on
the
of
development
back side of the storm system (Simpson and Scrase, 1937;
Predominantly
Simpson and Robinson, 1941; Kuettner, 1950).
positive
storms were seen and documented in the winter season in Japan, Sweden,
and
the
United
percentage
States.
positive
of
The
positive
correlation
between
the
flashes to the strength of the vertical wind
shear (Brook, et al, 1982), and also percentage of positive flashes to
isotherm
-10 C
height
(Takeuti,
1985)
detection networks further documented that
have
a
much
higher
ratio
of
positive
were recognized.
stage
late
flashes
Lightning
thunderstorms
in the final hour
(Orville, et al, 1983) and showed that the positive flashes
are
more
spatially dispersed than the negative flashes (Rust, 1986; Orville, et
al, 1988).
Positive lightning was found to
be
associated
with
low
radar reflectivities and light precipitation (Rust, 1986).
Yearlong statistics compiled from SUNYA network data (Orville, et
al,
1987)
have
revealed an 80 percent increase in the percentage of
flashes with positive polarity in the winter (see
this
region
(area
figure
1.4).
For
shown in figure 1.8) in the months of January and
-
22 -
Introduction
February 1985, positive lightning dominated over
Figure
1.5
the
shows
actual
number
flashes (solid line),
of
clearly indicates that the total number of flashes
follows
the
same
seasonal
cycle.
lightning.
negative
Because
of
each
dominant
polarity
the number of negative
flashes drops below the number of positive flashes there is
in
and
a
switch
polarity for the winter months of January, February, and
March.
Another result from this work is that on the
average
the
first
stroke peak current in a positive stroke is significantly greater than
the first stroke peak current in a negative stroke (figure 1.5
lines).
The
negative
peak
dashed
currents ranged in strength from a mean
radiation value of 135 (30 kiloamperes) to 290 (65 kiloamperes) with a
value of 140
negative
(30 kiloamperes) during the time of the maximum number of
flashes.
kiloamperes)
to
The
410
positive
(90
currents
kiloamperes)
ranged
with
a
kiloamperes) during the time of maximum number
of
from
value
of
positive
185
(40
210
(45
flashes.
This indicates that positive peak currents were 1.5 times the negative
peak currents.
As the area covered by lightning detection networks increased,
new
phenomenon
cloud-to-ground
"bipole"
was seen.
flashes
(Orville,
et
The existence of mesoscale organization of
was
al,
documented
1988).
The
and
became
bipole
is
known
1988).
lower
positive
The centers of
as
defined
systematic polarity pattern in the strike points of lightning
that
a
the
as "a
flashes
and negative charge to ground" (Orville, et al,
the
strike
points
- 23 -
for
each
polarity
were
Introduction
as
defined
the
of contours of constant flash density.
centers
bipole length is the distance between the centers
its
and
direction
determined
is
connecting the negative center
the
by
to
positive
the
each
polarity,
between
the line
of
angle
The
center
North.
and
are known to occur year round, but are more likely to be seen
Bipoles
in fall and winter.
individual
cloud
kilometers.
wind.
Bipole
sizes
lengths
and
have
The bipole direction
Cloud
heights
are
substantially
been
is
than
observed between 60 and 200
aligned
with
the
geostrophic
are lower over the positive part of the bipole
compared to those over the negative part.
show
larger
observed
to
present.
In at least one case, a
little
Upper level winds have been
directional shear, although speed shear is
surface
front
has
separating the positive and negative charge centers.
been
observed
(Orville, et al,
1988).
The new data obtained from lightning detection
combined
with
past
work
about
the
electrical
thunderstorms to evaluate hypotheses for the
lightning.
dipole"
shallow
the
hypothesis
lightning
1987).
Both
in
the
"inverted
predict
winter,
a
as
dipole"
which
and
relationships
points
between
toward
the
of
the
further
production
-
processes
of
of
preference
for
positive
has yet to be investigated.
hypothesis also requires wind shear for
lightning
be
positive
positive
been documented (Orville, et al,
They also suggest an association
clouds
production
can
hypothesis and the "tilted
seasonal
has
networks
24 -
of
lightning
The tilted dipole
production
of
investigation
positive
and
positive
into
lightning
the
and
Introduction
environmental characteristics.
- 25 -
H
+4 + +++
++
+ +++++
S+
+ +
+
+_+
+
+
n+,
+
+ ,+
I.4
+
+
+
+
+7-
++
+
+
+
+
+
+
++
+
.
+
-,
+
.
+
+.
.
+
+
4.
+
.--
+
4
..
*. .
......
(
e
Figure 1.1:
Life-Cycle of a thunderstorm (Krehbiel, 1986).
+
t
R"
0
o
o
+
+
+
4
4.
,
-
P77
++
+
+
d
'7
4
e-
+
+
.
+
c
4
+ +
-
+,.
+
I
77-/
4.
-
S--+
+
b
a
++++ +.,'.,.+..,
. +"+
,.+ ++
77*7
Introduction
Figure 1.2:
Inverted Dipole
4
K
1-
Figure 1.3:
+
Tilted Dipole
- 27 -
Introduction
100
C,
U,
80
LL
.)
60
U,
0
CL
>,
40-C
_ 20--
June 84
Aug 84
Oct 84
Figure 1.4: Percent of lightning
positive charge as a function of
- 28 -
Dec84
Feb 85
Apr 85
flashes to ground lowering
month (Orville, et al, 1987).
Introduction
500
200
400
150
100
200 CD
z
~~1
50-
100
0
June 84
Aug 84
Oct 84
Dec 84
Feb 85
Apr 85
Negative Flash Summary
Figure 1.5a: The variation inthe number of flashes (solid line)
and median peak radiation field (dashed line) by month for negative
flashes (Orville, et al, 1987).
500
5000
S
4000
/ \
t
^
/
4-1 400
300
N'
3000,
=
C,
2000,
200
1000.
100
0
June 84
Aug 84 Oct 84
Dec84 Feb 85
Apr 85
Positive Flash Summary
Figure 1.5o:
As in 1.5a except for positive flashes.
- 29 -
:
The Libraries
Massachusetts Institute of Technology
Cambridge, Massachusetts 02139
Institute Archives and Special Collections
Room 14N-118
(617) 253-5688
This page is intentionally blank.
SC)
Introduction
1.3
Data Sources
A
dataset
lightning
to
investigate
should
have:
the
(1)
time,
cloud-to-ground lightning events;
the
earth's
surface;
phenomena.
location,
of
of
(2) electric field records taken
at
(3) radar
data
of
and
positive
polarity
lightning; and (4) sounding data of
mesoscale
characteristics
the storms producing the
sufficient
resolution
to
study
The data should cover several storms during the
winter when positive lightning is most
prevalent.
In
addition,
it
would be highly desirable to have corresponding data for summertime to
examine differences in the
production
different environmental conditions.
of
positive
lightning
The observations taken during the
Genesis of Atlantic Lows Experiment make up an excellent
studying
under
dataset
episodes of positive lightning in the winter.
for
Data taken at
Huntsville, Alabama for a field study of microbursts, provide a highly
useful summertime complement.
The Genesis of Atlantic Lows
designed
specifically
1986.
(GALE)
at
synoptic
and
in
cyclogenesis.
mesoscale
features
a
project
was
aimed
East
Coast
It
of
Data collection occurred between January 15 and
"Operations
Carolina.
March
15,
Central" was located at Raleigh-Durham Airport in
North Carolina, with a secondary operations center at
North
was
to gather an extensive database to study scale interactions,
mesoscale phenomena and their role
cyclones.
Experiment
Sounding
Cape
Hatteras,
stations in the main area of interest are
shown in figure 1.6, including locations of special sounding
set up for the experiment and approximate location of ships.
- 31 -
stations
Introduction
were
observations
GALE
radar,
aircraft,
from
obtained
soundings,
atmospheric
Soundings were taken
and boundary layer platforms.
at three hourly intervals by 39 National Weather Service (NWS)
17
special
land
At seven Cross-chain LORAN-C
sites, and two ships.
Atmospheric Sounding System (CLASS) sites,
90
of
deployed.
Aircraft used were the
P-3
(NOAA)
Administration
Atmospheric
minutes
National
Oceanic
taken
at
Atmospheric
and
National
Citation;
and
Sabreliner,
(NCAR) Electra,
Research
were
soundings
Over 300 dropwinsondes were also
or more.
intervals
sites,
for
Center
King
and
Air;
University
of Washington (UWA) C-131A (Convair); National Aeronautics
and
Administration
(NASA) Electra
of
(MIT)
Space
Institute
Technology
scanning
Doppler
and
Baron;
ER-2;
North
Massachusetts
radars;
scanning
MIT
Doppler
SPANDAR; UWA vertically pointing Doppler radar; five NWS
two
State
Carolina
Radar data was taken by the NCAR
University (NCSU) Cessna.
CP-4
and
CP-3
radar;
radars;
NASA
and
Boundary
radars on board the NOAA P-3, one of which was Doppler.
Layer Platforms including lightning direction finders,
and
PAM
stations,
The Program was organized around a series of Intensive
Observing
NOAA buoys, and C-MAN platforms were also used.
Periods
(IOP).
soundings and
observations
There
aircraft
were
thirteen
missions
were
IOP's
coordinated.
case
which
special
These
special
were primarily aimed at times of rapid cyclogenesis, and
accordingly were not necessarily coincident
bipole
during
studies
done
here.
discretion of the individual site.
Radar
with
data
the
of
the
were recorded at the
Figure 1.7 shows the
- 32 -
times
location
of
Introduction
the various radar sites.
This study concentrated on the land sounding data, MIT radar data
and
lightning
detection
Other data used included the GOES-6
data.
satellite pictures, NWS surface and upper air maps, and radar
A
charts.
summary
point and tipping bucket rain gauge from near the
corona
MIT radar were used to supplement the MIT
radar
data
in
the
small
scale analyses (see chapter 3).
The lightning data was taken by the State University of New
at
York
(SUNYA) lightning detection network, which covers the East
Albany
and Gulf Coasts of the
commercial
direction
Detection,
Tucson,
United
States
(see
figure
1.8).
It
uses
finders (manufactured by Lightning Location and
Arizona)
that
provide
information
on
time,
location, peak magnetic field and polarity of the first return stroke,
and also the number of strokes per flash.
The coverage available
for
this work was limited to the area north of 29N.
Some problems associated with lightning detection
efficiency,
detection
at
Detection
efficiency
least 70 percent within a 400 kilometer range of the high gain
direction finders
corrections
were
systematic
errors.
positive
are:
systematic angle errors, and the acceptance of
intra-cloud flashes as positive return strokes.
is
networks
(Orville,
applied
to
Current
cloud-to-ground
et
1987).
al,
the
SUNYA
studies
flashes
by
In
network
SUNYA
1985,
site
error
to compensate for
indicate
that
the
recorded by the network are indeed
cloud-to-ground events and not intra-cloud events.
- 33 -
Introduction
The summer dataset used in this study was
Alabama (see figure 1.9).
recorded
by
at
Huntsville,
The data were taken as part of a project by
the Federal Aviation Association (FAA)
were
taken
to
study
microbursts.
the MIT C-band radar and an FAA radar.
Data
Ten corona
point recorders covered the main area of interest, a circular zone
10
kilometer
radius
around
the radar site (see figure 1.10).
from the SUNYA lightning detection network
within
56
kilometers
of
radar data were recorded.
was
added
for
the
of
Data
area
the MIT radar, which is the area for which
Eighteen storms occurring in Huntsville
in
1987 were used in the study beginning on the following dates:
May 30
June 1
June 14
June 21
June 23
June 25
June 26
July 4
July 13
July 23
July 27
August 10
August 18
September
September
September
September
September
The storms producing the greatest number
began
on
June
21
and
August
10,
electrically active storms began on:
September
13.
In
addition
to
of
and
6
10
11
13
17
cloud-to-ground
a
group
of slightly less
June 25, July 4,
these
storms,
beginning on July 12, 1988 was also included.
- 34 -
flashes
data
July
from
13,
and
a storm
Introduction
.1
ALB
FN'T
":I
..
"......
'"
...... : , "
",
:!" " ..".I- :..
... 1
.-
"
K
I
1p
/~
"1
.so
I
.
..
...
.
":
AYS
.;.:
--
d
-.
..
• :.
'" ""*""
'
;
"""4 : ""
..
.....
;::..:.'-z
- - - --'- -:
j
BNA
.
"I~
'""
"',
i.'
.
-.
I
.
"
-- MYR
- -i
:~i''
35N
...
."
.
.'.
-
RVC':'
.
>S
1
."
•)
..
40N
'
::
.o"
a
.........
.........
I-~
DRY
51.
....
.....
..
-.. -
f
..
'
'..D
30N
80W
85W
75W
Figure 1.6: Inner and Regional GALE Area rawinsonde network.
Open circles are surrounding NWS sites, closed circles are NWS sites in
the Regional Area. Stars are VIZ sites, squares are GMD sites and
triangles are CLASS sites. The Inner Area extends from LHW-SRL in the
southwest to WA in the northeast. The Research Vessels Cape Hatteras
(GALE Field Program Summary, 1986).
and Endeavor are also shown
- 35
Introduction
: Wal lops
(SPANDAR radar)
\
I
I
//k'Range of
SPANDAR
radar
Roleigh
NORTH
../
CAROLINA
Ocracoke
"- -r..AR radar # .
.
.'Wilmingtol
SOUTH i
CROLIN
\
\
....-
NCAR
radar #/
CaI
ope Hatteras
(NCAR radar
I#
anl U. of Wash. radar)
Range of
... ./
_\
dual-Doppler
---- coverage
Fort Fisher
(MIT radar)
Georgetown
,-Range of
.
Roneof
RCAge
/-
radar #2
/
ATLANTIC OCEAN
SChar leston
'/
v Range of
MIT radar
0
I
100 km
I
Figure 1.7:
The GALE Doppler radar network. Circles indicate the
ground-based Doppler radars: NASA SPANDAR (SPA), NCAR CP-3, NCAR CP-4,
and MIT C-band (farthest south at Fort Fischer).
Range circles are at
200,
140, 140, and 140km for the respective radars. The University of
Washington TPA-11 is shown by a square (GALE Field Program Summary, 1986).
- 36 -
Introduction
SUNY-Albany
Figure 1.8:
and their area
communication)
Lightning Detection
Network
Locations of the SUNYA lightning detectors
of coverage during GALE (Orville, personal
- 37 -
Introduction
Figure 1.9: Location of MIT C-band radar in Huntsville Alabama.
Circle indicates range of observation which was 56 kilometers.
Figure 1.10: Location of corona point recorder sites for the
experiment in Huntsville. Site number 25 is located near the radar.
- 38 -
CHAPTER 2
Climatological Phenomena
2.1
Introduction
typical
This study begins with analyses aimed at identifying the
Data taken in
of lightning during the winter season.
characteristics
et
summer in Huntsville, Alabama, and annual results by Orville,
(1987)
respective datasets will be
The
comparison.
for
used
are
All
referred to as "Winter", "Summer", and "Yearly".
the
areas.
section
positive
This chapter
will
between
and
between
latitude is confirmed.
the
winter
from
divided
sections.
three
into
The
first
are
of
densities
flash
both
polarities.
Yearly and Winter observations will be explained
by the difference in the
relationship
are
examine the geographic variability in the percentage of
lightning
Differences
is
data
network, but for different times and
detection
lightning
SUNYA
al,
area
by
covered
percentage
of
the
flashes
two
that
datasets.
The
are positive and
Geographic distributions of flash densities in
described.
These analyses provide evidence that the
Gulf Stream enhances cloud-to-ground lightning activity.
- 39 -
Climatological Phenomena
of
is
not
the
currents
It will show that the geographic pattern in peak
flashes
patterns
to
related
are
data
Winter
the
in
mean number of
or
density
flash
of
It will show that although nearly
thunderstorm days.
of
lightning flashes (hereafter referred to as "peak
cloud-to-ground
current").
currents
peak
The second section will consider the first stroke
percent
twenty
positive, the net charge
transfer to the earth over areas of 500 square kilometers or larger is
negative in all areas under coverage.
The
activity.
result
basic
lightning
in
cycle
diurnal
a
The third section will look for
was that no prominent cycle was found.
(1984),
This agrees with previous results by Shands (1948) and Colman
which also found no diurnal cycle in winter thunderstorm development.
was
During GALE there
through
continued
the
storm
a
that
Such a
highest flashrates and some of the highest peak currents.
the analyses done in this chapter were all
this
introduced
by
storm
this
to
assess
storm.
the
overall
The
performed
effects
of
could
and
For this reason,
any analysis done on the whole Winter dataset.
without
and
It was also responsible for
storm might not be representative of typical winter storms
bias
13
March
on
end of the experiment, producing more than 25
percent of the lightning flashes in GALE.
the
began
with
both
and
bias that could be
percentage
positive
of
lightning during this storm period was only slightly lower than during
the rest of GALE; 14 percent for this storm compared with
overall.
17
percent
Because the results without this storm were nearly identical
to results obtained for all of GALE, only results based on all of GALE
-
40 -
Climatological Phenomena
will be presented.
2.2
Geographic Variability
This section will consider the geographic variability in (1) the
of
percentage
positive
It will be
polarity.
polarity
positive
lightning
that
shown
increases
and (2) the flash density of each
the
of
percentage
with
flashes
with increasing latitude, and that this
relationship explains the difference in overall percentage of positive
between
flashes
Winter
the
This section will
and Yearly datasets.
show geographic distributions of flash density, which
the
the Gulf Stream in the winter is to enhance lightning
of
effect
that
indicates
activity.
The Winter data includes roughly 116,000
percent
are
positive.
flashes.
percent positive.
2
flashes
in
figure
1.4
in
the Introduction.
increase to 80 percent positive in the month
winter
season,
the
(80
which
are
in
1987,
The figure indicates an
of
February.
For
the
average Yearly percentage of positive flashes is
estimated as the mean of the percentages
February
17
A seasonal dependence in the percentage of
positive flashes detected was documented by Orville, et al,
shown
this
The Summer data has 6,000 flashes which are 2
percent positive, and the Yearly data has 738,000
also
Of
percent),
and
March
(40
for
January
percent).
(55
percent),
This mean is 60
percent, which is more than three times the percentage observed in the
Winter
data.
The first task of this chapter will be to explain this
discrepancy.
-
41 -
Climatological Phenomena
The main difference between the observations in
The coverage
observations is the area covered by the network.
Yearly
for the Yearly data is shown in figure 2.1a.
and
Winter
the
Winter
Coverage for the
data is the GALE area shown in figure 1.6, and for the Summer data the
Figure 2.1b shows
Huntsville area shown in figure 1.9.
of
the
three
of
areas
Winter area is not given a
The
coverage.
comparison
a
as
the network coverage.
The Yearly data is centered to the north of
The following analyses will show
the center of the Winter data.
the
that
positive flashes between the
of
percentage
the
in
difference
north
far
northern boundary since the observations did not extend as
Winter and Yearly data is due to the difference in the area covered by
the lightning detection network at these times.
Both the tilted dipole and the inverted dipole hypotheses predict
that
for
the percentage of flashes with positive polarity should increase
shallower
percentage
surface
latitude.
of
This
positive
results
in
temperature
the
should
flashes
tied
closely
is
depth
Climatologically,
temperature.
increasing
Cloud
clouds.
to
decreases with
that
expectation
increase
latitude, and is also consistent with previous results
surface
the
increasing
with
that
indicate
that positive flashes are relatively more prevalent in colder climatic
conditions (Orville, et al, 1987).
Figure
predicted
2.2a
trend
with latitude.
shows
that
the
Winter
observations
match
the
of an increase in the percentage of positive flashes
This confirms previous results and supports
tilted dipole and the inverted dipole hypotheses.
-
42 -
both
the
It is believed that
Climatological Phenomena
latitudinal
in
difference
the
is
coverage
the
that
reason
the
of positive flashes in the Yearly data is three times that
percentage
of the Winter data.
polarity
In order to justify that the relationship between flash
responsible for this difference, it is desirable to
is
latitude
and
Therefore,
quantify the magnitude of the effect of this relationship.
a
of total number of flashes with latitude will be used to
breakdown
estimate a mean latitude for the Winter data.
be
will
data
Yearly
an
in
estimated
A mean latitude for the
analogous
manner
and
the
3AN
the
percentages of positive lightning at these latitudes compared.
is
Even though the central latitude of the Winter data
further
is
latitude
mean
Figure 2.2b shows that the total
south.
number of flashes decreases with increasing latitude.
The Winter data
be dominated by the characteristics of flashes occurring between
will
30 and 33N, since more than 50 percent of the flashes occurred in this
latitude
belt.
Figure 2.2a indicates that in the range of 30 to 3N,
15 percent of the flashes were positive, which is nearly the
the
overall percentage of 17 percent.
38N.
through
that
This implies
the
Yearly
data
the
of
months
of
March should contain a higher percentage of positive
flashes than the Winter data, by as much as three times.
dependence
In
38N is 45 percent positive, three times greater than the
data
overall percentage.
January
as
By similar analogy, the Yearly
data is dominated by flashes occurring in the range of
Winter
same
polarity
of
flash
The latitude
predicts a percentage of positive
flashes in the Yearly months of January through March three times that
- 43 -
Climatological Phenomena
observed
in
the
Winter
data,
which
is
approximately
the actual
difference.
The latitude difference is not the only cause for a difference in
Winter and Yearly coverage areas is
flashes
over
Another difference between the
positive flashes.
of
percentage
the
the
warm
Gulf
that
the
Winter
data
Stream waters, whereas the Yearly data
coverage includes only a small portion of the Gulf Stream.
Stream
The
Gulf
location as represented by sea surface temperature is shown in
The outline represents an area of negative
figure 2.3.
density
point
includes
strike
greater than 100 flashes/square kilometer over the
of
GALE time period.
flash
temperature
The contours are of sea surface
means
for the month of February 1986, and are drawn at one degree intervals.
Sea surface temperature values are also shown.
the
Gulf
Stream
What is the effect
the production of lightning?
on
of
In the winter the
Gulf Stream is warmer than the land, so its effect would tend to be to
statically
de-stabilize
the lower atmosphere.
to increasing convection and lightning activity.
of
the
Gulf
This would contribute
The expected
effect
Stream is therefore an enhancement of the production of
lightning.
The effect of the Gulf Stream can most easily be seen
2.3,
which
in
figure
shows the correlation between the sea surface temperature
(thin lines) and the
negative
flash
relationship presented is clear.
densities
(heavy
line).
The
The warm sea surface temperatures of
the Gulf Stream yield high flash densities that will be shown to reach
twice
those
detected
over land.
It will be shown that these higher
-
44 -
Climatological Phenomena
positive
lightning.
relatively
with
flash densities are associated
of
percentages
low
Thus, the effect of the Gulf Stream is to lower
Again,
the overall percentage of positive flashes in the Winter data.
location differences between the Winter and Yearly data predict higher
percentages of positive flashes in the Yearly
compared
data
the
to
Winter data, which is the observed relation.
This section will next consider how typical
data
the
this
for
winter is compared to a winter-time climatology (Kessler, Thunderstorm
distribution
will
It
1983).
Dynamics,
Morphology and
the
compare
geographic
of flash densities of both polarities to the mean number
February,
of
of thunderstorm days for the month
averaged
the
over
If the Winter observations are typical, a
years of 1956 through 1975.
similarity between the patterns of flash density and
mean
number
of
thunderstorm days will be seen.
The negative flash density distribution is shown in figure
The
contours
are
in
drawn
logarithmic
intervals.
The outermost
contour for both polarities is 10 flashes/square kilometer.
negative
flash
densities,
contours
dashed
lines.
flashes/square
striped.
southern end.
the
kilometer
in
The area of greater than 1000 flashes/square kilometer
is dotted, and the area of greater than 100
is
For
100 and 1000 flashes/square
of
kilometer are drawn in solid lines, 500
2.4a.
The
area
of
flashes/square
kilometer
flashes was artificially cut at 29N, the
The main area of negative flash densities (greater than
100 flashes/square kilometer) forms a "V" shape in the southern end of
the detection area.
The pronounced maximum in the eastern
-
45 -
branch
of
Climatological Phenomena
Note that the
the "V" is coincident with the Gulf Stream (see above).
flash densities in this eastern part of the "V" are twice
in the western part which is over land.
those
seen
There is also an area of high
density of negative flashes in eastern North Carolina, extending
into
Virginia, surrounded by lower flash densities, and a few other smaller
areas of flash densities greater than 100 flashes/square kilometer.
The positive flash densities are more than an order of
less
than the negative flash densities, with no area reaching as high
as 100 flashes/square kilometer (figure 2.4b).
25
striped
(enclosing
(enclosing hatched area).
than
25
and
area),
50
The main area
kilometer)
flashes/square
flashes/square
flash
of
is
The contours are:
"V"
farther
extending
flashes/square
Virginia,
which
overlaps
in
North
eastern
with
10,
kilometer
(greater
similar to the
the
"V",
negative
There is a minimum (less than 10
northward.
kilometer)
density
shaped,
negative pattern, but centered one degree east of
and
magnitude
an
Carolina
and
middle
area of negative flash densities
greater than 100 flashes/square kilometer.
There is also a minimum in
positive flash densities in northwest South Carolina.
A comparison with the mean number of thunderstorm
month
land
available).
high
for
the
of February is used to support the idea that the Winter data is
typical of the winter season.
over
days
values
is
shown
in
The mean number
figure
of
thunderstorm
days
2.5 (values over the ocean were not
The pattern shows a maximum in Louisiana, and a tongue of
that
pushes northward and eastward up through Kentucky.
The Virginia - North Carolina area is
covered
- 46 -
by
a
tongue
of
low
Climatological Phenomena
values.
The "V" shaped area of high flash densities is located to the
east of the tongue of high numbers of thunderstorm days
five
degrees
of
by
at
least
It is believed that because the network
longitude.
detection area did not extend far enough to the west, the total number
of
of
If the east-west offset is ignored for the tongue
detected.
values,
Arkansas, and Missouri was not
Louisiana,
in
occurring
flashes
the
mean number of thunderstorm days and flash
of
patterns
high
densities are similar.
The exception to this similarity is the area of negative
of
greater than 100 flashes/square kilometer in Virginia and
density
North Carolina.
A possible explanation for this area is the effect of
the topography (shown in figure 2.6).
extent
north-south
As in the
is
convection
the
of
area
as
seen
the
of
density is to the southeast
Survey).
flashes
of
an
500
the
The area of high negative flash
centered
mountains,
meter
contour
(U.
Gulf
Stream,
an
increase
in
S.
along
the
Geological
enhancement
to
the electrical activity of
thunderstorms, primarily affecting the production of negative flashes.
The
lightning
of
effectiveness
networks
detection
has
not
yet
perfect.
likelihood of a flash being detected is related to
the
magnitude
studied
the
peak
and
revealed
current.
that
the
system
flashes.
The
of
Since positive flashes have higher peak currents,
their probability of detection at long
negative
is
been
This
in
results
ranges
the
is
greater
than
for
edges of the network area
appearing to be highly positive with large peak
currents.
The
peak
current distribution is shown in figure 2.9 and does exhibit a maximum
-
47 -
Climatological Phenomena
in peak currents around the edges.
section
2.3
first
on
stroke
More discussion will be
The increase in the
currents.
peak
in
given
percentage of positive flashes at the network edges
can
be
seen
in
figure 2.7.
Figure 2.7 shows the percentage of flashes with positive polarity
in
a
one
The dotted
latitude by one degree longitude grid.
degree
regions are less than 5 percent positive, hatched are 5 to 10
and
positive,
striped are 10 to 25 percent positive.
and 75 percent are also drawn.
are
an
order
the pattern in
densities,
Because the negative
percent
Contours of 50
flash
densities
of magnitude larger than the positive flash densities,
2.7
figure
very
closely.
2.4a,
negative
flash
resembles
figure
As
above, the areas of convective
noted
enhancement by the mountains and Gulf Stream are both seen as areas of
low percentages of positive flashes.
In
the
percentage
valley
Mississippi
there
is
a
probability
of
a
flash
being
in
can
detect
flashes.
detected is related to the peak
Since
current of the flash (Beasley, 1985).
higher
gradient
of positive flashes and this is attributed to the location
being at the outer reaches of where the network
The
strong
positive
flashes
have
peak currents than negative flashes (by as much as a factor of
1.5, see section 2.3), positive flashes are more likely to be detected
by
the network.
This would cause a bias toward more positive flashes
at the boundaries of the network.
It has also been suggested that the
polarity of the flash may not be correctly resolved at large distances
(Brook, 1988),
such
that
negative
flashes
- 48 -
would
be
reported
as
Climatological Phenomena
This bias in favor of positive flashes is clear at both the
positive.
The southern edge was artificially cut
eastern and western edges.
29N,
which
is still well within the range of the network, so no bias
The northern edge of lightning strike points
would be expected.
not
reach
at
northern
the
edge
of
does
network, so no bias would be
the
expected here.
The gradient in the percentage of positive flashes at the western
edge
is
The smoothing of the
much greater than at the eastern edge.
Since the effect of
gradient is due to the effect of the Gulf Stream.
the
Gulf
Stream is a decrease in the percentage of positive flashes,
this would temper the effects of the bias
edge
of
the network.
produced
by
reaching
The result is that the gradient on the eastern
edge is not as concentrated as the gradient on the western edge.
bias
in
detection
negates
the
only
two
lightning dominates.
southeastward
into
regions
where
Since
eastern Tennessee, the region will
Tennessee
region.
it
can
The
second
southward into Virginia, and
will
There
are,
said that positive
be
The first region extends from
Georgia.
This
ability of the network to determine
whether or not positive lightning dominates at the edges.
therefore,
the
western
Kentucky
the center of this region is in
be
referred
region
be
Pennsylvania region.
- 49 -
to
from
extends
referred
as
to
as
the
eastern
Pennsylvania
the
western
Climatological Phenomena
between
coverage
of
area
This section has shown that the difference in
the Winter data area and the Yearly data area would cause the
of
dependence
latitude
difference in the percentages of
which is present in the Winter data is
which
lightning,
positive
associated
the
is
The Gulf Stream,
difference (60 percent versus 17 percent).
observed
greater
a factor of three
predicts
polarity
flash
The
lightning.
positive
Yearly data to have a higher percentage of
area
an
with
of
500 negative flashes/square kilometer which is a factor
than
of two larger than the negative flash densities observed over land and
an
magnitude
of
order
observed.
then
greater
the
positive flash densities
percentage
The Gulf Stream's effect is to decrease the
of
positive lightning in the Winter data relative to the Yearly data.
In
to
explaining
the
difference
addition
responsible
for
associated
low
with
flashes/square
(approximately
kilometer)
25
in percentage of positive lightning
the
that
this section has documented
average
flashes/square
Carolina - Virginia region is
eastern
flash
negative
and
are
differences
location
how
is
than
10
flash
positive
with
region
(less
densities
kilometer).
associated
Tennessee
eastern
North
negative
flash
The
high
densities
densities (greater than 100 flashes/square kilometer) and low positive
flash densities
(less than 10 flashes/square kilometer),
effect of the mountains to the west.
- 50 -
due
to
the
Climatological Phenomena
Figure 2.1a: Spatial coverage of the SUNYA lightning
detection network in 1985 (Orville, et al, 1987).
SUNYA
by the
area of coverage
of
Comparison
Figure 2.1b:
datasets.
Yearly
and
lightning detection network for the Summer, Winter
- 51 -
Climatological Phenomena
Percentage of Positive Flashes versus
Latitude
80
55
'50
-45
s4o
I-
40
IL
U
-i
c3s
a
U30
a
(L2S
20
15
10
3
30
31
32
33
34
36
35
La titude
37
38
39
40
41
4
Figure 2.2a Number of positive flashes observed within a one degree
latitude belt divided by the total number of flashes observed within
that same belt for the entire GALE time period.
Number of Flashes versus Latitude
2400
2
220
2100
20
1900
1800
a 1700
01600
_j1400
-
1300
0
LI1
,
39000
7000
2S1
,33
34
35
36
Lat tude
37
38
39
40
41
Figure 2.2b Total number of flashes of either polarity observed
a one degree latitude belt for the entire GALE time period.
52 -
within
Climatological Phenomena
/L
"
.-
4
5/ .X,
-
2
a
S..
, , C .20I
6,-,
/
.1 20.
.414 19 6
39.
19
9,A
49.8
3.5 1.6 1.3
" 5"33.8
19.6
0.
3.4
9
s
8
sI
.
s9.
S.
I
1.5 15.3 19.5 19.7 39.
19.1
1s 1 1.3 1.4 .
.5 1
5
19.4
.5 18 2
.1.2.18.4.18.1
8.3
186.1
.I3. 1
.
,
I. 0"
0 l 6 8.3 18.2 3.4
9.4 19.5 19.4 9.
19.6 19.9 19.6
40N
.
8
to
2 1 .0 18.6 10.4 18.4 16.4 18.4
35N
I
33i
t
6 23.0 22.4
ZX2
2.8 22.
L 23.2
1-
i
.
3.2
6 2
S
22.
0
.322
3.8 23.\
3.
23.
23.5s9i
-
.
1"
2.
20
139.819.7 3.5 13.5 13.6 13.4 3 .9 38.8 38.5 18.3
20.0 20.1 20.1 2
9.6
20
208
.7 2.2 22.2
23.7
4
4.4
.19.8
s9
.6 .1.7 19.8 3
.
0
.
.
3
a
20.7
.
-
l
23.
1.
2
2
.4 2.11
30 N
2i2.,
23.3 231.
3.2 23 3 22.5 22.6 22.5 22.5 22 4 22.7 22 9
3.3 23. 23.1
23.3 23.7 2
23.4
. 2
.3 22.5 22.6
21.6 i-
22 5 22
.5 2
.21.
1
3.19
I3s.SIs9.7
.9 20.6 20.4 20.3 20. 1 20.5 20.6 20.7 20.6
29
20.6 20. 1 20.5
9 21.6 21
23.4.23.9 23.G.
.
9
.52.2 213 21
21
2-3.9 23.2 23.5 23.9 23.0 23.5 23.1
-
20 2 20 4
21.8 21.7 21.4 21.
22.322.
23.2
19.6 19.
20.3 20.1 20.4 20.3 20.4 20.3 20.3 20.2 2
21.4 21.
22 4 22.4 22.4
23.
3.1
21
.8 2
1.8
22.8 4.G 21
1 20
0.7 20.1
.0
Z?23
-.
I3
.
22.
9
.9 23.6 23.4 23.4 23.5 2.23.7
24.6 24.
,
4.0
2
24
.
0 22.9
23.5 23.4
23.8
25N
25N
Figure 2.3 Sea surface temperatures and contours drawn for every degree
Celsius, with an outline of the area of negative flash densities
greater than 100 flashes/square kilometer during the GALE time period.
- 53 -
Climatological Phenomena
37
38
J
-C
z
34
33
0
32
31
30
29
96
94
92
90
88
86
84
82 80
78
Degrees West Longitude
76
74
72
.70
68
Figure 2.4a
Negative flash densities per square kilometer for
the entire GALE timeperiod. Striped area indicates flash densities of
greater than 100 flashes per square kilometer, dashed lines are drawn
at 500 flashes per square kilometer, and dotted area is greater than
1000 flashes per square kilometer.
41
40
39
38
37
36
S35
o
z
U
)
o
0
34
33
30129
96
94
92
90
88
86
84
82
80
78
Degrees West Longitude
76
74
72
70
68
Figure 2.4b: Positive flash densities per square kilometer for
Striped area indicates flash densities
the entire GALE timeperiod.
between 25 and 50 flashes per square kilometer, and hatched area is
greater than 50 flashes per square kilometer.
- 54 -
Climatological Phenomena
00
37
-j
36
o0
Z 35
I
01.0
,34
33
3231
300
29
96
.
94
92
I20
90
88
86
84
Degrees
82
80
78
West Longitude
76
74
72
70
68
Figure 2.5: Mean number of thunderstorm days in February
(Kessler, 1986).
for the years from 1951 through 1975
#0
-O
40
39 38 S37
0
36 -
, 34
4
33
32
31
29
96
___
94
92
90
88
86
84
Degrees
82
West
80
78
76
74
72
70
68
Longitude
Figure 2.6: Topographic map of the eastern United States, with
contours drawn at 200 and 500 meters. (U.S. Geological Survey).
-
55 -
V
37
-
36
- 35
0
Z
t, 34
a,
29 '96
'
I
94
I
I
92
11
-II
90
,
88
i
i
m.Ir
86
84
82
1
80
I8 s -
1
78
IV
76
I
74
I
a
72
I
rI
70
Degrees West Longitude
Figure 2.7:
Flashes Striped
Geographic Distribution of the Percentage of Positive
area indicates between ten and twenty-five percent
positive, hatched area is between five and ten percent
dotted area is less than five percent positive.
positive,
and
68
Climatological Phenomena
2.3
First Stroke Peak Currents
The relationship between the cloud's electrical structure and the
peak
current
of
well understood.
currents
the
No theoretical argument for the differences in
between
flashes
scientific community.
related
to
first stroke of a cloud-to-ground flash is not
has
been
accepted
It is here suggested
environmental
by
that
peak
a majority of the
peak
currents
are
conditions (affecting cloud structure) and
not factors such as topography.
Past work by Orville, et al, (1987) reported that the mean of the
positive
peak
currents is substantially greater than the mean of the
negative peak currents for the Yearly data.
the
This data is dominated by
summer environment, more than 80 percent of the flashes occurring
during the months of June, July, August, and September.
This section will confirm the finding by
positive
currents.
peak
currents
et
to
contribute
Winter and Yearly data.
been
identified
to
the
extent
that
not
be
differences in peak currents between the
Since areas of predominantly positive flashes
(see
section
2.2),
an estimate of a quantity
directly proportional to net current per unit area
determine
al,
of greater magnitude than negative peak
It will be shown that geographic variability would
expected
have
are
Orville,
to
which
these
significant.
-
57 -
will
anomalous
be
areas
made
to
may
be
Climatological Phenomena
For the Yearly data (presented by
positive
peak
currents
et
Winter positive flashes also
times the negative peak current average.
believed
that
conditions
environmental
difference of the Summer data, the
low
number
averaged
Summer positive flashes
While
show the same average peak current as the negative flashes.
is
the
1987),
al,
averaged 1.6 times the negative peak current
average (see table 2.1 below).
1.6
Orville,
it
are responsible for the
of
positive
flashes
available for this analysis reduces the confidence in this belief.
Data
Figure Source
Flash
Number of Flashes
Polarity (thousands)
2.8
2.8
2.8
2.8
2.8
2.8
positive
negative
positive
negative
positive
negative
a
b
c
d
e*
f
Yearly
Yearly
Winter
Winter
Summer
Summer
Mean Peak
Current (camp)
Mode Peak
Current (kAmp)
55
35
90
55
30*
30
18
720
20
96
0.1
6
30
25
60**
33
5
17
Table 2.1: Number of flashes and average first stroke peak
currents for the Yearly, Winter and Summer observations. *Note that
Summer positive flashes are only 100 in number, which may not be
sufficient for a reliable estimate of peak current. **Note that the
mode for the Winter observations was taken to be midway between the
two peak current values having 1420 flashes each.
Although the ratio of the averages of the positive peak
to
the
negative
currents
peak currents is the same for the Yearly and Winter
observations, the actual peak currents are approximately 1.5 times
large in the Winter observations than in the Yearly observations.
number of flashes observed is related to peak current in
The
mean
as
The
figure
2.8.
and mode of the peak currents are marked with arrows.
Note
that the mode for the Winter
observations
- 58 -
was
taken
to
be
midway
Climatological Phenomena
between
two
the
current values having 1420 flashes each.
peak
The
analyses show that in all cases (1) the mode is less than the mean and
(2) the
difference
between the mode and the mean is greater for the
positive flashes than for the negative flashes.
described
This
could
also
be
the positive peak current distribution having more of a
as
tail of high peak
values
current
than
the
negative
peak
current
distribution.
variation
One aim of this section is to eliminate geographic
current
peak
as
explanation
an
currents average 1.5
Toward
this
aim,
times
the
Yearly
the
geographic
of
maps
of
fact that the Winter peak
and
(figures
peak
Summer
first
stroke
currents.
peak currents
longitude
averaged over a one degree latitude by one degree
presented
in
box
are
The positive peak currents (figure
2.9 a and b).
2.9b) do not make a simple pattern, but the
centers
two
of
minimum
(less than 60 kiloamperes) cover the North Carolina - Virginia border.
The maximum is at about 39.25N, 72.0W, off the coast
(Recall
that
of
New
Jersey.
high current values at the edges of the network are due
to network detection biases; see section 2.2.)
The negative first stroke peak currents were analyzed in the same
manner.
The
results
(figure
2.9a)
present
a
smoother
picture,
ostensibly because there are five times as many observations.
clear
that
there
is
Carolina and Virginia.
this
area.
a
It
is
in negative peak currents in North
minimum
The peak currents increase radially away
from
This predicts higher negative peak currents in the Yearly
data compared to the Winter data.
This is not what is observed.
-
59 -
The
Climatological Phenomena
mean
of
the
Yearly
data
negative
peak currents is 35 kiloamperes
compared to the Winter average of 55
factors
kiloamperes.
Since
geographic
are not responsible for the observed relationship in negative
flashes and there is not evidence against
positive
flashes,
(including, but
it
not
environmental
factors
is hypothesized that environmental differences
limited
to
season)
are
differences in first stroke peak currents.
responsible
through
its
control
of
the
for
the
The picture presented here
is one in which the environment influences the magnitude of
current,
for
thunderstorm
the
structure
peak
and
development.
It
has
different
been
from
shown
that
summertime
lightning (Orville, et al,
percent
of
all
winter
lightning
1987,
lightning
lightning
in
section
is
significantly
the percentage of positive
2.2).
Since
Models
100
occurs in the summer, past work involving
lightning has been aimed at summertime lightning which is
negative.
almost
98
percent
of atmospheric electricity fields are based on the
concept of the normal thunderstorm where negative charge is lowered to
the
earth's
surface
and
positive
charge
question this study will attempt to answer is:
positive
cloud-to-ground
results of models that do
lightning
not
in
include
lightning?
- 60 -
the
the
is left in the air.
Is
winter
there
The
sufficient
to question the
production
of
positive
Climatological Phenomena
In order to answer this
quantity
believed
question
an
estimate
was
made
Areas of net positive charge transfer could be
from
quantity
to
identified
the magnitude of charge and areal
determine
coverage of the positive region.
about
a
to be directly related to total charge transferred
to ground.
this
of
The following assumptions were
made
net charge transferred by cloud-to-ground lightning in the
the
estimation:
(1) it is directly related to first stroke peak current
(2) it is directly related to number of strokes per flash
which leads to:
(3) net charge transfer is directly related to the product of first stroke
peak current and the number of strokes.
A one degree latitude by one degree longitude
and
the
was
defined
sum of the products of peak current and number of strokes in
each box was plotted.
noted
grid
before,
The results are contoured in figure
2.10.
As
artificial areas of positive charge transfer are to be
expected at the edges of the data region due to detection inaccuracies
at
large
distances.
There
are
two
areas where by this estimate,
positive charge was transferred to ground during the GALE
The
areas
are
the
predominantly positive
same
as
flashes.
those
noted
These
- 61 -
are
experiment.
in section 2.2 as having
the
eastern
Tennessee
Climatological Phenomena
region
and
the
western
Pennsylvania
square kilometers in size.
these
areas
is
region,
each of which is 500
The magnitude of net
charge
transfer
in
an order of magnitude less than the land maxima, and
therefore would not be significant for analyses
averaged
over
areas
first
stroke
peak
larger than 500 square kilometers.
The
current
into
investigation
has tried to determine the effect of environmental factors on
lightning production in general.
also
affecting
factors
It has
been
established
here
and
in previous literature (Orville, et al, 1987) that positive peak
currents are larger than negative peak currents.
Do these larger peak
current flashes occur at the same rate as lower peak current flashes?
Figure
flashrate.
2.11
shows
a
graph
of
average
were separated into one hour bins and the number of
minutes yielded the flashrate, and
was
flashes/minute).
flashes of a given polarity occurring at
summed
and
divided
flashes
of
each
The total number of flashes divided by sixty
polarity was determined.
(0.1667
versus
All the Winter flashes
The graph was created as follows.
flashes/hour
current
peak
rounded
All
a
to
the
flashrate
specific
10
currents for
peak
the
nearest
were
by the number of flashes in the same category to
give the "average peak current"
value
used
as
the
ordinate.
The
results of this analysis show that positive first stroke peak currents
decrease with increasing flashrate where
entire
storm
system,
and
the
is
flashrate
averaged over an hour.
over
The peak currents
ranged from 97 kiloamperes at 0.2 flashes/minute to 73 kiloamperes
2.6
flashes/minute.
No
positive
flashes
- 62 -
an
occurred
at
at
storm-wide
Climatological Phenomena
averaged
flashrates greater than 2.6 flashes/minute
in
change
The
above
percent
hour.
negative first stroke peak currents was significantly
less than for that of positive peak currents.
10
one
over
The initial values were
mean value, decreasing to 3 percent below the
the
mean value between 6 and 7 flashes/minute.
The analysis presented suggests
lightning should emphasize low flashrates throughout a storm
positive
system for flashrates sustained for an hour.
of
producing
for
mechanisms
that
production
As the storm
more flashes/minute), the mean
positive lightning increases (i.e.
of first stroke peak currents of those flashes should decrease.
Other results of this section confirm that positive peak currents
are larger in magnitude than negative peak currents, and that there is
Peak currents in the Winter
a factor of 1.5 difference.
were
times
1.5
the
observed peak currents.
Yearly
in
distribution of peak current is similar
minimum
in
the
currents
peak
the Yearly data.
peak
currents
electrical
a
area, and increasing
why
averaged 1.5 times the average peak current in
in
This suggests that the difference
the
between
activity,
with
The geographic pattern does not explain
magnitudes on all sides.
Winter
The geographic
polarities
both
Virginia
and
Carolina
North
observations
is
Winter
tied
to
and
the
structure and development.
- 63 -
Yearly
magnitude
flashes,
environment
i.e.
through
of
the
cloud
Climatological Phenomena
Peak Current (kA)
0
200
100
I
L
400
300
~&
I
100 I
I
p
p
P
50+
t I I I II -I II
I II II
III
I
i
I
i
-
i
I
i
i
i
I
Figure 2.8a: Number of flashes of positive polarity observed
within the region of the lightning detection network in the Yearly
dataset (Orville, et al, 1987). Mode (left arrow) and mean (right arrow
of the distribution are marked.
Peak Current (kA)
400
300
200
100
0
1000-
D 500
E
0
l
ibIIlili
1
i
1
;
1
1
;,
Figure 2.8b: As in 2.8a except for negative flashes
(Orville, et al, 1987).
- 64 -
Climatological Phenomena
Nu
Ir-M
-- w
ber of
Positive
I I 1''''I''T'I 177 ri ~r rjrrrrrr
Flashes
versus
Average
r~rr rr II rrrlrrrr(rrr rl r r r rl r r rr I ( r Ti-
Peak Current
I II1
'
11'
I
1" '
1400
1300
1200
1100
£1000
908
Lam800
o
700
.
608
z 50 r
C
400
300
200
100
2
0
%Th7
25
50
75
Figure 2.8c:
12;VinR¢
-w"
100 125 150 175 200 225 250 275 300 325 350 375 480 425 45(
Average Peak Current (kiloamperes)
As in 2.8a except for the Winter dataset.
NJuI ber of Negative Flashes versus Average Peak Curren..
. ....
11i0
1000
S000
7000
4-
o 6000
O 5000
S4000
3000
1000
0
100 125 150 175 200 225 250 275 300 325 350 375 400 425 450
Average Peak Current (ki loamperes)
Figure 2.8d:
As in 2.8c except for negative flashes.
65 -
Climatological Phenomena
34
Number of Positive Flashes versus Average Peak
,1,
1 I . . . . . I I
. . . .
Current
I
32
30
28
-I
- 24
o20
"*.
L16
14
=12
10
8
6
4
2
0
0
25
50
75
100 125 150 175 200 225 250 275 300 325 350 375 400 425 450
Average Peak Current (kiloomperes)
Figure 2.8e:
As in 2.8a except for the Summer dataset.
Number of
N
Negative
Flashes
versus Average
Peak
Current
,,,
,111
11,, ,J
,11111,]
1000
900
o
,800
o
400
300
/1
25
I -". ,
50
It
Itl ltI Ili tII..li1
I,,I,, , , ,,11111
it ... i I
75 100 125 150 175 200 225 250 275 300 325 350 375 408 425 450
Average Peak Current (kiloamperes)
Figure 2.8f:
As in 2.8e except for negative flashes.
66 -
Climatological Phenomena
37
, 36
Z
35
z
34
33
a,
31
96
94
90
92
88
78
80
82
84
Degrees West Longitude
86
76
74
72
70
68
Figure 2.9a: Average first stroke peak current for
negative flashes in kiloamperes for the Winter dataset.
6
o
31
-J
Z
31
a)
a,
3:
0
120
.1
96
94
92
90
Figure 2.9b:
88
86
84
82
80
78
Degrees West Longitude
76
74
f.j
72
70
68
As in 2.9a except for positive flashes.
- 67 -
Climatological Phenomena
2
Net Current per KM
Figure 2.10:
Estimate of net current per unit area.
Average
Peak
Current versus
Storm-wide
Flashrate
SS
9100
aesso
85
L
L
S80
S
75
aoaiS
S 70
V
* 60
0
1
2
3
4
Flashes per
S
'6
7
8
9
10
11
Minute (over whole Gale area)
12
Figure 2.11 Average first stroke peak current versus flashrate. Dashed
line with triangles represents positive flashes, solid line with boxes
represents negative flashes. Solid horizontal lines mark the mean
first stroke peak current for each polarity.
-
68 -
Climatological Phenomena
2.4
Diurnal Effects
One change in the environment occurs every
the
sun
day.
Every
morning
rises and heats the earth's surface, and every night it sets
allowing the earth's surface to cool
heating
of
the
earth
tends
the
planetary
general,
the
to de-stabilize the planetary boundary
layer, increasing the likelihood of
stabilizes
In
radiatively.
The
convection.
night
cooling
layer, and is often related to a
boundary
What is the diurnal effect on the production
decay in the convection.
of lightning during the winter environment?
for
The diurnal variation can be looked
positive
2.12a
Figure
flashes.
flashes for each hour of the day.
Eastern
The
UTC.
Zone,
Time
solid
All
line
with
70 and
between
capacity than vegetation
expected
to
be
as
boxes
located
data
90W.
and
large
Since
soil,
the
represents
is
in
the
Winter
flashes
water has a much higher heat
diurnal
effects
would
not
be
in magnitude over the ocean as over land.
Therefore, the data was divided into two parts:
points
of
so sunrise is near 1100 UTC and sunset near 2200
o
occurring
Winter
the
percentage
percentage of positive
the
shows
the
in
flashes
with
strike
on the land surface (dashed line with triangles), and
those located on the water surfaces of the Atlantic Ocean and the Gulf
of Mexico (dash-dot line with diamonds).
The Winter data showed
very
little
diurnal
variation
in
the
percentage of flashes with positive polarity between 0400 UTC and 0000
UTC.
For both land and ocean, there was an increase in the percentage
- 69 -
Climatological Phenomena
of
flashes
with
positive
polarity
approximately four hours after sunset.
that
thunderstorms
percentages
of
hypothesized
around
Since it has
positive
that
There
flashes
(Orville,
et
is
coincides
constant
been
al,
EST),
documented
1983),
also
an
with
a
percentage
it
is
within
of
positive
increase in the percentage of positive
Figure 2.12b shows that
this
decrease in the total number of flashes.
Around this time the number of positive flashes
remains
(21
after sunset the storms begin to decay and thus at
flashes over land at 1700 UTC (12 EST).
increase
UTC
in their late decaying stages tend to have higher
that time there should be an increase in the
flashes.
0200
25
percent,
flashes decreases by 40 percent.
occurring
while
This is not
each
hour
the number of negative
the
normal
summertime
pattern of a peak in negative flashes in mid-afternoon.
The total number of flashes (figure 2.12b) also
diurnal
variation
shows
is confined to one part of the day.
the
hour
of
2300
UTC
(18 EST).
in
the
winter,
flashes
The two graphs in figure 2.12
indicate that diurnal effects are not visible
lightning
the
There is a 30
percent increase from the mean value of 5000 per hour to 6500
in
that
in
the
production
of
except after sunset from 2300 UTC to 0400
UTC.
As noted before, this is in contrast to the summertime situation
and
may reflect that winter thunderstorms are more related to lifting
of potentially unstable
air
(as
over
de-stabilization by surface heating.
- 70 -
frontal
surfaces),
than
to
Climatological Phenomena
Support for this finding can be found
(1948)
and
Colman
(1984)
in
the
works
of
Shands
on storms occurring in the winter season.
Shands found "a tendency toward the equalization of
all four (six hour) periods of the day".
probabilities
in
Colman confirmed this result
and went on to state that the phase and magnitude of the diurnal cycle
of the convection are dependent on the frontal type.
observations include storms of all frontal types, the
and magnitudes would obscure a diurnal cycle.
-
71 -
Since the Winter
mix
of
phases
Climatological Phenomena
Positive Flashes versus Hour of Day
Percentage of
38
36
-4
34
32
30
I
a 28
-26
A
04
12
0
2
4
8
6
10
12
14
Hour (UTC)
16
18
20
Figure 2.12a Percent positive versus-hour of the day.
boxes
represents
all
GALE
flashes,
dashes
22
24
Solid line
line
with
represents flashes with strike points on land, dash-dot line
diamonds represents flashes with strike points over the ocean.
Number of Flashes versus Hour of Day
6500
6000
V
a
4600
,,
0
h~
S4000
A&..A
L 24.
p.
w3000
ZSO
.4... --1-,1.4.. "' -
;-- .-
.0
A
-t "
-.
" -
. , 4
a
-cI
C
-41
.. *
44
20001
00
I
'
1000~, I
2
II
4
I
6
I
I
I
I ,
I
10
12
14
Hour (UTC)
8
T
16
18
I
20
I
22
Figure 2.12b As in 2.12a but for number of flashes.
-
72 -
with
triangles
t
24
with
Climatological Phenomena
Summary
2.5
The
analyses
presented
picture
climatological
of
in
this
chapter
have
created
lightning characteristics.
positive
a
The
dependence on latitude of both the likelihood of producing a flash and
flash
polarity
results.
The
been documented, and is consistent with previous
has
effect
demonstrating
that
of
the
the
warm
Gulf
Stream
has
This
is
explained
an enhancement to the convection which is seen as deep clouds that
are more electrically active.
of
determined,
ocean surface is associated with flash
densities that are twice that observed over land.
as
been
positive
lightning
Geographic areas showing a predominance
are noted in the eastern Tennessee region and
the western Pennsylvania region.
densities
but
high
negative
An area showing low
flash
Carolina and Virginia, and believed
densities
to
be
located in North
is
a
flash
positive
result
of
convection
enhanced by topography.
An analysis of first stroke peak currents confirms
result
previous
that positive peak current averages are 1.5 times the negative
peak current averages.
times
the
Winter currents of
both
polarities
those observed in the Yearly and Summer data.
are associated with storm system-wide
positive
less
an
than
2.6
flashes/minute
positive peak currents decrease
flashrate increases.
over
significantly
increasing
averaging
and it is noted that
as
the
storm
system
The geographic distributions of peak currents of
both polarities average minimum peak currents in
Virginia,
1.5
Positive flashes
flashrates
hour,
are
in
magnitude
-
73 -
in
all
North
Carolina
directions.
and
Results
Climatological Phenomena
presented here suggest that the environment is the controlling
in
the
electrical
activity,
through
its
relationship
factor
with cloud
structure and development.
An analysis
lightning
of
shows
the
diurnal
relatively
variation
little
in
the
shortly
after
values
by
0000
for
sunset, from 2200 to 0400 UTC.
the total number of flashes increases 30 percent
normal
UTC.
UTC.
the
period
At 2300 UTC
decreasing
back
to
At 0100 UTC the percentage of positive
flashes increases by 10 percent, falling back to its normal
0400
of
variability in the percentage of
positive flashes and total number of flashes, except
occurring
production
value
by
This lack of diurnal cycle in winter thunderstorm activity
is consistent with previous work by Shands (1948) and Colman (1984).
-
74 -
CHAPTER 3
Convective Scale Phenomena
3.1
Introduction
Both the tilted dipole and the inverted dipole hypotheses account
for
the
of a positive cloud-to-ground flash by describing
existence
the electrical structure of a single thunderstorm cell.
will
chapter
This
look at phenomena associated with occurrences of cloud-to-ground
lightning on the scale of a single convective cell, or group of cells.
predicts
The tilted dipole hypothesis
emanates
that
positive
This predicts that
from the upper part of a tilted updraft.
positive lightning is associated with anvil clouds, or
"down-tilt"
side
of
the
main
The
updraft.
lightning
least
at
cloud
producing flashes of either polarity, but the production
the
is capable of
of
positive
flashes is restricted to the down-tilt area beneath the upper positive
charge region.
radar
Positive flashes will occur
reflectivity
than
that
of
the
negative
associated with the main part of the updraft.
the
same
height
beneath
areas
flashes
of
lower
which are
The top of the cloud is
for both polarities of flashes in this case, but as
the cloud begins to decay, it may shrink in height and the lowering of
- 75 -
Convective Scale Phenomena
the
positive
charge region associated with this decay may lead to an
increased likelihood of the
lightning.
A
production
relationship
between
would not necessarily be clear
flashes
may
in
of
this
case.
Since
the
positive
be associated with high anvil clouds, there may not be a
field
at
positive
flashes.
The
the ground for a group of tilted dipoles would be
dominated first by the lower negative charge
directly
cloud-to-ground
cloud height and flash polarity
direct link between the shallower clouds and
electric
positive
overhead.
The
field
would
then
polarity as the downtilt side of the updraft
region) moves overhead.
regions
when
switch
they
are
to fair weather
(upper
positive
charge
This predicts positive lightning occurring in
fair weather polarity electric fields at the surface.
produce
The
inverted
dipole
can
positive
polarity
only.
Negative
cloud-to-ground
flashes
of
flashes would come from separate
upright dipole cells existing in the same environment.
The
inverted
dipoles are expected to be associated with weaker updrafts and smaller
clouds.
in
This would result in lower reflectivities than would be
upright
dipoles,
and
possibly a stratiform-like cloud structure
characterized by small vertical gradients in
dipoles
are
seen
reflectivity.
Inverted
expected to be shallower than the upright dipoles in the
same environment, meaning that a relationship between cloud height and
flash
polarity
would
be
expected.
Beneath an inverted dipole the
field at the ground would be dominated by the
region.
This
predicts
that
positive
lower
charge
lightning will occur in fair
weather polarity electric fields at the surface.
- 76 -
positive
Convective Scale Phenomena
Four storms will be presented in
extent
to
which
two
summertime
There
The storms will be
representing the wintertime environment.
January
30
June 21 (1987), July 12
The relationships
and March 14 (1986).
(1986),
the
environment
thunderstorm
referred to by the dates on which they began:
(1988),
examine
to
the observations support the two hypotheses.
are two storms representing the
and
chapter
this
that will be shown are those between:
(1) the
points
any altitude, (2) location and
and
maximum
reflectivity
at
location
of
strike
vertical gradient of reflectivity, (3) location and cloud height,
(4)
the
field
electric
which
radar,
was
1987-1988 summers, and Fort
(1986)
Data
data.
from
corona point recorder
located
near
ground
the
at
in
located
Fischer,
with
associated
The data presented
cloud-to-ground events.
C-band
at
were
taken
Huntsville,
North
Carolina
by
Alabama
for
and
positive
the
MIT
for the
the
GALE
the SUNYA lightning detection network, the
Fort
Fischer,
and
corona
point
network
the MIT radar in Huntsville, will be used to relate the
electrical activity to the convective phenomena.
Section 3.2 will present two summer storms, and section
winter
storms.
3.3
two
Considerations of boundary layer temperature lead to
the expectation that the summer storms will have more energy available
for
convection.
This
would
lead
higher maximum reflectivity values.
to taller summertime clouds with
This convection would be expected
to by typified by distinct thunderstorm cells.
As was seen in chapter
2, peak currents in these storms are observed to be lower than in
winter
storms.
the
The winter boundary layer temperatures are less than
-
77 -
Convective Scale Phenomena
summer temperatures, and this would be seen by the
stratiform
structure
summer storms, the
maximum
than
winter
reflectivities
is
seen
storms
and
radar
in the summer.
should
higher
be
as
more
Compared to the
shallower
overall
a
with
lower
percentages of positive
flashes.
3.2
Summer Storms
This
section
Huntsville,
will
associated
with
dipole
lower
flashes and (2) positive
gradients
in
two
summer
storms
occurring
in
The relationships predicted by both the tilted
Alabama.
dipole and inverted
present
hypotheses
maximum
the
negative
relationship
flashes
than
negative
smaller
vertical
values
with
associated
than
relationship to be examined is
(1) positive
reflectivity
flashes
reflectivity
are:
flashes.
The
between
third
cloud
top
and flash polarity, and the fourth is the relationship between
height
positive strike points and the sign
of
the
electric
field
at
the
surface prior to the lightning.
The first summer storm took place on June 21, 1987.
observation from 2030 UTC to 2130 UTC.
time is shown in figure 3.1.
pressure
areas
It was under
The synoptic situation at this
Huntsville lies between two surface high
with light (10 knots) southerly winds at the surface.
Winds aloft are westerly at about 30 knots and the 0000 UTC
sounding
shows that the air is very moist at all altitudes except for
a dryer layer around 500mb.
shows
Nashville
The radar
summary
chart
that the convection is very cellular in nature.
- 78
for
2035
UTC
There are many
Convective Scale Phenomena
pockets of tall clouds with high reflectivity
particular
interest
values.
The
area
of
(outlined with a heavy line) includes the 46,000
foot radar top noted in northern Alabama.
Detailed radar observations for this storm cover a radius
kilometers.
evolution
Four
of
the
approximately
20
constant
storm
elevation
in
minutes
figure
angle
3.2.
of
56
scans (PPI's) show the
The
observations
are
apart and cloud-to-ground lightning strike
points for the period from three minutes before to three minutes after
the
observation time have been added.
An "x" is used to indicate the
location of a negative strike point and a "+"
location
of
a positive strike point.
45, and 60 dBz.
used
to
indicate
the
Contours are drawn for 15, 30,
The first PPI is taken while the storm
is
decaying;
where decaying is defined as decreasing in maximum reflectivity value,
or as a decrease in the area of the maximum reflectivity region.
cloud
The
continues to decay over the next 40 minutes (figure 3.2b and c)
as the storm loses its distinct cells and becomes a more uniform band.
Twenty
minutes later the observations show that although the northern
part of the band continues to decay the southern part begins to grow.
The overlaying of strike points and reflectivity values was
to
determine
the
reflectivity
associated with each cloud-to-ground
lightning event occurring within three
There
used
minutes
of
the
observation.
were 401 negative flashes and 43 positive flashes meeting these
criteria.
kilometers
For
was
each
flash
recorded.
the
The
maximum
reflectivity
within
two
mean
reflectivity
values
were
approximately the same for both polarities
- 79 -
--
41
dBz
for
negative
Convective Scale Phenomena
flashes and 40 dBz for positive flashes, with standard deviations of 9
and 11 respectively.
A comparison of mean reflectivity values for all
storms will be presented in section 2.4.
Constant azimuth scans (RHI's) depict the vertical
the thunderstorm.
of
of
Figure 3.3a shows an RHI taken at 2049 UTC which is
closest in time to figure 3.2b.
points
structure
The RHI is overlain with
the
strike
lightning flashes in the vicinity of the observation.
The
criterion used for acceptable strike points was that the difference in
azimuth between the radar RHI observation and the flash must result in
a difference of less
events
occurring
included.
many
than
within
7.5
kilometers.
three
cells.
The
seem
to
merge
distinct cells.
so
cell
is
composed
As time progresses
the
Figure 3.3b shows an RHI taken at 2115 UTC, about
after
the
of
that it becomes more difficult to identify
previous
observation.
The
azimuth
part
of
of
the
25
this
cloud
producing positive lightning, as shown in figure 3.2c, a PPI
taken 30 seconds after the RHI.
tilted
is
at 20 to 25 kilometers is actively
observation is two degrees, which is through the
that
lightning
minutes of the observation time were
producing negative flashes (shown as arrows).
minutes
only
The RHI shows that at this time the storm
distinct
cells
Again,
dipole
in
This picture seems
to
be
an
ideal
that the main reflectivity region (greater than 50
dBz) shows a tilt away from the radar with
down-tilt region.
-
80 -
positive
flashes
in
the
Convective Scale Phenomena
The RHI in figure 3.3c, which is closest in time to
figure
3.2c,
shows
positive
and
different cells in the same storm.
closer
to
the
main
negative
at
30
points
occur
flashes
seen
in
figure
degrees azimuth, 20 kilometers, that would be associated
with the tallest clouds seen at 5 and 15 kilometers in the RHI.
could
be
in
region than the positive flashes.
This RHI is prior to the flurry of negative
3.2c
PPI
flashes associated with
The negative strike
reflectivity
the
This
a case where the timing of the observation would make a big
difference in the
observed
relationship
between
the
strike
point
locations and cloud characteristics.
The
RHI-lightning
relationship
being
plots
examined
vertical reflectivity gradient.
was
calculated
were
used
to
evaluate
in this chapter:
the
second
flash polarity versus
The vertical gradient of reflectivity
as the difference in reflectivity between the maximum
and values divided by the minimum difference in height between the two
values.
Figure 3.4 and table 3.1 show an example.
representation of
the
reflectivity
values
for
Figure 3.4 shows a
each
kilometer
of
altitude for a negative flash (solid line with circles) and a positive
flash
(dashed line
with
squares).
To
calculate
the
vertical
reflectivity gradient:
(1) determine the maximum reflectivity value (table 3.1 line 1)
(2) determine the highest altitude at which this reflectivity is
observed (table 3.1 line 2)
(3) determine the minimum reflectivity value (table 3.1 line 3)
(4) determine the lowest altitude at which this reflectivity is
observed (table 3.1 line 4)
(5) subtract the minimum reflectivity from the maximum reflectivity
(table 3.1 line 5)
(6) subtract the height of the maximum reflectivity value
from the height of the minimum reflectivity value (table 3.1 line 6)
-
81 -
Convective Scale Phenomena
(7) divide (5) by (6) to get the vertical gradient of reflectivity
(table 3.1 line 7)
Line
1
2
3
4
5
6
7
Quantity
maximum reflectivity
max height of max reflectivity
minimum reflectivity
min height of min reflectivity
reflectivity difference
height difference
vertical gradient
Positive Flash
45 dBz
0 km
15 dBz
7 km
30 dBz
7 km
4.3 dBz/km
Negative Flash
35 dBz
4 km
15 dBz
8 km
20 dBz
4 km
5.0 dBz/km
Table 3.1: Values used in calculating the vertical gradient of
reflectivity associated with the strike point location of a negative
flash and of a positive flash.
June 21 contained 14 negative flashes
12
and
flashes
positive
The mean
that were considered "good" by the criteria described above.
vertical reflectivity gradients were:
and
4.4
dBz/km
for positive flashes with standard deviations of 2.7
associated
negative flashes.
with
positive
flashes
vertical
the
As predicted by both hypotheses,
and 0.8 respectively.
gradient
5.8 dBz/km for negative flashes
less than that for
was
For June 21 the vertical gradient
was
24
percent
less for positive flashes than for negative flashes.
The RHI-lightning plots can also be used to compare cloud
and
For
polarity.
flash
each
of
height
the 14 negative and 12 positive
flashes the height of the cloud was defined as the maximum altitude of
reflectivity
kilometer.
negative
of
15
dBz or greater, and was taken to the nearest 0.1
The means of the cloud heights
flashes
and
11.2
standard deviations of 1.3
kilometers
and
1.6
were
for
positive
kilometers
- 82 -
9.5
kilometers
for
flashes
with
respectively.
This
Convective Scale Phenomena
result
is
in
hypothesis,
contradiction
where
the
to
the
concept of the inverted dipole
flashes
positive
are
associated
with
the
shallower clouds.
The fourth relationship to be investigated in this chapter is the
relationship
between positive strike points and the electric field at
the ground.
Ten corona point recorders
in
figure
1.10)
used to measure the electric field at the ground.
Figure
were
(locations
3.5 shows the output from these ten recorders.
such
that
the
recorder
farthest
west
shown
The figure is arranged
(number
22) is at the top,
progressing downward to the recorder farthest to the east (number 30).
The
figure
shows
cloud-to-ground
that
an
were
flashes
"active
phase",
confirmed
by
when
the
network, existed from 2030 to 2150 UTC.
detection
SUNYA
field),
there
were
cloud-to-ground
evidence
of
an
flashes
A
lightning
few
cloud-to-ground
detection
(foul
of
both
polarity
"anomalous phase", when intra-cloud or positive
cloud-to-ground flashes are recorded, is seen from 2150
UTC.
lightning
overhead
The electric field then changes to fair weather
polarities.
and
and
negative
The electric field
at the surface is positive indicating negative charge
weather
many
network
are
flashes
during
the
confirmed
period
and
through
by
2240
the SUNYA
these
have
substantially greater currents than the events shown in figure 3.3 for
the active phase.
the
network
ground.
The positive cloud-to-ground
occurred
in
strong
flashes
detected
by
fair weather electric field at the
This evidence is consistent with either the tilted dipole
the inverted dipole hypothesis.
- 83 -
or
Convective Scale Phenomena
Noted in chapter 2 but worth repeating here is the
the
positive
lightning
figure 3.3).
for
in the active phase of this summer
currents
storm to be systematically smaller than the normal
(see
tendency
currents
negative
This result contrasts with the findings for winter
storms (recall figure 2.8).
The second summer storm occurred on July 12, 1988 and
observation
for
the
from
hour
0055 UTC to 0155 UTC.
was
The synoptic
At this time, Huntsville was
picture is shown in figure 3.6.
under
in
the
western part of a surface high pressure region with surface winds from
the
south
to
southwest
west-southwest
around
to
10
knots.
aloft
Winds
are
about
1000
kilometers
to
the
west.
The
moist air with the tropopause above 150mb.
very
shows
5
20 knots below 400mb and west-northwest above.
There is a 500mb trough
sounding
at
The
radar summary chart shows that the area of interest has reported cloud
tops
as
high
as 50,000 feet at the tallest point.
The southeastern
portion of United States is covered by many scattered clouds, although
three-quarters of Alabama is covered by the same cloud-mass.
Detailed observations of the storm were taken for a 40
radius
from
radar
the
site
in
Huntsville.
During the hour under
of
observation the storm was mainly composed of two bands
shown
in
3.7a.
figure
The
clouds
as
band to the northeast of the radar has
reflectivities as high as 55 dBz and the clouds reach
The
kilometer
14
kilometers.
cloud band to the west is much shallower, reaching no more than 9
kilometers
and
is
a
stratiform-like
cloud
with
reflectivity gradient either horizontally or vertically.
-
84 -
very
little
Convective Scale Phenomena
The movement of the bands is shown in figure
3.7.
The
eastern
band is decaying over the 30 minutes it was under observation (figures
3.7 a through c),
area
observation
activity
and the cloud-to-ground lightning
The
decreases.
shown in the RHI in
figure
3.8a.
the
in
vertical structure of the band is
The
shown
cells
are
producing
is
decaying,
exclusively negative lightning at this time.
The evolution of the western band
it
that
shows
in the northern part of the area of observation.
especially
some cloud-to-ground lightning associated with
the
eastern
There is
edge
at
0058 UTC plus or minus three minutes (figure 3.7a) and to the south of
the radar at 0114 UTC
(figures
3.7
b
and
0154
plus
UTC
and d, respectively).
or
minus
three
minutes
The vertical structure of the
band is shown in figure 3.8b, which shows the location of the negative
flash shown at 30 kilometers, 135 degrees azimuth in the PPI in figure
3.7d.
The cloud appears quite homogeneous below the bright band at
4
kilometers.
The relationship between reflectivity and flash
be
conclusively
positive flashes.
examined
polarity
cannot
for this storm because there were only two
The mean reflectivity values were 19 dBz for the 33
negative flashes and 30 dBz for the two positive flashes with standard
deviation of 20 and 8 respectively.
that
satisfied
the
There were
no
positive
criteria to be mapped to an RHI.
negative flashes that were considered "good" flashes and
to RHI's.
flashes
There were ten
were
mapped
For the ten flashes the mean vertical reflectivity gradient
was 4.0 dBz/km and the mean cloud height was 11.1
- 85 -
kilometers.
These
Convective Scale Phenomena
values
be
will
compared
to
the values from June 21 and the winter
storms in section 3.4.
The corona point records for this storm are similar
the
previous
storm
to
fair
weather
field at the ground.
positive.
area,
From
indicating
the
that
standpoint
the
of
dominant
followed
by, a
This change coincides
with the movement of the western band of this storm
network
of
There is an active phase with
(see figure 3.9).
confirmed cloud-to-ground flashes of both polarities,
change
that
to
over
charge
storm-relative
wind
the
corona
overhead
shear
is
and
electrostatics, this case strongly supports the tilted dipole concept.
>From the standpoint of the storm relative lightning locations, a more
confused
picture
is
presented, but accuracy of locations is problem
for comparisons on these scales.
- 86 -
millibar
500
Figure 3.1a:
chart for June 22, 1987, 0 UTC.
Solid lines are height contours,
dashed lines are temperature.
Figure 3.1b: Surface chart for June
22, 1987, 0 UTC. Solid lines are
isobars, fronts are analyzed.
Convective Scale Phenomena
N55
Figure 3.1c: NWS radar summary chart for June 21, 1987 20:35 UTC.
levels are contoured, maximum cell heights are underlined.
870622/0000
72327
VIP
BNA
.......
...................................
...................................
........................
....................
.......................
.............
...
..............................
...........
...............................................
.................
...
.........
ur,
SJ
L3J
.................
... .:..............,:...........
.........................
~5$/
oY
\Y,
............
........
........
.........
.....
........
....
1000
-40
Figure 3.1d:
1987, 0 UTC.
..-..
. .. . .. . ..
-20
..
uL,
uy
\y,
...
'
850 ............ . . ..............................
....
....
\yv
s~
uv
~
luL
:...
. ..
......................................................
.....
..--. .. .---.
0
.....
..
...................
....
.....--..............--.
....
.....
20
Sounding from Nashville Tennessee
(BNA)
Temperature and dewpoint are plotted.
- 88 -
.
40
for
June
22,
Convective Scale Phenomena
6/21/87
20:56 35
6/21/87
6/21/87
21:16:06
21:35:57
Figure 3.2 a through c: Constant elevation scans from the MIT radar in
Huntsville. Dates and times (in UTC) are labelled; a is earliest time,
c is latest. Range is 56km, elevation angle 0.7 degrees. Contours are
(lightly dotted), 30 (densely dotted), and 45 (striped)
drawn at 15
dBz.
- 89 -
Convective Scale Phenomena
6/21/87
20:49:26
Azimuth
300.0 degrees
,:: =
0
5
10
25
30
35
40
45
50
55
-60
6/21/87
21: 15' 34
+15+8
6/21/87
21:28:38
Azimuth
2.0 degrees
+8
Azimuth
35 - 40
27.0 degrees
45
50
55
Figure 3.3 a through c: As in figure 3.2 but for constant azimuth
Contours are drawn for 15, 30, and 45 dBz. Arrows indicate
scans.
range and first stroke peak current in kamps of lightning flashes
within 7.5km and 3 minutes of the observation.
- 90 -
Vertical
0
5
10
15
Distribution of
20
25
Reflectivity
30
(dBz)
Reflectivity
35
40
45
Figure 3.4 Sample of reflectivity versus height for a negative
flash strike location (solid line with boxes) and a positive flash
strike location (dashed line with triangles).
50
Convective Scale Phenomena
06/21/8'7
14
23
+3.5
z
16
Z
24
FOUL
28
25
-35 FAIR
18
20:10:00
20:30
20:50
21:10
21:30
TIME
(UTC)
2150
22:10
22:30
22:50
Figure 3.5 Corona point records for June 21, 1987. Stations run
(top) to west (bottom). See figure 1.10 for exact locations.
- 92 -
east
II0
C)
is.~~~?
.
"56,2.
65.so
t
5 po
are temperature.s"
..
.
c
a
•?-20
59
R1
,7
,
- 6-
-
"
Z
;.
CT
"
3
-
121
(D
/
.
J*q-
159
(D
sl
jj 59-- Cs
10
*r
.*t*l(
62
fI
-; .Iss
7
OQ~
i
.
6 o68
'IIG
-/
i4
14b
222,
59
U
0(
are
isobars,
to
57i; analyzed.
65fronts
90,~c
65
-14
.
G li
~
IJ4
\v
G4-*jej
-1
6
11
f.,,''-
11Q0
rs
52V
r~l;~qh
T.)75 143L
,
j . "
02 '-';.
1(752
"
-
......
.
4
48"1
+J.
+~
c:
3Li
.
72 .8i f 11I
£
0!..
4 4
I
\L
76t 2 3
ts
-051~-
;/vr
L7
-e., ,5
7P" '~ t
I !I
72
is
i7
7
IG a
I'
513+1411
t%
,
are
contours,
173 0-
?7-107'f
......
Figure
3.6b:
Surface
chart .72for July
.
74t 3
78
0 UTC. Solid lines are
12,
1988,
a
74;o~I
77;0
temperature.
dashed
lines
Figue 36b:
12f
Figure
11
5
82: i 24
are height
-
%?
7-
a,,
5
_6.1
9
21
for
chart
500 0millibar
Figure
lines
Solid
TC
July
123.6a:
4 1988,
""
" - " -"
-': "i ,.
' =
L5 .
6
6
183 60,3
3.6a
5isobars,
urfce hartforJul
UC. Slidline ar
988,0
fronts are analyzed.16
'i
~
(D
Convective Scale Phenomena
Figure 3.6c: NWS radar summary chart for July 12, 1988 1:35 UTC.
levels are contoured, maximum cell heights are underlined.
72327
880712/0000
.. .100
-40
....
BNA
.......-
-20
VIP
°............
0
40
20
Figure 3.6d:
Sounding from Nashville
1988, 0 UTC.
Temperature and dewpoint are plotted.
- 94 -
Tennessee
(BNA)
for
July
12,
Convective Scale Phenomena
7/12/88
7/12/88
0: 58:13
7/12/88
1:27:03
7/12/88
1:1434
r54:20
Figure 3.7 a through d: Constant elevation scans from the MIT radar in
Huntsville. Dates and times (in UTC) are labelled; a is earliest time,
d is latest. Range is 40km, elevation angle 0.7 degrees. Contours are
(lightly dotted), 30 (densely dotted), and 45 (striped)
drawn at 15
dBz.
- 95 -
Convective Scale Phenomena
7/12/88
0:59:49
Azimuth
54.3 degrees
15-
10-
:
= =:
:?c.
~f~i~f~i~
=
~~
5
10
15
20
Z5T
30
35
40
-15
7/12/88
15
1:53:18
Azimuth
134.0 degrees
-
10
5
S
I0°
5
10
15
20
25
30
35
40
-36
Figure 3.8 a and b: As in figure 3.7 but for constant azimuth
Contours are drawn for 15, 30, and 45 dBz. Arrows indicate
scans.
range and first stroke peak current in kamps of lightning flashes
within 7.5km and 3 minutes of the observation. Dashed parallel lines
indicate bright band.
- 96 -
Convective Scale Phenomena
TIME
(UTC)
Figure 3.9 Corona point records for July 12, 1988. Stations run
(top) to west (bottom). See figure 1.10 for exact locations.
-
97 -
east
Convective Scale Phenomena
3.3
Winter Storms
occurred
that
storms
winter
This section will present two
on
January 30, 1986 and March 14, 1986, respectively, with detailed radar
Fischer,
and corona point observations based* at Ft.
The
th,e pattern
follow
will
presentation
lower
with
associated
flashes
positive
reflectivity than negative flashes,
top height
flash
and
vertical
in
gradients
(3) the relationship between cloud
(4) the
and
polarity,
cloud-to-ground
positive
smaller
with
associated
flashes
than negative flashes, (2)
values
reflectivity
maximum
Carolina.
of the previous section
(1) positive
looking for the same relationships:
North
between
association
and electric field at the surface.
flashes
>From several considerations, it is expected that
the
storms
winter
be much shallower, with lowex vertical gradients in reflectivity
will
than summer storms.
The first storm occurred on January 30, with
in
area
observation
the
from
back
are
winds
The synoptic
The storm under study occurred
situation is shown in figure 3.10.
the
UTC to 0315 UTC.
0300
side of a cold front which moved in from the west.
northerly
west-southwesterly
at
at
40
15
knots,
winds
The
knots.
aloft
are
The
figure 3.10d.
the
rest
of
closest
other soundings.
or
storm studied here is at the
the
500mb
sounding was taken at Hatteras and is shown in
The layer from 600 to 650mb looks quite different
the
on
Surface
westerly
coast of North Carolina at 0300 UTC, and so is just behind
trough.
activity
lightning
from
sounding, but this feature is not supported by any
Since not all radar sites reported at 0335 UTC,
-
98 -
the
Convective Scale Phenomena
The storm
summary chart shown in figure 3.10c is for 0235 UTC.
radar
-
Carolina
North
of interest is the band along the
Carolina
South
coast.
approximately
are
PPI's
17
minutes
last
and
first
observation
apart, with no cloud-to-ground
The
times.
throughout this period, where decay is
cloud
as
defined
around
windows
minute
six
lightning activity occurring in the
The
3.11.
figure
in
The evolution of the cloud band is shown
the
band
is
decaying
in
the
previous
section.
Two RHI's are presented in figure 3.12 to illustrate the vertical
structure of the band.
The first RHI was taken at 57 degrees azimuth,
midway in time between the first two PPI's.
of
positive
a
flash
with
associated
Assuming the accuracy of the strike point,
inverted
dipole.
example
The RHI shows an
the anvil part of the cloud.
this
is
clearly
not
an
The second RHI is also taken midway in time between
the first two PPI's and was taken at 269 degrees
azimuth.
This
RHI
shows a negative flash associated with the main reflectivity region of
the largest cell.
There were eight negative and 15 positive flashes associated with
the
PPI's.
The mean reflectivity values were 42 dBz and 43 dBz with
standard deviations of 11 and 7 respectively.
As was the case in
the
summer storms there is no reflectivity difference between positive and
negative flashes.
This is contrary to expectations
based
on
the tilted dipole hypothesis or the inverted dipole hypothesis.
-
99 -
either
Convective Scale Phenomena
to
One negative and four positive flashes were able to be mapped
For the negative flash the vertical reflectivity gradient was
RHI's.
7.0 dBz/km and the cloud height was
positive
flashes,
6.0
the vertical reflectivity gradients were 4.0, 4.0,
5.0, and 5.0 dBz/km, yielding a mean of 4.5 dBz/km.
The
cloud
mean
was 5.5 kilometers with a standard deviation of less than half
height
With so few observations there is little
a km.
four
the
For
kilometers.
results,
but
these
in
confidence
the results do agree with the predictions in this case.
The positive flashes are associated with lower
vertical
reflectivity
gradients and shallower clouds than the negative flashes.
3.12a
The vertical structure shown in figure
tilted
dipole,
and
be
to
appears
a
it would therefore be expected that the dominant
an
electric
charge over the radar site would be negative,
producing
field at the ground of foul weather polarity.
The corona point record
for Ft.
Fischer at this time shown in figure 3.13 is consistent
with
this expectation.
The second storm to be presented in this section is part
storm occurring on March 14.
characteristics
the
As was noted in chapter 2, although this
storm produces much more lightning than
electrical
of
are
typical
any
of
other
all
GALE
GALE
storm,
its
storms.
The
detailed radar observations used here will cover only the period
1445 UTC to 1800 UTC.
- 100 -
from
Convective Scale Phenomena
The synoptic situation shown in figure 3.14
cloud
band
under
study
3.14b.
By
(figure 3.14d).
and c).
0000
UTC
surface
the
analysis
in
UTC on March 15, the two fronts have merged
There is a broad trough at 500mb (see
figures
3.14a
The sounding in figure 3.14f shows that although winds at the
surface are light, there is a low level jet from the
55
that
is associated with the area between the two
occluded fronts shown in the March 14 0000
figure
indicates
knots at 950mb.
south,
reaching
Above this the flow is south-southwest, gradually
increasing from 30 knots at 725mb to 95 knots just above
200mb.
The
radar summary chart from 1635 UTC shows that the area under study here
is an intense section of a continuous cloud
region
that
covers
the
northeastern United States.
The evolution of the storm in shown by four PPI's in figure 3.15.
The area of observation covers a radius of 100 kilometers.
Throughout
the period the cloud band broadens and develops two sections
UTC
by
1730
(figure 3.15c), with the front section containing an area of very
high reflectivity (50 dBz) and the back section having less horizontal
gradient
in
reflectivity.
This
is
similar
to the July 12 storm,
though on March 14 there are not two distinct cloud bands.
One phenomenon shown by the four PPI's is that
can
occur
in
any
part of the cloud-mass.
positive
In the first PPI (figure
3.15a) there is a positive flash at 10 degrees azimuth, in the
PPI
second
a positive flash at 190 degrees azimuth, and in the fourth PPI at
30, 90, 155, and 270 degrees azimuth.
there
flashes
was
On the large scale (not shown),
a preference for positive flashes downwind (to the north).
-
101 -
Convective Scale Phenomena
A detailed discussion of
the
relationship
between
and
wind
flash
polarity will be given in the next chapter.
In the time period under study there were
flashes
positive
negative
and
47
met the criteria for being overlain on PPI's.
that
An analysis of the reflectivity
values
point
reflectivity
locations
314
yielded
mean
associated
with
values
of
the
strike
33
dBz for
negative flashes and 27 dBz for positive flashes, which is 20
percent
For this storm there was a tendency for positive flashes to be
less.
associated with lower reflectivity than negative flashes.
The vertical structure is shown in
shows
the
3.15a).
cell-like
structures
The cloud band is
negative
producing
figure
3.16.
Figure
that comprise the first PPI
growing
cloud-to-ground
at
this
flashes.
time
and
Figure
is
3.16a
(figure
actively
3.16b shows a
tilted reflectivity region of reflectivity greater than 40 dBz, and
positive
flash in the vicinity.
a
This observation supports the tilted
dipole hypotheses, but since the flash is not unambiguously associated
with the high reflectivity region it does not rule out the possibility
of an inverted dipole.
Figure 3.16c was included to show the similarity between the back
section
of
the
March 14 cloud band and the stratiform-like cloud of
July 12 (see figure 3.8b).
horizontal
and
vertical
The reflectivity is fairly uniform in both
directions below the bright band.
can be contrasted with one taken through the front section 38
later (figure 3.16d).
This RHI
seconds
The back band is shown from 0 to 15 kilometers,
- 102 -
Convective Scale Phenomena
a slight break in reflectivity of 30
kilometers,
and
or
greater
exists
at
15
then a more cell-like structure is seen farther away
from the radar.
associated
dBz
As is typical of this storm, the
negative
flash
is
with the highest reflectivity region (at 35-45 kilometers)
and the positive flash is in cloud of lower reflectivity to the north.
Figure 3.16e again shows the two bands,
Both
stratiform
flashes.
is
and
convective
sections
but
20
minutes
are
producing
later.
positive
There is nothing to indicate that the flash at 40 kilometers
associated
with
a
tilted
The
dipole.
cloud
reflectivities and stratiform-like structure expected of
has
an
the
low
inverted
dipole.
There were 17 negative and 8 positive flashes that were mapped to
RHI's.
4.0 dBz/km
standard
mean vertical reflectivity gradients were 4.8 dBz/km and
The
for
negative
deviations
and
positive
flashes
were 2.0 and 1.9 dBz/km.
respectively.
As in all the previous
cases the vertical gradient in reflectivity was greater
flashes
than
for
positive flashes.
The
for
negative
The mean cloud heights were 6.9
kilometers and 6.0 kilometers for negative and positive flashes,
with
a standard deviation of 1.7 for both polarities.
The corona point record for this storm is very detailed
to
the
compared
other storms, and also much longer, spanning three hours.
the storms under study it is the best
source
evidence
The
of
the
inverted
dipole.
of
data
look
for
inverted dipole hypothesis
predicts positive cloud-to-ground flashes occurring
- 103 -
to
Of
in
fair
weather
Convective Scale Phenomena
polarity
field
at
the
ground.
A sample of an "anomalous phase" of
this storm is shown in figure 3.17.
confirmed
by
cloud-to-ground
Seven
field
3.17.
The
flashes
could
be
strikes detected by the SUNYA network.
These seven field changes are noted with a number
figure
changes
associated
with
in
the
parenthesis
in
field changes are
described in table 3.2.
Flash
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Time (UTC)
17:21:27
17:26:50
17:34:13
17:37:14
17:40:23
17:44:14
17:46:12
Polarity of Flash
positive
positive
positive
positive
positive
positive
negative
Range (km)
41
26
14
5
44
45
12
Polarity of
Electric Field
foul
foul
foul
foul
foul
foul
foul
Table 3.2: Cloud-to-ground flashes detected by the SUNYA network
associated with field changes in the corona point record form Ft.
Fischer.
A similar analysis was done for the entire record.
cloud-to-ground
flashes
confirmed
by
flashes
occurred
inconsistent with both
in
foul
weather
hypotheses
for
lightning and remains a puzzle.
-
104 -
positive
the SUNYA lightning detection
network occurred with a fair weather polarity
All
No
field
fields.
the
at
This
production
the
ground.
evidence is
of
positive
millibar
500
Figure 3.10a:
0 UTC.
1986,
30,
January
chart for
contours,
height
are
Solid lines
dashed lines are temperature.
Surface chart for
Figure 3.10b:
Solid
1986, 0 UTC.
January 30,
are
fronts
isobars,
lines are
analyzed.
Convective Scale Phenomena
Figure 3.10c: NWS radar summary chart for January 30, 1986 2:35
VIP levels are contoured, maximum cell heights are underlined.
72304
860130/0000
UTC.
HAT
100 -
. ...........................................................
250
300 *.
400
*
*
............
...............
..................
................
..............
.................................
.......................
700 .*.....
-40
. .............
.................
-20
0
20
40
Figure 3.10d: Sounding from Cape Hatteras, North Carolina (HAT)
for January 30, 1986, 0 UTC. Temperature and dewpoint are plotted.
- 106 -
Convective Scale Phenomena
1/30/86 3:00:47
1/30/86
1/30/86
3:17:43
3:35:59
Figure 3.11 a through c: Constant elevation scans from the MIT radar
in Ft. Fischer. Dates and times (in UTC) are labelled; a is earliest
Range is 100km, elevation angle 0.0 degrees.
time, d is latest.
Contours are drawn at 15 (lightly dotted), 30 (densely dotted), and 45
(striped) dBz.
-
-
107
-107
-
Convective Scale Phenomena
1/30/86
15
30
3'8:13
45
57.0 degrees
Azimuth
60
75
+230
1/30/86
0
3"10:2 Azimuth
45
15
60
269.0 degrees
75
-51
Figure 3.12 a and b:
As in figure 3.11 but for constant azimuth scans.
- 108 -
0
POINT DISCHARGE
FT
FISHER,
CURRENT
o-
N.C.
.-
JAN 30 1986
IO1A
FOUL
FAIR
400
300
UTC
Figure 3.13 Corona point record for January 30, 1986 from Ft.
North Carolina.
Fischer,
0
0(D
3
1.,6
5 0' .-
'..
,6
_
"t?
a-s
Z}
...
P'o?
G7
29
" .'
-:.
°- .1.
14
16
'1
k,
6-1
,9
•
I
85A
o
J0
C.
7I
0
o.-3
.46
429
.-
/
, vQ s
* ."-l-31,
/-
7P"-, .
S2
57-
.
N
., . Qr1~
4S.-
2j1
10ur
LOW
31
,-
Figure7
5052lia
30
3.14a
-3
fr
Marh
1F
,
ar
51 b
S
10
s-Q(
2-)
L7
2
s
01fo
T
March
1
mllibar50
186,0 UC.
3
43
14a:
a
c
1
t -)'
./
are47 sobars 18f
Mrch
U
50
0
-
;a
0
ts6~Z
3sati
'V
s,,
1
S? 0~,~?
1
4
Figure
500
br3491
4
Sr.
J
18
chart
14,
3
j-
v-I
a1
IH31o
chat
5-43
1
.
\
0
S
~S
75 1
74 1
7
7
2
Figur
3.1b:
Srfac
Marh
4,196,0
TC
are isobrs, frots are
chat
fo
Sli
lne
nalyzed
75 1606
LO
0
0
.5(1
\16
)9epJ,5g
: 869
91
S
),
3-,--/
.i
/-4
.
U r t
/5982!13
O As
8
H
.,
C
*.
43
/7
6a
\
,
Hi6
263a
0
30
P_1.
(D
2.7"'
'ON, *.A
q7))~
'
',.
'-'
l
?7
.l
ES
36
are isobars, fronts are analyzed.
64
, -Q..
,-.,'
.
-3044
/CT
ns;S ./
/
rt
i
314
q6
8'(s63s
34
1
-
153~
,43 , ,,,"
*-i.. ,.5wi
I~S
log
41.,,
:"-35
T.5x+"1;
).
5 "!q" 4
321
8
-4
4.52,~
o,2
91l
'
7O, G
5344?
P\j
17
5
Solid
ae
21,
8
22
1-
2Q;6
34Z
46
~33
__
0141
52
lnontort
.34
7
17
1VJ
))
4 57*4L3
I
'3,dI
-14
4411
4-~69
ei"1i"
I'S
44 I1
33r
20
1
T
~
5
7
..
4A
HIGH
h0i
i"
2
bts
fo.r1
1
15
char
Mac
24
2913
o
5;0 'j_
45
01-
IQ
3
' 4
fj
-
52
i~ te
3
Figure3.14:
3 4.
50 0 UTC.
March
15,
1986,
,14 lines
Solid
4?
1&3,_
53
.,1
M0
March
nr
2,.s
ioam
5 t
15~~~5,
i
53.s~
56
796
fo
T,
oi
r
aayzd
ie
\Q\1~
Convective Scale Phenomena
Figure 3.14e:
NWS radar summary chart for March
14,
1986
6:35
UTC.
VIP levels are contoured, maximum cell heights are underlined.
860314/1200
99903
ILM
150
****.
. .
200
150
....
*
400
...
.-
-
-
.....
l
i
.
..
.... ..
....................
*..."
/
112. -.
112
...
....
"
...
.."..."
..
....
..." ,
50
..
"
......:.....................
.....................
......
2
...
"
....................
.. ...
............................
..... .......
."""
.
700.....
... ..............
........
..........
.
.
........................
..
100...
..
....... ..
....................................
:
.................................
1 50..............
..
."
..."
..
Figure~
~
~~..
2
...
"
...
"
rm~timntn
•1 i
.. ":Sudn
.................. :.-,::n
..................
T e.-:-::r
...............
100
-.0
-20
0
arln
ot
p l
20
IM
o
ac
td
40
Figure 3.14f: Sounding from Wilmington, North Carolina (ILM) for March
14, 1986, 0 UTC. Temperature and dewpoint are plotted.
Convective Scale Phenomena
3./14/86
3/14/86
3/14/86
14:46: 35
17:29:14
16;17:52
3/14/86 18:01:10
Figure 3.15 a through d: Constant elevation scans from the MIT radar
in Ft. Fischer. Dates and times (in UTC) are labelled; a is earliest
Range is 100km, elevation angle 0.0 degrees.
time, d is latest.
Contours are drawn at 15 (lightly dotted), 30 (densely dotted), and 45
(striped) dBz.
-
113 -
-113
-
Convective Scale Phenomena
3/14/86
14:50-6
Azimuth 205.0 degrees
15
105
O
O
-86
3/14/86
15:7:16
Azimuth 25.0 degrees
IS-
5-
60
O
75
+58
Figure 3.16 a and b:
As in figure 3.15 but for constant azimuth scans.
-
114 -
Convective Scale Phenomena
3/14/86
0
17:25 4
30
15
3/14/86
0
15
Azimuth
45
17:25"42
145
30
300.0 degrees
60
Azimuth
60
0
15
30
17:45:41
7 45
45.0 degrees
-75
1
Azimuth
60
45.0 degrees
75
+72
+86
Figure 3.16 c through e:
scans.
As in figure 3.15 but
- 115 -
90
+65
-117
3/14/86
90
75
for
constant
azimuth
-1720
-1730
- 1740
+ 45 FOUZ
/1)5)
- 45 FAIR
Figure 3.17 Corona point record for March 14, 1986 from
North Carolina.
Ft.
Fischer,
Convective Scale Phenomena
3.4
Comparison And Summary
This chapter has considered phenomena related to
of
positive
lightning
relationships chosen to
inverted
dipole
that
confirm
hypotheses
on
occur
or
the
refute
of
production
convective
the
tilted
scale.
The
dipole
and
(1) the location of strike points
were:
and maximum reflectivity at any altitude, (2)
gradient
the
location
and
vertical
reflectivity, (3) location and cloud height, and (4) the
electric field at the ground associated with positive
cloud-to-ground
Predicted results for the tilted dipole hypothesis are:
events.
positive strike
positive
points
strike
associated
points
reflectivity, (3) slight
positive
and
lower
reflectivities,
(2)
with lower vertical gradients of
associated
difference
strike
negative
with
(1)
cloud
in
top
height
between
point locations, with positive strike
points being associated with lower cloud tops, and (4) positive strike
points
associated
with
fair weather polarity electric fields at the
surface.
Predicted results for the inverted dipole are the
all
the
but
third
relationship.
The
inverted
predicts that positive strike points will
be
same
for
dipole hypothesis
associated
with
lower
cloud tops than negative strike point locations.
The first relationship is between strike points and
values.
In
both
storms presented in section 3.2 and also in all of
the Huntsville storms studied, there is
no
reflectivity
strike
polarities.
3.3.
There
values
associated
with
The actual mean reflectivities
is
reflectivity
difference
are
points
in
for
presented
the
the
in
mean
two
table
considerable difference in the actual measured means
-
117 -
Convective Scale Phenomena
for the two storms (approximately 10 dBz),
but the difference
between
the mean reflectivities associated with the two polarities is only one
dBz.
Mean Positive
Reflectivity
41
30
37
Number of
Negatives
401
33
828
Mean Negative
Reflectivity
41
29
38
6/21
7/12
Huntsville
Number of
Positives
43
2
66
1/30
3/14
GALE
15
47
62
8
314
322
43
27
31
42
33
33
Total
128
1150
34
37
Time Period
associated
Summary of arithmetic mean reflectivities
Table 3.3:
with strike point locations for Huntsville and GALE data.
For the winter storms the relationship between mean
polarity is not clear.
flash
and
values
polarities,
but
the
table 3.3).
The
mean
table
in
presented
representative.
flashes
is
for
values
The
3.3.
means,
by
values
March 14 has been
two
these
storms
combined
a
for
shown
is
however, are not necessarily
negative
March 14 which contained 98 percent of the
combination
flashes accounting for 24 percent.
characteristics
both
The mean reflectivity
is
the
March 14 storm shows a difference of 6 dBz (see
dominated
flashes
reflectivity
for
values
The mean reflectivity value associated with
negative flashes.
positive
The January 30 storm shows
reflectivity
essentially no difference in mean
reflectivity
of
value
the
associated
be
This invalidates a
typical
of
the
two storms, January 30
comparison
the combination of the two storms.
to
with
winter
storms
of
Because
in
the
examined in chapter 2, it is believed that it is also
- 118 -
Convective Scale Phenomena
typical of the winter storm in these results.
The result of this analysis is
different
exhibit
that
characteristics.
summer
The
and
summer
does
support
the
relationship
between
storms
Huntsville
storm
The March 14 winter
results are inconsistent with the two hypotheses.
storm
winter
flash
polarity
and
reflectivity predicted by the two hypotheses.
be
The second relationship to
between
flash
considered
is
the
relationship
and vertical gradient in reflectivity.
polarity
Both
hypotheses predict that positive flashes will be associated with lower
vertical
gradients.
The
observations do support this relationship.
Only three of the storms had positive flashes meeting the criteria for
this
analysis.
All
three
reflectivity associated
with
storms had smaller vertical gradients in
positive
flashes
than
with
negative
flashes (see table 3.4).
Time Period
6/21
7/12
1/30
3/14
Mean Cloud Top
Height (km)
Positive
Negative
9.5
11.2
11.1
--6.0
5.5
6.9
6.0
Mean Vertical Reflectivity
Gradient (dBz/km)
Negative
Positive
5.8
4.4
4.0
--7.0
4.5
4.8
4.0
Table 3.4: Summary of cloud top height and vertical reflectivity
gradient means for the two summer and two winter storms presented in
this chapter.
-
119 -
Convective Scale Phenomena
The third relationship is
height
and
flash polarity.
the
relationship
between
support
the
predicted
mean cloud top
height
locations
greater
was
top
Again, only one of the summer storms had
positive flashes meeting the criteria for analysis.
not
cloud
relationship.
associated
with
This
storm
does
For the June 21 storm the
the
positive
strike
point
than that associated with the negative strike
point locations by 15 percent.
The winter situation
is
exactly
the
For both winter storms the positive flashes were associated
opposite.
with lower mean
cloud
of
interpretation
top
heights
(by 8
and
14
The
percent).
this analysis is that since only the tilted dipole
hypothesis can explain positive flashes associated with
higher
cloud
heights, inverted dipoles were not responsible for the production
top
of positive flashes on June 21.
The
final
observations
relationship
and
predictions
shows
inconsistency
between
based on the two hypotheses.
Both the
tilted dipole hypothesis and the inverted
fair
polarity
weather
flashes.
electric
fields
an
dipole
at
hypothesis
predict
the ground for positive
No flashes confirmed as positive cloud-to-ground strikes
by
both corona point data and SUNYA lightning detection data originate in
a fair weather electric field at the ground.
There is electrostatic documentation on July 12 that the
cloud
band
is
a tilted dipole.
western
The inverted dipole hypothesis does
not allow for the production of negative flashes and
yet
band
This is based on
is associated with flashes of both polarities.
flashes observed by the SUNYA lightning
- 120 -
detection
this
network,
but
cloud
not
Convective Scale Phenomena
confirmed
with
matching
field
changes
in the corona records.
lightning results present a more confused picture and do
not
The
provide
clear support for the tilted dipole picture.
The results
hypothesis
over
of
the
this
chapter
other.
against each of the hypotheses.
do
not
convincingly
favor
one
Indeed, there is evidence both for and
A weakness in all the
analyses
done
here, however, is that they assume (1) that the strike point locations
are accurate to within a few kilometers and (2)
propagation
between
not significant.
the
that
the
horizontal
origin of the flash and the strike point is
The extent to which flashes
propagate
horizontally
and that effect on the above analyses has yet to be determined.
- 121 -
CHAPTER 4
Mesoscale Phenomena
4.1
Introduction
In this chapter, we consider mesoscale aspects of
its
organization,
and
focusing on the relationship between lightning and
the synoptic and mesoscale
also
lightning
environmental
characteristics.
We
will
examine the mesoscale organization of lightning strike locations
which Orville, et al, (1988) have termed the "bipole".
The first section will
hypothesis.
separating
deal
directly
with
the
tilted
dipole
The environmental characteristics related to horizontally
the
considered.
charge
We
will
regions
through
wind
shear
will
then
be
also examine the relationship between positive
flashes and the height of the -100C isotherm, which is
a
measure
of
the height of the positive charge region.
The heart of the tilted dipole hypothesis is
charge
region
charge region.
conditions
were
is
displaced
horizontally
that
relative
Such a displacement could result if the
such
that
the
positive
to the negative
environmental
the updraft of the cloud was tilted with
- 122 -
Mesoscale Phenomena
given
a
in
speed
direction
containing the charge regions.
larger
the
The more the wind speed increases
As the
negative charge regions.
height over the layer
with
increases
centers
the
between
displacement
the
wind
One way for this to happen is that the
respect to the vertical.
tilted
the
increases,
shear
wind
the positive and
of
dipole hypothesis predicts that the probability of the occurrence of a
relationship
This
increase.
also
will
has
been
positive
flash
examined
by several investigators, with results summarized in a graph
by Takeuti (1984) which will be presented in section 4.2.
It has been suggested that positive lightning is
in
temperature,
charge regions are tied to
conditions
wintertime
charge region comes
and
closer
height of an isotherm, and
to
the
ground,
prevalent
al, 1987) because the
et
therefore
lead to lower charge regions.
positive flash increases.
because
(Orville,
conditions
climatic
colder
more
colder
that
As the positive
probability
the
of
a
Climatic temperature can be assessed by the
Takeuti
has
chosen
the
of its link with the negative charge region.
isotherm
-10 C
His results and
GALE observations will be presented in section 4.3.
One way to displace the positive charge region is with
to displacements in
symmetrically
can be
gradients.
and
a
tilted
slanted
or
unstable (see appendix A).
assessed
momentum
tilted
tilted updraft is possible if the atmosphere is unstable
A
updraft.
a
by
potential
i.e.
it
is
Moist symmetric instability
cross-sections
constructing
equivalent
direction,
temperature
of
pseudo-angular
and comparing their
This analysis will be shown for four cases in section 4.2.
- 123 -
Mesoscale Phenomena
The results of these analyses show that, in most cases, the atmosphere
In only one case is there evidence
is unstable to upright convection.
that
instability
symmetric
may play a central role, suggesting that
occurrence
symmetric instability is not a necessary condition for the
of positive lightning.
lightning
bipolar
negative
shallower
a
in
flashes
predominantly
occur
flashes
positive
The
part
of the
There is not a clear division between regions of positive
cloud-mass.
and
the
from
the
that
indicate
downwind
of
Three are located over the ocean and three over land.
patterns.
results
cases
six
Section 4.4 will then examine
strike
negative
points,
upwind
the
but
end
has
the lowest
percentage of positive flashes in the storm, and the downwind end
Results
The life-cycle of the bipole is then presented.
the highest.
has
of this analysis indicate that a decrease in bipole size is coincident
with an increase in the negative flashrate of the storm as a whole.
4.2
Environment Of Anomalous Lightning Events
dipole
tilted
The
responsible
of
percentage
charge
region.
positive
flashes
It
compiled
by
wind
shear
is
of
positive
Takeuti
has
been
Figure 4.1 shows a graph of
events versus wind shear.
(1985)
that
observed
the
increases with increased wind shear
(Brook, et al, 1982; Takeuti, 1985).
probability
that
for a displacement of the positive charge region relative
negative
to the
predicts
hypothesis
shown
are
by
a
the
Previous results
variety
of
symbols
GALE results are shown as circles.
indicating the source of the data.
-
124 -
Mesoscale Phenomena
The filled circle
is
based
on
sounding
data
from
the
NOAA
P-3
aircraft, while other GALE results are based on routine sounding data.
The wind shear is calculated by subtracting the wind speed
lower
level
This
environment
balance.
shear
wind
and
the
of
is
system
that
storm
4.1) suggested the idea of
production
estimate
results
Previous
a
to represent the negative charge region from the
chosen
wind speed at an upper level chosen to represent the
region.
at
positive
(triangles
a
cutoff
positive
of both the
representative
perturbs
charge
the
geostrophic
and filled squares in figure
in
wind
shear
for
necessary
When all results including GALE
lightning.
observations are considered a cutoff at 1.0 m/s/km is indicated, which
is
in
reasonable
agreement
with
the
1.5 m/s/km cutoff originally
proposed by Brook (Brook, et al, 1982).
One idea central to the tilted
dipole
hypothesis
is
that
the
closer the area of positive charge is to the ground the more likely it
Since the heights
is to produce a positive event.
regions
of
charge
been
have
shown
center
and
thus
should
a
main
lower
positive
be more likely to produce anomalous
of the -10 C isotherm and the probability of the occurrence of
a positive flash (using the same symbols as in figure
GALE
two
Figure 4.2 shows the relationship between the
events (Takeuti, 1985).
height
the
to be tied to temperature (see
section 1.2), a lower -10*C isotherm would suggest
charge
of
data
there
are
anomalous events is very
relatively
high.
a
4.1).
In
the
number of cases in which the probability of
high
even
though
the
-10 C
isotherm
is
This is interpreted to indicate that while there is
- 125 -
Mesoscale Phenomena
a weak relationship between the height of the -100 C isotherm
probability
of
anomalous
events,
and
the
it is not likely to be of primary
importance.
- 126 -
Mesoscale Phenomena
%0 (5)
00. 0..
In
o
x
cio)
x
(3)
80.0-
L.W
Symbols
60.0O (4)
O W
40.0
02
o
0-
(11)0
L-
0
Gale
X
Norway Winter
Takeuti 1984
Japan May - Sept
Funuki
/
020.0
O
0. )
.0
Sweden Summer
Takeuti
1980
'
(12) (6)
0
2.0
0I
I
I
I
4.0
3.0
Wind Shear
5.0
I
7.0
km)
(m/s/
SJapan
Winter
Brook et al,
6.0
A
P 3
March
22:30
Figure 4.1: Probability of the
a positive flash
of
production
related to magnitude of wind speed
shear (Takeuti, 1984).
Numbers in parenthesis
indicate date during GALE
(00.
(1o) (5) (3)
A
(I)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
-n
u,
U-
80.-0 "0
.0-
(2)
0
.0
60.
N
0
a
(It)
0O
40 .0-
0
20 .0
(12)
o=0
.0U%
1.0
2.0
3.0
Height of -IOOC
(12(
4.0
5.0
Isotherm
6
6.0
7.0
(Km)
the
Probability of
Figure 4.2:
flash
a positive
of
production
related to the height of the -10C
isotherm (Takeuti, 1984).
127 -
Jan.
Jan
Jan
Jan
Feb
Feb
Feb
Feb
Feb
Mar
19
00o
20
12_
Mar
12
122
14
12 2
Mar
26
30
2
I 22
5
6
002
17
122
21
122
OOZ
12
Mesoscale Phenomena
4.3
Symmetric Instability
charge
is
region
that
is
One idea
charge region.
wind
the
causes
shear
charge region to be advected downwind of the negative charge
positive
An idea suggested in this thesis is that
region (Brook, et al, 1982).
the
vertical
positive
to the negative
relative
displaced
horizontally
the
that
The heart of the tilted dipole hypothesis is
if,
result
would
updraft
operates
mechanism
charging
a tilted updraft.
along
for
example,
to
assessing
the
A tilted
were
atmosphere
symmetrically unstable.
In the simplest approach
stability
is
assessed
whether
determining
or
displacement
the
along
instability,
convective
displacing a parcel of air vertically and
by
net
the
buoyancy
force
displacement.
is
opposed
In the case of symmetric
instability (see appendix A), the parcel is displaced in
direction,
rather
than
in
the
the
to
vertical direction.
a
slantwise
Pseudo-angular
momentum (M) is defined as the sum of the velocity and the product
the
Coriolis
parameter
(f) and
of
the distance along the line of the
cross-section from the origin (x) to
the
evaluated, i.e.
M = v + fx
- 128 -
point
where
M
is
being
Mesoscale Phenomena
In symmetric instability theories, the quantity M is
be
conserved
slantwise
than
the
slope
the
of
slope
the e
surface,
a
If, however, the slope of the
displacement will be stable.
eesurface is greater than
to
Thus, for a moist process, if the slope
in the motion.
of the M surface is greater
assumed
of
the
surface,
M
slantwise
displacements may be unstable (see appendix A).
Four upper air sounding times were chosen to evaluate whether the
convection
is
stable or unstable for moist processes.
symmetrically
The times were chosen as having:
hour
surrounding
the
(1) lightning
during
activity
the
sounding launch time, (2) sounding stations in
and beyond the extent of the
lightning
activity,
and
(3)
sounding
stations aligned roughly perpendicular to the direction of wind shear.
Figures 4.3a-d show the lightning activity for
the
sounding
launch time.
the
The symbols used are:
hour
surrounding
"x" for a negative
flash strike point, "+" for a positive flash strike point, and a
indicating
line
the location of the cross-section used for the analysis in
direction
figure 4.8, which is chosen as the perpendicular to the
of
wind shear.
January
26,
0000
UTC
(figure
exclusively positive lightning.
in figures 4.4a-c.
areas
of clouds.
4.3a)
represents
a
case
of
The synoptic situation is illustrated
The lightning is associated
with
three
separate
The location is east of a trough at 500mb, but west
of a front associated with the surface low in Georgia.
- 129 -
Mesoscale Phenomena
The second case, February 6, 0000 UTC (figure 4.3b), has a pocket
of
positive
(directly north
lightning
illustrated
is
situation
synoptic
the
of
cloud,
and
cloud band.
station
pockets of negative
with
a
The
TBW).
identifier
The primary
figures 4.5a-c.
in
associated
region of lightning activity is
two
between
sandwiched
lightning
single
band
of
near the deepest cells in the southern end of the
occurs
This area is in the warm sector of the
cyclone,
surface
and east of a trough at 500mb (figure 4.5).
1200
The third case, March 12,
(figure
UTC
is
4.3c),
highly
anomalous, with fifteen positive flashes and four negative flashes (80
percent positive).
of
cases
highly
The region of analysis was chosen to
anomalous
lightning
production,
associated
with
the
warm
front
lightning
is
edge of the long cloud band extending
northern
from Texas to Ohio (see figure 4.6a-g).
the
two
and two cases of
The region of
primarily negative lightning production.
provide
This cloud band is
north
of
associated with the surface cyclone, and east of the
500mb low.
The fourth case, March 14, 1200 UTC (figure 4.3d),
the
most
electrically
strike points.
shown
in
the
storm
active
frontal
eventually
become
of
bands
as
analysis for March 14, 0000 UTC, figure 4.7b.
The eastern band moved slowly compared to the western band
they
part
in GALE and shows two bands of
This storm was associated with two
surface
was
one
cloud-mass,
chart for 1235 UTC (figure 4.7f).
such
that
shown in the radar summary
Both bands are organized such
that
the center of the positive strike points is north of the center of the
- 130 -
Mesoscale Phenomena
negative strike points.
front.
The lightning is associated with the
surface
As in the other three cases, the 500mb height field indicates
the lightning region is east of a trough, in an area of
southwesterly
winds.
The potential for moist symmetric
taking
a
cross-section
instability
temperature.
The
figures 4.8a-d, respectively.
points
assessed
by
perpendicular to the direction of wind shear
and contouring the fields of pseudo-angular
potential
was
momentum
and
equivalent
results for the four cases are shown in
The latitudes
of
the
at which x equals zero are listed in the figure captions.
The
latitude and longitude coordinates
appendix
B.
of
and
the
longitudes
stations
are
listed
in
The orientation of the cross-section in degrees is also
noted in the figure captions, and is taken in coordinates
those used for wind directions.
similar
to
The locations and orientations of the
cross-sections are illustrated in figure 4.4.
January 26, 0000 UTC,
slantwise
4.8a).
shows
a
near
neutral
configuration
to
convection through the region of lightning activity (figure
Between AYS and TBW there is very little change with height in
the values of equivalent potential temperature, indicative of vertical
convection in the recent past.
instability
in
There is an
indication
of
symmetric
a layer from roughly 650 - 400mb over AYS and of weak
conditional instability to upright convection above 600mb over TBW.
- 131 -
Mesoscale Phenomena
The cross-section for the
(figure
over a
4.8b),
case,
depth
for
the
of
the
troposphere.
6,
0000
UTC
In
the
contrast,
third case, March 12, 1200 UTC (figure 4.8c),
shows evidence of slantwise adjustment,
slantwise
February
is suggestive of the potential for upright convection
substantial
cross-section
second
i.e.
the
cross-section
neutral in the area of 900 to 500mb around the station HTS.
This suggests that slantwise convection is likely to have occurred
the
is
recent
past
in
the
area
of lightning activity.
problem with this interpretation is the absence of
NWS
in
One possible
radar
echoes
(figure 4.6f) and the lack of rainfall (based on NWS surface analyses,
not shown) In the vicinity of the observed
which we do not understand.
lightning
(figure
The southeastern end of the cross-section
between 99A and GSO shows evidence of past vertical convection.
14,
4.3c),
March
UTC (figure 4.8d), also shows evidence of the potential for
1200
upright convection, with slantwise instability suggested in the
lower
troposphere between SSC and AHN.
The conclusion from these analyses, is that the winter
that
occurred
during
GALE
was
associated
lightning
primarily with vertical
have
instability, although conditional symmetric instability may also
contributed
in
some cases.
Although there was evidence of slantwise
instability without upright conditional instability in
the
other
one
case,
in
three, conditional instability also existed, and therefore
symmetric instability does not appear to be a necessary ingredient for
the production of positive lightning.
dipole hypothesis requires a
tilted
This implies that if the tilted
updraft,
- 132 -
there
must
be
other
Mesoscale Phenomena
sources of tilted updrafts besides symmetric instability.
The analyses done in these two sections are directly
the
tilted
dipole
hypothesis.
related
to
The original result by Brook, et al,
(1982) indicating that wind shear in the vertical is directly
related
to the percentage of flashes with positive polarity has been confirmed
along with a cut-off value for the production of positive lightning at
1
m/s/km
of vertical wind speed shear.
that the height of the -10oC isotherm
percentage
The result of Takeuti (1985)
is
inversely
related
to
of flashes with positive polarity is also confirmed.
the
This
continues to support the idea that positive lightning is produced when
the
positive
charge region is displaced horizontally relative to the
negative charge region, and is lower to the ground than in the typical
summertime thunderstorm.
The idea of a tilted updraft resulting from symmetric instability
was
also
examined.
Although
cases show that the lightning
convection,
and
the
analysis is not conclusive, most
activity
is
associated
with
upright
that symmetric instability does not appear to play a
necessary role in the production of positive lightning
lightning times.
- 133 -
during
active
Mesoscale Phenomena
x -negative, + positive
Figure 4.3a: Location of cross-section and stations for symmetric
instability analysis for January 26, 1986, 0 UTC. Lightning strike
points are also noted.
Kx
x
LSF
x
x
AQw
TBW*
FEBRUARY 5, 1986
23:30 - 0:30 UTC
x - negative, + positive
Figure 4.3b: Location of cross-section and stations for symmetric
instability analysis for February 6, 1986, 0 UTC. Lightning strike
points are also noted.
- 134 -
Mesoscale Phenomena
MARCH
12,1986
11:30-12:30UTC
x-negative
+ positive
Figure 4.3c: Location of cross-section and stations for symmetric
instability analysis for March 12, 1986, 12 UTC. Lightning strike
points are also noted.
AUL*
x
+
x
x
x.x
MARCH 14 ,1986
11:30 -12" 30 UTC
x- negative + positive
Figure 4.3d: Location of cross-section and stations for symmetric
instability analysis for March 14, 1986, 12 UTC. Lightning strike
points are also noted.
- 135 -
Figure 4.4a: 500 millibar chart for
Solid
January 26, 1986, 0 UTC.
lines are height contours, dashed
lines are temperature.
Surface chart
for
Figure 4.4b:
Solid
January 26, 1986, 0 UTC.
are
lines are
isobars,
fronts
analyzed.
Mesoscale Phenomena
UTC.
Figure 4.4c: NWS radar summary chart for January 25, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
- 137 -
500 millibar chart for
Figure 4.5a:
Solid
February 6, 1986, 0 UTC.
lines are height contours, dashed
lines are temperature.
for
Surface chart
Figure 4.5b:
Solid
UTC.
0
February 6, 1986,
are
fronts
isobars,
lines are
analyzed.
Mesoscale Phenomena
__
Figure 4.5c: NWS radar summary chart for February 5, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
- 139 -
UTC.
1C(D
42
CO
0
-2474
X
Be
6 1
"7~r
14 a1
2
S3 a
\16
1
5
26
'42
S18
3
\k2
It'4
;ij
i3
'
--
'r,
22-'2;
:
I-
25,4
22 M4
z-23
r
'0
-0
176
a
.94
13(
24-3
3
o
S18
A
1-1=5s;~e
33
2
Z v
36
27 240
-~-sff1
.
Marc
a re i
n
f
a
6616
'8 & 1;-;
'14
62
I
m
1(
a
:j
IQio
46
3a
-M
I
11
1
13?
46
-165;ut~
95O741.
h-'"e
(D
0
1
233k
ft'4'
13
20
0,
O
/
b
. .:'~
8
.
1
3~~53y4I
2
Z41
if
01
S.
'I~~~~
u'
:LO'
/
/
L:
l
1
27o
3
-- ;
?.~
/ 51
are tepeau.
I
~H1W62-
"I:!
6
;P
66
?2 JB
61
.-
71
\ 12
I
.
7~51
sln
Q3'
118*33
1 113
63
"
6
68G6
Figure
March
500 millibar chart for
4.6a:
Solid lines8218
12, 1986t 0 UTC.
are height
contoursf
are temperature.
dashed
lin
m
63~
044
7
6-L~l-oZ
Soo18I
7,\
?S
_Ole861I~
fr
Figure
March
are
4.6b:
12,
059\
Surface
1986, 0 UTC.
isobars, fronts
chart
for
Solid lines
are analyzed.
(D
CO
*
.Cj/.C
/
(
-12
C
I-l
!7~~
10=;i
r-
B ,2 6919 1.a
2 A-0-1
/
S1
! ,
-3 :14
a,
-1
3
8
T
S4
-
(D
36,
Utz,
Fiur
4.d
uraechrpo
34?-
;_
41
2
fib
-32
31
54
MarchA
are
i4, a1
-1
12
a
312
57
69:4
1963.T.
a
f
4 -J':
45
1~64
ie
(
64
24"bs'
LW
noi
..
-'II
'2V 3
HIGH
C---------------
V-1
1
?
-07
~L,/K'
j
58 1
S? 09
OCOam
42
Figure 4.6c:
March
13r
69
551C
51-5
1 !
75156
16idlie
1986r
are height
are temperature.
.12
j
500 millibar chart for
0 UTC.
dashed
contoursr
2
7
lines
~6~~7-~
i~
5
J
for
chart
Surface
4.6d:
Figure
lines
Solid
UTC.
0
13t 1986,
march
are isobars, fronts are analyzed-
Mesoscale Phenomena
Figure 4.6e: NWS radar summary chart for March 11, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
-
142 -
UTC.
Mesoscale Phenomena
Figure 4.6f: NWS radar summary chart for March 12, 1986 12:35
UTC. VIP levels are contoured, maximum cell heights are underlined.
Figure 4.6g: NWS radar summary chart for March 12, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
- 143 -
UTC.
(D
0
fo
EO
zl
19261(
14,
.'
l;,
8066
),
t ...
32
g.
-r.,
I
t .
.
it
i2
%,
I
)
!1
701"
576
'f;
.3M
F
43414
- 4
1
R
I
53;
5 1
40
6
401
4,4.
Mach14.186.
UC.Soidli/
3'J
+,~ 326
*
8 4
41IN
46
mprtr
are te 1,.
1
Marc
02
'
(:-'€:
44
,4q
, '- I),.',0
o\
_.,.-.
3 1"''
4-'11'r
3201~i'!
.
%5 .0o,
r
S1
+ - / .. . .•
to
,
<-
ar
jj
f l
a ioranl0
aree53LN)552
~I
M1~'
'
4",8
512
34
rr
5?t
41
d
j
Yl
53
~ltu
43
l
50
;8
58X6
t
1 -14-~P
70 6/1.
'5 :6
11
ia
Figure 4.7a:
500 millibar chart for
14f 1986f 0 UTC. Solid lines421
\
132(
4-
March
areheihtcontoursr
are temperature.
dashed
lines
5
10
Figure
4.7b:
Surface
March
14, 1986r 0 UTC.
are
chart
for
Solid lines
isobars, fronts are analyzed.
0
ZII
-
~
W /lc/<
'
....
6
*~
'$G ?
.'-4 --
'_*.
s.
IGI
Z,
das
areeight
tat"" II
'~
(D
13
1--
" :"""'''Sufc41d
Figur
HIG
><~
a
F_ M1
0
1
15,
MarIc
582
7
Ilie
aly
r,
lo
Figurje
c
Marh
sbr,
S_
r .
"
4.7d:
50i so; f
4
W'!
/101
100
5,
oi
cart
j,'r
m
tepeatre
42 5'
5
2
31:
,1
IN
~
ar
i1
are isbas
are~J
0(D
4
3
_OV
"
""
.t
""I
"--- I~~ls~s~~c
\
- /1-IV
12,
1
6
-,3-
.3s*
I.-Io
(P
O
D
o
arre
?
rnsae
Sufc
0Uo 53-s chart
C1
har
51.
lne
?4 114li
nlzd
fo
li
fo
Mesoscale Phenomena
Figure 4.7e: NWS radar summary chart for March 13, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
-
146 -
UTC.
Mesoscale Phenomena
Figure 4.7f: NWS radar summary chart for March 14, 1986 12:35
VIP levels are contoured, maximum cell heights are underlined.
UTC.
Figure 4.7g: NWS radar summary chart for March 14, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
UTC.
-
147 -
Mesoscale Phenomena
E 500
-40
for M.
850
300
1000
AHN
T8W
AYS
LSF
Figure 4.8a:
Symmetric instability cross-section for January 26, 1986,
Origin of x at 35.0 north latitude, 84.7 west longitude,
0 UTC.
aligned with wind from 344 degrees. Units are Kelvin for E , and m/s
for M.
E 50(
60
tJ
... ,
i,i
:D
hi
0
(n
30
50
0.
40
CKL
LSF
'
AQQ
TBW
Figure 4.8b: As in 4.8a but for February 6, 1986, 0 UTC. Origin of x
at 34.0 north latitude, 87.2 west longitude, aligned with wind from 335
degrees.
-
148 -
Mesoscale Phenomena
Z50
325
I
325
S
S500
320
w
1
50
60
70
700040
DAY
I
I
HTS
99A
I
GSO
Figure 4.8c: As in 4.8a but for March 12, 1986, 12 UTC.
at DAY, aligned with wind from 322 degrees.
Origin
of
x
250
330
60
325
E 500
X60
S700
50
40-
325
30,
AYL
SSC
AHN
Figure 4.8d: As in 4.8a but for March 14, 1986, 12 UTC. Origin of x
at 36.4 north latitude, 85.0 west longitude, aligned with wind from 317
degrees.
- 149 -
Mesoscale Phenomena
4.4
Bipolar Lightning Patterns
This section will present six cases of bipolar lightning patterns
in
Appendix C lists the characteristics of each of the bipoles and
wind.
its
relationship
the
orientation and the direction of the geostrophic
bipole
the
between
large scale and
their
emphasizing
environment,
mesoscale meteorological
to
them
relate
will
It
orientation.
the
size
as
such
similarities and differences in their characteristics,
and
examine
will
It
locations.
lightning
cloud-to-ground
the
to
information
provided
reader
than
the
of
source
a
as
is
It
environment.
contents
of
more detailed
graphical
the
presentations in this section.
A bipole is defined as a
strike
points
in
figure
"+"s
4.9.
An example of a
of
negative
flashes.
The
of
mean
is
between
the
strike
locations of each
positive
will be referred to as the geographic centers.
is defined as the distance
bipole
represent
the
polarity is marked by "N" for negative and "P" for
and
the
indicate the strike points of positive
flashes and tend to be northeast of the "x"s, which
points
in
pattern
lightning flashes that lower positive and negative
of
charge to ground" (Orville, et al, 1988).
given
polarity
"systematic
geographic
flashes,
The bipole's size
centers
of
the
positive and negative flash locations, and is illustrated with a thick
line.
In this
is
orientation
example
the
bipole
is
size
as
is
kilometers.
Its
determined by drawing a line from the negative center
to the positive center and taking the angle
line,
110
done for wind directions.
- 150 -
between
north
and
this
For example, if the positive
Mesoscale Phenomena
center
was
orientation
directly
would
north
be
of
180
the
negative
degrees.
For
center
the
the
example
bipole
bipole
the
positive
and
orientation is 230 degrees.
In order to accurately define the locations of the
negative
geographic
centers
number of flash locations.
at
least
ten
interest.
one
flashes
it
is
necessary
to have a sufficient
For this study "sufficient" was defined as
of
each
polarity
per
hour
in the area of
For this reason the evolution of the bipole was studied
hour intervals beginning and ending at the aforementioned cutoff.
Only bipoles lying entirely within the
main
area
of
the
lightning
detection network and within the GALE time period were studied.
were six bipoles meeting the above criteria.
the
date on which it began, e.g.
bipole
locations
the bipole beginning at 1700 UTC on
illustrated
are
represents a bipole.
The arrows
center
the
and
end
at
There
Each case was "named" by
January 23, 1986 will be referred to as the January 23 bipole.
six
in
begin
in
at
figure 4.10.
the
negative
These
Each arrow
geographic
positive geographic center to indicate each
bipole size and orientation.
The centers were determined as the
mean
location of all flash strike points detected in the bipole area during
the lifetime of the bipole, i.e.
when there were "sufficient" flashes
to define a bipole.
When studying lightning it is helpful to have a
of
the
synoptic
environment.
As
general
was demonstrated in section 4.2,
lightning is usually associated with upright convection.
of
the
bipoles,
the
picture
In the
case
convection was sufficiently intense to produce
-
151 -
Mesoscale Phenomena
lightning for several hours; the shortest bipole
lifetime
six
being
hours, and the longest fourteen hours.
of
The lightning area for the January 23 bipole is off the coast
South
and
North
Carolina.
The radar summary chart for 2335 UTC on
with
January 23 (figure 4.11c) shows that the lightning is associated
an
area of clouds that is entirely over the ocean by this time, which
occurs in the middle of the bipole lifetime.
cold
front
at
the
surface,
moved well out to sea.
The
area
behind
is
a
associated with a surface low that has
height
The 500mb
field
trough
a
indicates
coincident with the lightning area.
to
similar
the
January 23 environment in that (1) the bipole occurs
in
the
500mb
height
area
bipole
behind a cold front and (2) the
trough
is
4.12)
The synoptic picture for the January 30 bipole (figure
field.
is
coincident
with
a
The area of cloud in this case
hour
seems to match the shape of the coastline at 0000 UTC, the first
of the bipole lifetime.
The February 5 bipole lifetime surrounds the February 6 0000
sounding
time
shown
in
figure
4.5.
As noted in section 4.2, the
lightning occurs in a band of clouds beginning in the Gulf
and
running northward along the east coast.
- 152 -
of
Mexico
The lightning is located
in the warm sector of a surface cyclone and just east of a
the 500mb height field.
UTC
trough
in
Mesoscale Phenomena
The February 21 bipole had the longest duration of the six bipole
It
hours.
fourteen
--
cases
is associated with a cloud mass that
moved through Ohio, Pennsylvania, and New England, ending
east
coast
the
at
at 2335 UTC, two and one half hours before the end of the
with
is
As in other cases, the bipole
bipole lifetime (see figure 4.13g).
associated
up
a trough in the 500mb height field, and is initially
associated with an occluded front at the surface (see figure 4.13c).
for
The March 6 bipole is located over the ocean and is too far
east
coverage at 2335 UTC (see figure 4.14c).
The
significant
the
lightning is in
associated
radar
warm
of
sector
with a trough at 500mb.
the
surface
and
cyclone
is
The March 10 bipole is associated
with a cloud band in Kentucky and Ohio at 2335 UTC (see figure 4.15c).
The area is just west of a surface cold front and is associated with a
height field trough at 500mb.
All six bipole cases are associated with a trough
height
surface.
in this section.
cyclone
does
500mb
the
in
This feature will be discussed in more detail later
The connection between the bipoles and
not seem very direct.
the
surface
Two bipoles occurred in the warm
sector of the surface cyclone, three behind the
cold
front
and
the
remaining bipole was associated with an occluded front at the surface.
Electrically active convection in the winter can
any front or sector of a surface cyclone.
- 153 -
be
associated
with
Mesoscale Phenomena
The most striking feature of the bipole and
its alignment with the geostrophic wind.
predicts an alignment with the upper
upper positive charge.
a
winds
which
direction
difference of five degrees for the six bipole cases.
For
Figure
The
winds
4.16
arrows
with
vary
in
direction
shows the geostrophic winds at 850, 700,
indicate
speed
and
direction
any way.
and
all
the
directions
is
not
February 5 has the least amount of directional
shear, with a variation of only ten degrees in direction with
where
of
(see appendix C for actual speeds and directions).
The dashed line is the mean of the four wind directions,
in
the
and
500, and 250mb.
weighted
advect
height,
height.
geostrophic
is
The bipole orientation is always within twenty
all six bipole cases the geostrophic winds do not
with
environment
The tilted dipole hypothesis
level
degrees of the geostrophic average wind
averages
its
are
to the nearest five degrees.
height,
There is no
pattern in the change of direction with height found amongst
the
six
bipoles.
In this environment where the wind direction varies
than
twenty
degrees
with
height,
the
4.17
shows
no
more
six bipole exist and show a
general alignment with the geostrophic wind as
Figure
by
well
as
wind
shear.
a comparison between the average direction of the
wind labeled "W" corresponding to the dashed line in figure 4.16,
the
bipole orientation, labeled "B".
The lines labeled "C" represent
the cloud distribution and will be discussed later.
January
23,
January
30,
February
5,
and
In
four
154 -
cases,
February 21, the bipole
direction differs from the geostrophic wind direction by no more
-
and
than
Mesoscale Phenomena
five
degrees.
In
more north-south
the remaining two cases the bipole orientation is
than
the
wind
direction.
In
other
words,
the
positive geographic center of strike point locations is north of where
the center of positive charge would be advected to by the
wind
if
the
negative
geographic
center
was
geostrophic
coincident
with the
negative charge center.
Both the inverted dipole and tilted dipole hypotheses
between
relationship
shallow
To check this
lightning.
performed.
A
cloud
relationship
the
was
following
analysis
was
determined by comparing the
relative locations of the center of the tallest clouds and the
of
a
and the production of positive
clouds
"orientation"
predict
center
the shallowest clouds in the same cloud-mass, as documented by the
NWS radar summary
different
charts.
vertical
The
development
intent
in
was
to
compare
clouds
of
similar environments in order to
isolate the effect of vertical development
on
lightning
The
if
the taller clouds were
orientation
was
defined
such
directly south of the shallow clouds
degrees.
that
the
orientation
by
would
be
180
This orientation was determined for each radar summary chart
available during the bipole lifetime, and the average
shown
production.
the
line
labeled "C" in figure 4.17.
orientation
(See appendix C for
actual orientations, and the variation in orientation over the
lifetime.)
-
155 -
is
bipole
Mesoscale Phenomena
For January 23, the bipole and the cloud are
the
average
geostrophic
wind.
bipole
or
wind
aligned
with
Four out of the five remaining cases
show a cloud orientation with a stronger
the
both
orientation.
north-south
component
than
The final case, January 30, seems
backward in terms of the relative directions of the bipole, cloud
wind.
geostrophic
the
Since
relative
majority
to
of
the
of
explanation of this has not been determined.
the
bipoles
an
idealized
orientations
by
a
that
bipole,
of
similar
orientations
was
taken
and
"P",
and
from
the
The bipole
the six bipole cases.
geographic centers are located at "N"
emphasized
exhibit
environment, an idealized picture can be created.
their
Figure 4.18 shows
averages
The
and
the
bipole
thick line drawn between the two centers.
is
The wide
arrow represents the geostrophic wind directional average with height.
The
contours
are of reflectivity, the hatched area indicating higher
reflectivities, as would be seen on an NWS radar summary
higher
reflectivities
were
The
chart.
assumed to be associated with the taller
clouds, in order to orient the cloud-mass relative to the
bipole
and
wind directions.
The fact that the bipole is aligned with the geostrophic wind
a
is
strength of the tilted dipole hypothesis which relies on wind shear
to advect the positive charge region relative to the
region.
charge
The tilted dipole hypothesis goes beyond just predicting the
orientation of the bipole, by predicting
more
negative
that
positive
flashes
likely in a more highly sheared environment (see figure 4.1).
test of this relationship is now made for the six bipoles.
- 156 -
are
A
Mesoscale Phenomena
Figure 4.19 shows the wind speed shear in three layers
bipole.
The
difference
for
each
in geostrophic wind speed between the upper
and lower surfaces of the layers 850 - 700, 700 - 500, and 500
-
250
mb is plotted halfway between the heights of the two pressure surfaces
and was rounded to the nearest m/s/km.
noted
for all six bipoles.
- 700mb is nearly the same:
500mb
is
greater
than
Two
systematic
features
are
(1) The speed shear in the layer from 850
(2) The shear from 700 -
about 3 m/s/km.
the
shear from 500 - 250mb.
Because of the
relationship between temperature and electrical structure of the cloud
(see
section 1.2), the negative charge region would be located in the
layer from 700 - 500mb, and the positive charge in the layer above, so
that
the
negative
charge
region is typically in a layer of greater
speed shear than the positive charge region.
There is not a pattern between the overall percentage of positive
flashes
and
wind shear as predicted by the tilted dipole hypothesis,
when only these six observations
are
(see
considered
appendix
C).
Figure 4.1 however, showed that these six observations do fit into the
larger picture and support the idea that the percentage
of
lightning
that is positive increases with increasing wind speed shear.
The objective in this portion of the
effects
of
the
of
that
examination
as
dipole and tilted dipole hypotheses.
difference that has already been shown to play
production,
is
is
to
examine
the
environment on the production of lightning, and then
consider the results
inverted
study
the
they
relate
to
the
One main environmental
a
role
in
lightning
difference between land and ocean at the surface
-
157 -
Mesoscale Phenomena
period, and provides a source of low level moisture
GALE
the
during
ocean and half over land.
the
the
be made between
two
to
types
located
were
bipoles
In this study, half of the
during convection.
over
land
The ocean surface is generally warmer than the
(see chapter 2).
This allowed for a comparison to
the
of
effects
the
determine
different locations.
by
the
lightning,
largest
land
covering
bipole
square
330,000
covers
which
kilometers
(see
average the ocean bipoles cover twice the area of
The
bipole
size shows a similar pattern:
ocean
bipole,
appendix C).
On the
the
averaging
the
twice
the largest land bipole is
sizes.
land
bipoles.
land
There
correlation between bipole lightning coverage and
is
bipole
ocean
150 kilometers and the smallest is 180 kilometers, with
sizes
square
360,000
smallest
kilometers, which is roughly the same as the
covered
area
The three ocean bipoles are larger in terms of the
pronounced
no
bipole
The
size.
longer bipoles are not associated with the largest areal coverage, nor
the shorter with the smallest areal coverage.
Another way this ocean/land surface environmental
difference
is
of the bipoles:
ocean
bipoles moving southeastward and land bipoles moving eastward.
Figure
observed
4.20
in
is
direction
the
illustrates
of
movement
bipole paths, January 30 and February 21.
two
negative geographic center is represented by an "x" and
center
by
a
"+".
A
positive
dashed line connects the two illustrating the
bipole size and orientation.
which
the
The
Each bipole is labeled by the
the observations were used, i.e.
- 158 -
hour
for
the "2" bipole was determined
Mesoscale Phenomena
using all the flash locations occurring in the bipole area between 2 and
0300 UTC on January 30.
The distances traveled presented in appendix C
were derived by tracking the midpoint of the line connecting the geographic
positive and negative centers.
Although the paths of the positive and
negative geographic centers are not smooth the orientation of the bipole
seems consistent for January 30, varying only thirty degrees in orientation
(see appendix C).
Four of the six cases varied by no more than forty
degrees in orientation.
The other two cases varied by 135 degrees (March
6) and 150 degrees (February 21, see figure 4.20).
Consider the variation in the January 30 bipole first.
The initial
hour (0000 UTC) shows a bipole about 10 kilometers in length just north of
36
N, 73
W (see figure 4.21 for more exact bipole size).
deviation in the geographic centers
than 10 kilometers.
this case.
The standard
(mean of flash locations) is greater
Obviously, the true orientation cannot be resolved in
Now consider February 21.
The hours not showing an orientation
within 30 degrees of the overall orientation are 1300, 1800, 1900, and 2000
UTC.
The bipole sizes for these times are: 55, 30, 30, and 30 kilometers,
respectively.
Elementary trigonometry shows that the orientation can be
resolved only within a range defined by the
arctangent of the standard
Table 4.1 shows
deviation divided by half the size.
the bipole sizes,
orientations, and accuracy to which the orientation can be determined.
The
results show that the distribution of flashes cannot explain the majority
of the deviation of the bipole unless instrumental error is included.
-
159
-
Mesoscale Phenomena
instrumental error is included.
hour
13
18
19
20
unexplained
deviation
155
120
100
40
deviation from
overall bipole angle
165
165
145
75
accuracy of
orientation
10
45
45
35
standard
deviation
5
15
15
10
bipole
size
55
30
30
30
bipole
Determination of magnitude of deviation of
Table 4.1:
determine
orientation that can be explained by inability to accurately
geographic centers.
The central
question
is:
what
environmental
characteristics
shape the bipole evolution, and how does this relate to the production
one
will first be presented.
life-cycle
bipole
of
a
maintaining
its
moves
that
bipole
or
eastward
21.
shows
the
of
the
Thus far, the picture is
southeastward,
orientation (see figure 4.20).
roughly
No discussion of the
evolution of other bipole characteristics has been
4.21
picture
To approach this question, a
of positive lightning?
included.
Figure
evolution of bipole size for January 30 and February
These plots are representative of all six cases.
The bipole
has
a general characteristic of being initially relatively short, and then
The maximum size occurred in the hours of 0300
fluctuating in length.
UTC
on
January
30
and
1600 UTC on February 21 (marked with dashed
lines on figures 4.21, 4.22, and 4.23).
The next largest size
during 0800 UTC on January 30 and 2200 UTC on February 21
with a dashed line).
-In both cases the
after the largest size.
second
largest
What happens in the interim?
- 160 -
occurs
(also marked
size
occurs
Mesoscale Phenomena
Figure 4.22 shows that in the interim
(over the
entire
decreases again.
storm
area)
increases
Thus, large bipole sizes
the
negative
to
its
are
flash
maximum and then
associated
hours of higher percentages of positive lightning.
with
the
Higher percentages
of positive lightning are associated with shallow clouds in
environment.
rate
the
same
Are the clouds shallower at the times of largest bipoles
than during the shorter bipole times?
Figure 4.23 presents evidence that the cloud
some
point
in
top
is
at
higher
the interim than during the hours of longest bipoles.
On January 30, the hour of maximum cloud top height occurred one
after the maximum bipole size, and was 1,000 feet higher.
hour
On February
21, the hour of maximum cloud top height occurred four hours after the
maximum
bipole
size,
and is estimated to be 1,500 feet higher.
maximum in negative flashrate always occurred in the
the
interim
between
middle
The
hour
of
bipole maxima, but only for February 21 did the
maximum in cloud top height occur in the middle of the interim period.
For
January
30,
the
maximum
in
cloud
top height occurred in the
beginning of the interim period.
Why
do
bipolar
that
locations
convection.
they
lightning
do?
patterns
Lightning
occur
occurs
at
- 161 -
times
and
in regions of vigorous
Bipoles occur when vigorous convection is
the upper level colder air directly aloft.
the
combined
with
Mesoscale Phenomena
The
Figure 4.24 shows this relationship for the March 10 bipole.
the minimum in the height field at the same latitude as
of
longitude
the
a
the cold air aloft for
west,
farthest
Each line represents the location of
field was analyzed.
height
different
the
with
time,
The layer
initially
850-500mb
from
time
earliest
The striped area
and progressing eastward with time.
represents the bipole lifetime.
which
at
the March 10 bipole (39.5N) is plotted against the pressure
shows a westward tilt with height in the location of colder air aloft,
and then loses its tilt.
As
the bipole occurs.
eastward,
tilt
to
starts
aloft
air
cold
the
The bipole ends, and the cold air gains
a greater eastward tilt with height.
This pattern is visible in all six cases to varying degrees.
case
"average"
each level and
of
lengths
was
created by taking the mean of the longitudes for
bipole.
time,
initial -- at least
the
data
twelve
bipole,
the
Because
were
hours
bipoles
bipole.
If more than two sounding times occurred
two
slots for each bipole.
4.25.
a
during
the
bipole
of them were averaged in order to have four time
Missing data were linearly interpolated.
The result is a picture of the movement of
during
beginning
and final -- at least twelve hours after the
ending
then
bipole,
the
to
varying
for
four time slots:
into
organized
prior
endured
bipole,
lifetime,
An
the
cold
air
aloft
hypothetical average bipole lifetime and is shown in figure
The figure shows that the bipole occurs when
has no east-west tilt between 850 and 500mb.
the
coldest
Geostrophy requires that
the wind direction at all levels be the same in this situation.
- 162 -
air
This
Mesoscale Phenomena
is
believed
to
be
the
explanation
for
the
lack
of directional
variation in the wind shown in figure 4.16.
The results of this section have led to a picture of an idealized
bipole.
not
The
vary
bipole
appreciably
development
in
the
is aligned with the geostrophic wind which does
in
cloud
direction
with
geographic
center
The
vertical
is such that taller clouds are associated
with the negative geographic center, and
positive
height.
(to
within
shallower
ten
clouds
degrees).
with
the
The bipole
organization occurs when the cold air aloft is between a westward tilt
phase
and
an eastward tilt phase.
The bipole life-cycle consists of
growing and shrinking in length, with marked increases in the negative
flashrate between times of maximum bipole length.
- 163 -
Mesoscale Phenomena
Example Bipole
..
.
.
.
. ...
.
.
.
. .
.
..
X
:x
8s
81
82
..
.;
~
+
+.
'$qX
.
....
..
40
w
... .. ..:. . .. .. . . . ...
x
x
:
XX
)x
X
/'rbruary 21,
:10 30-
X
X
negoatve
4-positive(
N : negative geographic center
P : positive geographic center
Bipole size = 110 km.
Bipole orientation = 230 degrees.
Figure 4.9 Example of a bipolar lightning pattern.
- 164 -
S1986
1Z30 GMT
Mesoscale Phenomena
Bipole Locations
*
70
..
4
.35
January23
.38
..........Six Dpole
29
Areao
+.positive
Cases During GALE
79.8
42 ON
9. 0W to
xanegative ctr
ctr
ON
Figure 4.10 Locations of the six bipoles that will be examined in this
chapter.
Bipoles are labelled by the date on which they started.
Arrows point from the negative geographic center to the positive
geographic center.
- 165 -
500 millibar chart
Figure 4.11a:
for January 24, 1986, 0 UTC. Solid
lines are height contours, dashed
lines are temperature.
Surface chart for
Figure 4.11b:
Solid
January 24, 1986, 0 UTC.
are
fronts
isobars,
lines are
analyzed.
Mesoscale Phenomena
Figure 4.11c: NWS radar summary chart for January 23, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
- 167 -
UTC.
Figure 4.12a:
500 millibar chart
for January 30, 1986, 0 UTC. Solid
lines are height contours, dashed
lines are temperature.
Surface chart for
Figure 4.12b:
Solid
0 UTC.
1986,
30,
January
are
fronts
isobars,
lines are
analyzed.
Mesoscale Phenomena
Figure 4.12c: NWS radar summary chart for January 29, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
- 169 -
UTC.
-2
-45
45
1.9 ',6,'
-30
,30
r,40
34
a
1
-27 3N
5 a
2
11-112
2-1
\2
4.1
\L
-4 2-
-~
ar
~33
lanalyzed
........
.B
'
A
~;
V"32:2%i.nF-I
421
i!
.
iL1;l69
194\
2
593
525S
gr61
1.
41
Y~c3
~P
1
6
54
50
0
- !1\
.
IrntEr
o
55;
'12
1
5
8
q lr
2
-0 !9
1,
Ila
a i , ,
,
Figure 4.13a:
500
millibar
chart
for February 21, 1986, 0 UTC. Solid
lines are
height
contours,
dashed
lines
are temperature.
=
-6
1
SLe
7
lines~~~~L1
cotors ar dahe
hegh
Z~
I
3-
7
February
21
196
254,
41
lines ar
isbrs
28-9
..
3
.1b4',
-8*
vi>-
i',,
23
5
tepraue.f- Fiur
.7
13'(2Q
C:)
I
3
's*
a'~b
~
3
/
I
15
7"-__
,o
(D
'LJSQ 18
21
f2r-./NtJ
-159-1
91 ,~
W7
S--
lines~~~t
~
-
-3
-32!
(-
-201
/I
I-)'L
(
19
L
.
CD
21
291
HIGH
.J\
00
27&.-IGH
28 3%,
OU
-42,
IL9
'InI
8
I77
213
6
6
7
4
66
;
6'513,
4=rr
5
.1610
7
1
Figure
4.13b:
Surface
chart
for
February
21,
1986r
0
UTC.
solid
lines
are
isobars,
fronts
are
analyzed.
(D
2
.
.. .
56I
I
/
I
o
552
--
"' "14
*1
481~i
_
1
1.4Z.61~ 57I--
238
1
46
43
-2
14,
*,
4~? b
1%
s-
I
J,,,1
1
1,;
-
4=2
i
I3,2
-9
0.
(D
.
'
&
4
*P
941
1
to
0
(D9
227 2I1
3 Ce!
a
All
Icnalyze.20
3
I
.
-31~ !4ZY
a
ot
14
0;
\\r
lines
are
heightI
s
contours
dashe
Z","
33s
2
u',
?5s
8?
Ia
;Ir4
analyzed.
4 13 0
Figure~~~
mliarcat
fronts
'g,
11-1-
l~
~~
~.- il~ldr
65 (
.
63
5-
'9' m,
63a
42
500 millibar chart
Figure 4.13c:
for February 22, 1986r 0 UTC. Solid
lines are height contours, dashed
lines are temperature.
11162
.81
4,0
,\E
80
n
?5 IV
4
are
Q
2892
!73 ie
'
45-9(
0o
1400-i
35LOZI
14.
112~
206-e
/
..
4
A0I
,J
3" I
-
l1ies2,
-4
,
X
2'
1028 1 ,3, U
5-
,q
-
I
c"
-
54
fo21
2.I
e1
,e
LOW
if.-,,,
- "<' '
I""
.
*
-2&
-00A)
4
2121
5
-40
/
5S I
\\2
ISO//
16,
,
;30
-45
,
~
.,
69,~
A
Figure 4.13d:
Surface
chart
for
February 22,
1986F
0 UTC. Solid
lines are
isobars,
fronts
are
analyzed.
Mesoscale Phenomena
Figure 4.13e: NWS radar summary chart for February 20, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are underlined.
Figure 4.13f: NWS radar summary chart for February 21, 1986 12:35 UTC.
VIP levels are contoured, maximum cell heights are underlined.
-
172 -
Mesoscale Phenomena
Figure 4.13g: NWS radar summary chart for February 21, 1986 23:35 UTC.
VIP levels are contoured, maximum cell heights are underlined.
-
173 -
N
124
0"0
0
(0
I
@
iuG
*132022
I
25
27
--
"LI
,,'--I,
I _I
IQ
-C2
a+,2l
.c
K+r.;l---1
.
--
Z4
t
2"-4
!26,
14I
S
ii
linear
/4e
for
Mac
7, 1986,
0 UC.
Figure!6C
ilba
line
ar tmertue
3
oli
4q3
hr
lc~:La!
are
34f
50
4
4
~S
7,
1986,
0 UTC.
58 IS
are nalyzed.
Surfce
chrt
fo
Figure
4.14b
Figure
4.14b:986r Surface
chart lne
for
Sli
0 TC
7,
March
areisoar,
frMarch
4n,
le, 0.
ronts
srr
4
41
2
551
l
S
a
/277
tepertur.
4
9
'C8
392
.
05QI,"
Q U10
f 261(?-
2
are
45-i
..
41
b
1948
-2
--
N
...
0.
A l -I'1
4
1 0
...
3
,
,CD'41
49ra:AN
LM
1:-
2
-,39,
.1
,1
;45'
(D
(.-31
7.I
4 ,
Solidlie
font
ae aalyed
Mesoscale Phenomena
Figure 4.14c: NWS radar summary chart for March 6, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
-
175 -
UTC.
98
-,j
'
/0
a
42
6522
..""
,
1
92 98
0%
(D
.
O0
-
v
-" :
5IQ -
"i-h
23~
N4~
~37c1721,
-
y
0
_A
4
!-P
(D
&4
4 4P
'(D
0-
324
55'
.
">
60£4
,
"7
'o2
'-"4
19Ira
2
3-0-
L
.,_,.--
~ 211
i
i
'2
0= i
LL4Q £4 jC
1.1
'0
.
,
I~~l?
4
~;
6
5:44
.4
are isobars, fronts 6 are
analyzed.
r:/
:
60&N'.S1S
c 45a
Fiur bar ---
forMarh
for
March
lines are
1,
500.
height
1986, 0 UTC.
,
fo
~I~b-!
5ljf
Surfce
4.15b
Solid
contours, dashed
are temperature.
,
chrt
6
415a:
500millibarchart
Figure
1986, 0 UTC. Solid~r:
ii,
6 5
1
E5
Figure
March
Marh
ar
sbr,
Figure
2,i
1986,:, 0 UTC. Solid lines
rnsae
3. nlzd
.lines
4.15b:
4
Surface
chart
for
Mesoscale Phenomena
Figure 4.15c: NWS radar summary chart for March 10, 1986 23:35
VIP levels are contoured, maximum cell heights are underlined.
-
177 -
UTC.
Mesoscale Phenomena
-
Geostrophic Winds
10 m/s
February 5
January 23
February 21
January 30
March 10
March 6
Figure 4.16: Comparison of geostophic wind speed and direction for the
levels of 850, 700, 500, and 250 millibars.
- 178 -
Mesoscale Phenomena
B
January 23
januav 23
w
February 5
W
B
February 21
January 30
March 10
March 6
Figure 4.17: Comparison of orientations of mean geostrophic
bipole, and cloudmass. See text for further explanation.
- 179 -
wind,
Mesoscale Phenomena
Idealized Bipole Picture
North
Figure 4.18: Fictitous bipole that was created to illustrate the mean
relative orientations for the six bipole cases. Large arrow indicates
geostrophic wind direction, solid line represents bipole with negative
center N and positive center P. Contours are of cloud reflectivity in
imitation of NWS radar summary charts.
-
180 -
Mesoscale Phenomena
Wind Shear vs
11
I
I
I
I
Height
i
I
I
10
9
8
,7
3
2
1I
1
I
I
I
0
2
3
4
Wind
9
8
7
5
Speed Shear (m/s/km)
10
11
12
Figure 4.19: Comparison of the wind speed shear in the layers of 850 700, 700 - 500, and 500 - 250 millibars for the six bipole cases.
- 181 -
Mesoscale Phenomena
t,.
. . . . . .
.
..
..
*
...
.
.............
2'
rr
. . . . . .•. . . . . .
-L... .. _ , .
-
Y
I
.
*
•s
.
'.
..
7r...
',1,6
*.II
.
..
.
1/ I
v.
.3:..
.
.
...
76e6
.r.,@ITr
.r
.,.t.r
.rr
Figure 4.20a: Movement of the January 30 bipole.
Each dashed line
represents the size and orientation of the bipole for a given hour,
which is labelled at the negative geographic center.
Positive
geographic centers are marked with a "+".
Figure 4.20b:
As in 4.20a, but for the February 21 bipole.
- 182 -
Mesoscale Phenomena
t
10
S70
N SO
o
/
Figure 4.21a: Distance between the geographic centers as a function of
time for the January 30 bipole. Error bars were estimated using the
standard deviation of the mean latitude and longitude for each
geographic center.
110
180
70-
E 60
a s
0
-; 4 i
Figure 4.21b:
As in 4.21a, but for the February 21 bipole.
- 183 -
Mesoscale Phenomena
Fiosn-a te
vs
Time
7
1
5
4L
4
-
AY
J3
CE
0L
-
1
2
I
3
4
S
Hour
7
E
(UTC)
10
8
I11
Figure 4.22a: Storm-wide flashrate for the January 30 bipole as a
Negative flashrate is shown as a solid line with
function of time.
boxes, positive flashrate is shown as a dashed line with triangles.
vs
7
I
Time
asrroI
1'
5E
5C -
o--I
L
I
I.
i
0
10
~..
12
1...
i
12
13
Figure 4.22b:
14
15
18
117
hour (uTC)
19
20
21
2 ,3
24
As in 4.22a, but for the February 21 bipole.
- 184 -
Mesoscale Phenomena
P '
Cloud
.
,
'-
,I
uv
-lmQ
I
'
I
I
I
- r
I
--
7
.
'KI-
rur
I1
o
Clod
SBi
l~l
I
-I
@
I
I
I
1
2
3
4
5
H
igh
I
t
-1
vsTim
\
6
7
5
or( UTC )
1
15
i
I
I
5
11
12
15
IA
Figure 4.23a: Maximum cloud top height of the cloud~mass as reported by
NWS radar summary charts for the January 30 bipole.
34
0
Cloud
ke~th
us
Time
"1 i
C30
o2
-
o
I
\
I
-i
.L
-
L
L
!I
1i
11
1i
Figure 4.23b:
13
14
5S
1S
17
1
Hour (UTC)
is
2
21
3
24
As in 4.23a, but for the February 21 bipole.
- 185 -
Mesoscale Phenomena
ne,VL
3/10
.0n n2
197
I
3/11OE
"C
sin 12
3/12
~-3/12
02
Oz
Sr
S500
W
r
a
i,
iv
W
Cr
nu
e"f
!
I
I0001
I10
sIEn
100
i
•
\
\
Degrees
... 1
80
90
West
I
I
60
t(
L onllude
Figure 4.24: Example of the westward tilt of the cold air aloft as
represented by a trough in the height contours at a standard pressure
level. Each line represents a single time spaced twelve hours apart.
Hatched area indicates time during which a bipole existed.
250
700
a-
850
1000 100
Figure 4.25:
95
90
85
West
Degrees
80
75
Longilude
70
65
As in 4.24, but averaged over the six bipole cases.
- 186 -
Mesoscale Phenomena
4.5
Summary
The
first
analyses
of
this
chapter
show
that
the
GALE
observations are roughly consistent with previous analyses that assess
how the probability of
varies
with
producing
a
positive
cloud-to-ground
flash
wind shear and -10°C isotherm height (which is a measure
of proximity between the charge regions in the cloud and the
ground).
Both results are consistent with the tilted dipole hypothesis, as well
as the relationship between shallower clouds and increased
likelihood
of positive flashes.
The
idea
displacement
that
a
between
tilted
positive
updraft
may
produce
a
horizontal
and negative charge regions led us to
investigate whether symmetric instability was an important source
the
convective
activity.
for
Since a symmetrically unstable atmosphere
accelerates parcels in a slantwise direction, the resulting convection
would
have
a
tilted
updraft.
The result of these analyses suggest
that, although symmetric instability may occur in some
cases,
it
is
not necessary for the production of positive lightning.
Confirmation of the existence and basic
bipole
were given in section 4.4.
ranging up to 310 kilometers.
geostrophic
characteristics
is
the
Bipole lengths are confirmed to be
The alignment of the
wind is visible in all the observations.
bipole
with
the
In addition, an
alignment between the bipole and a "cloud orientation" is
It
of
documented.
shown that the orientation of the cloud top heights associated
with the bipole has a larger north-south
-
187 -
component
than
the
bipole
Mesoscale Phenomena
itself.
The
life-cycle of the bipole reveals that the length of the
bipole is related to the lightning activity.
are
observed
near
Minimum
to
lengths
maxima in the storm-wide negative flashrate.
bipole itself occurs when the cold air aloft has
west
bipole
moved
in
from
The
the
be directly overhead, resulting in winds that do not vary in
direction with height by more than thirty degrees.
- 188 -
CHAPTER 5
Conclusions
characteristics
and
of
analyses
of positive lightning.
both the likelihood of
documented,
with
began
thesis
This
a
producing
The dependence on latitude of
flash
is
the
active.
polarity
associated
convection
with
flash
the
that
seen
as
deep
twice that
clouds that are more electrically
latitudes in the GALE area, where temperatures
of lightning flashes is positive.
expectations based on both
the
inverted
are
activity
is
At the northern
influenced by the environmental conditions.
percentage
ocean
This is explained qualitatively by an enhancement
The picture presented is one in which lightning
strongly
warm
are
that
densities
was
The effect of
largely agrees with previous results.
observed over land.
of
flash
and
the Gulf Stream was examined, and it was found
surface
climatological
the
lower,
a
higher
This is consistent with
dipole
and
tilted
dipole
hypotheses.
An analysis of first stroke peak currents confirms
result
the
previous
(Orville, et al, 1987) that positive peak current averages are
1.5 times the negative peak current averages.
- 189 -
Winter currents of both
Conclusions
polarities are 1.5 times those observed in the Yearly and Summer data.
shallower
average,
this
considerations,
environmental
to
Due
and
12,
July
1986
14,
June
on
21,
(figures 3.3 and 3.8) associated with tall
1988
observed
summer clouds, and the larger currents
March
the
are associated with larger peak currents.
clouds
In support of this are the small peak currents observed
1987
on
that
suggests
and
3.16), associated with shallower
and
3.12
(figure
30
January
on
winter clouds.
More support for the idea that the environmental conditions shape
electrical
the
Huntsville
summer
scale
convective
studying
by
found
Evidence has been presented of three
phenomena.
the
is
activity
between
differences
The first
storms and the winter GALE storms.
difference is that on the average in the summer storms, positive flash
locations
were
than the tops associated with
3.4).
This
result
not
is
cloud top heights that were greater
with
associated
locations
flash
negative
consistent
inverted
the
with
hypothesis
hypothesis, and is consistent with the tilted dipole
if
it
observations
show
flash
positive
clouds.
anvil
locations
dipole
only
of
lightning
is
located
on
flash location accuracy can
rather
the
convective
affect
-
which
Firm conclusions on these points
consistent with both hypotheses.
are prevented on account
Winter
associated with
shallower clouds than negative flash locations (see table 3.4),
is
table
is assumed that in the case of a tilted dipole the majority of
the positive flashes are associated with tall
storm
(see
poor
scales.
the
190 -
accuracy
with
which
An example of how the
observed
cloud
top
height
Conclusions
associated
with
that location can be seen by looking at figure 3.3c.
An error of five kilometers in one
of
the
reported
flash
negative
locations could change the cloud top height associated with that flash
by as much as four kilometers, a thirty percent change.
second
The
difference
observed
lightning is the summer evidence for positive flashes
cloud-to-ground
field
with an initially large fair weather polarity electric
(see
surface
winter
and
summer
between
3.5).
figure
Winter positive flashes are noted during
times of foul weather polarity electric field (see figure 3.17).
winter
the
at
This
result is not consistent with expectations based on either the
inverted dipole or the tilted dipole hypotheses.
The third difference between summer
lightning
current.
of
1.5
in
is
phenomena
the
winter
and
cloud-to-ground
of the first stroke peak
magnitude
This was documented by season in figure 2.8, where a
difference
in
currents was observed.
current
factor
magnitudes between summer and winter
In chapter three the current magnitudes can be
compared between four storms shown in figures 3.3, 3.8, 3.12 and 3.16.
Results indicate a difference in seasonal magnitudes of
first
stroke
peak currents of both polarities.
In support of a common
phenomenon
in
both
seasons
are
three
characteristics
that winter and summer episodes of positive lightning
have in common.
These are (1) positive flash
negative
flash
locations
(see
section 4.4 for winter example),
figure
3.2
locations
for
downwind
summer example and
(2) a predominance of negative
- 191 -
of
flash
Conclusions
locations
in deep convection to the east and positive flash locations
in a trailing mesoscale band as was seen in
the
July
12,
1988
and
March 14, 1986 examples given in chapter 3, and (3) the association of
some positive flashes with bright bands seen in the trailing mesoscale
bands
of
July
12,
1988
and
March 14, 1986 (see especially figure
3.16e).
The existence of the bipole yields three points that can be
used
to evaluate the tilted dipole and the inverted dipole hypotheses.
(1)
Bipole lengths are on the order of hundreds of kilometers, an order of
magnitude
larger than the convective scale (see appendix C).
bipole is aligned with
4.17).
(3) The
upper
region
level
geostrophic
winds
(2) The
(see
figure
of positive flash locations is not distinct
from the region of negative flash locations in most cases, as seen
in
both chapters three and four.
A tilted dipole structure can be created in
way
several
ways.
One
is for the updraft itself to be tilted, such as would result from
The results presented in section
symmetric instability.
support
this
idea.
A
inside the cloud to vary
negative
second
with
direction
of
the
geostrophic
height
such
that
the
or
may
wind.
not
The
be
positive
charge
region.
This
not
third
interpretation
same
way
is
away
as
the
for
the
from
the
of the tilted dipole
hypothesis predicts a cloud electrical structure that would
- 192 -
and
The direction
the
geostrophic wind to advect the positive charge region
negative
do
way is for the horizontal velocities
charge regions become horizontally displaced.
of displacement in this case may
4.3
force
an
Conclusions
A
alignment of a bipolar lightning pattern with the geostrophic wind.
the
test of this interpretation is whether or not
is
wind
able
to
displace the positive charge far enough to explain the observed bipole
lengths.
This test can be approximated by a
that
the
simple
calculation.
time of peak positive space charge production occurs at the
time of peak lightning rate (0600 UTC on January 30 and
February
1900
January 30 and 2200 UTC on February 21).
wind speed as the average of the geostrophic
millibars:
UTC
winds
at
500
and
250
on
The estimate of advective distance is 400 kilometers
and 375 kilometers on February 21.
30
These estimates are a
does
not
idea of advection of space charge as the explanation for
the
bipole lengths.
scales
(0800
Estimate the geostrophic
factor of four larger than the actual bipole lenghts, which
support
on
55 m/s (200 km/hr) on January 30 and 35 m/s (125 km/hr) on
February 21.
January
UTC
Then consider the advection of this space charge from
21).
this point in time until the time of maximum bipole length
on
Assume
would
Advection of
tend
to
the
positive
space
charge
on
these
isolate the positive flash locations from the
negative flash locations, which is not consistent with observations.
A single tilted dipole would produce distinct regions of positive
and
negative
flash
locations (see figure 5.1a).
The negative flash
locations would be found associated with the main updraft region.
The
positive flash locations would be located downwind of the main updraft
region.
a
It is necessary then to consider a group of dipoles
bipolar
lightning
pattern (see figure 5.1b).
- 193 -
creating
Figure 5.1b shows a
Conclusions
sketch of a top view of a group of tilted dipoles.
all
negative
flashes
directly
beneath
The sketch depicts
a confined convective scale
region associated with the updraft (drawn as a circle).
dipole
is
exactly
relative
to
the
circular
updraft
located
region.
cloud-to-ground lightning presented is one of a
and
negative
flashes.
tilted
identical to the others in that it produces three
negative flashes and two positive flashes, all
place
Each
flash
location
regions,
The bipole created is much
with
larger
in
The
the
same
picture of
mixture
of
positive
clusters
of
negative
than
the
size
of
any
individual cloud, but not necessarily an order of magnitude larger.
One explanation of how the tilted dipole hypothesis
for
can
bipoles that are an order of magnitude larger than the individual
cloud sizes is to realize that not all tilted dipoles will
This
is
depicted in figures 5.2 a and b.
be
flash.
In
this
case
the
bipole
convective scale of the cloud.
changes
when
the
dipoles are not equal.
have shown that the probability
directly
related
to
higher
of
shows
right
side
how
this
picture
The results of this thesis
producing
a
positive
flash
is
Figure 5.2b takes
The left side of the picture
is
the
temperatures, and therefore the deepest cloud, and
the cloud with the least probability of producing
The
negative
exactly equal to the
is
temperature and cloud depth.
this relationship into account.
with
length
Figure 5.2b
equal.
Figure 5.2a shows a row of
identical tilted dipoles each producing one positive and one
side
account
has
lower
temperatures,
highest percentage of positive flashes.
-
194 -
a
positive
flash.
shallower clouds, and the
In these sketches the
bipole
Conclusions
sized
has been increased by a factor of 3 by including the effects of
temperature.
from
This effect combined with the increase in bipole
considering
a
group
of
length
tilted dipoles could result in bipole
lengths an order of magnitude greater than the convective scale.
The inverted dipole hypothesis lends itself
scales
to
organization
on
larger than the convective scale because it requires a mixture
of upright dipoles and inverted dipoles in order to create
lightning pattern.
figure 5.3.
a
bipolar
A simple picture resulting in a bipole is shown in
The upright dipoles are shown to the left (and are likely
to be associated with warmer air).
The bipole depicted in this sketch
is on a much larger scale than the convective scale, and is created by
two
distinct
regions
of flashes.
An alignment with the upper level
geostrophic wind is not evident from this picture.
mixture
of
upright and inverted dipoles, again with a preference for
the inverted dipoles to be
bipole
Figure 5.4 shows a
is
length
much
to
toward
larger
right.
the
than
Again,
that
the
alignment
of
the
expected
scale, but no
convective
It
alignment with the geostrophic wind is in evidence.
therefore,
the
is
concluded
bipole with the upper level
geostrophic wind favors the tilted dipole hypothesis over the inverted
dipole hypothesis.
The results presented in this thesis do
either
not
convincingly
favor
the tilted dipole hypothesis or the inverted dipole hypothesis
as being
lightning.
the
dominant
There
is
mechanism
evidence
dipoles and tilted dipoles.
The
for
for
the
the
production
positive
existence of both inverted
characteristics
- 195 -
of
of
the
bipole
do
Conclusions
tilted
the
favor
dipole
hypothesis,
but
not
conclusively.
question of the accuracy of the reported lightning strike points,
the
of
neglect
Further
results.
deeper
A
understanding
mechanisms and charge
dipoles
and
tilted
convective
the
work in this area could be done to link the
magnitude of the peak current and
cloud.
and
the effects of propagation between the origin of the
flash and the strike point lend skepticism to some of
scale
The
the
of
distributions
dipoles
might
electrical
of
structure
the
the differences in the charging
associated
yield
with
both
differences
distinguish between cases of the two electrical structures.
- 196 -
inverted
that
would
Conclusions
+ ++
+ ++
4-
-
+
+
+
4
bipole length
convective scale 1
Figure 5.1a:
Sketch of a single
tilted
dipole.
Cloud-to-ground lightning flashes are noted, with "x" for
negative strike point and "+" for positive strike point.
bipole length
convective scale
Figure 5.1b: Top view of a
Symbols as in figure 5.1a.
group
- 197 -
of
tilted
dipoles.
Conclusions
i
I---
bipole length
convective scale
Figure 5.2a: Row of identical tilted dipoles.
as in figure 5.1a.
bipole length
warm
I
i
Symbols
cold
convective scale
Figure 5.2b: Row of tilted dipoles illustrating
effect of temperature. Symbols as in figure 5.1a.
- 198 -
the
Conclusions
cS44
f
-+
V
--
.L
-4-
bipole length
convective scale
Sketch of one way of creating a bipole with
Figure 5.3:
a combination of upright and inverted dipoles. Symbols as in
ficure 5.la.
xx
- i-
)
--
bipole length
convective scale
Figure 5.4: Sketch of a second way of creating a bipole
with a combination of upright and inverted dipoles. Symbols
as in figure 5.1a.
- 199 -
APPENDIX A
SYMMETRIC INSTABILITY
Bennetts and
Results from several studies (e.g.
1979;
Hoskins,
Emanuel, 1983b; Wolfsberg, et al, 1986) have provided evidence for the
atmospheric
instability
symmetric
of
importance
phenomena.
a
variety
of
mesoscale
appendix will briefly review the basic
This
assessing
toward
ideas and approaches
in
symmetric
instability.
the
description that is adopted closely follows the basic approach
parcel
described in Emanuel (1983a,b).
Symmetric instability can
Consider
a
an f-plane.
--
= -
dt
assessed
in
the
following
way.
basic flow v(x,z), that is in geostrophic wind balance on
The equation of motion for v can be written as:
1 ap
dv
(1)
be
-
--
p ay
Fx
-
fu + --
=
-fu
FX
+ --
where t is time, p is density, f is the Coriolis parameter, u
is
the
in the x direction, and FX is the frictional acceleration in the
flow
x direction.
dx
u = -dt
But
df
and -- = 0
dt
- 200 -
SYMMETRIC INSTABILITY
so in the absence of viscous effects
(2)
(F = 0)
d(v+fx)
dv
-- + fu = ------- = 0
dt
dt
In this case a quantity which is a local approximation to the absolute
angular momentum M can be defined by
(3) M = v + fx
>From (2), M is
conserved
following
the
motion.
The
inviscid
u
momentum equation can be written as
(4)
du
-- = fv,
dt
= f(v-v3 )
Assuming that the basic state v(x,z) is in geostrophic balance,
M =
+ fx
then
(5)
du
-- = f(M' - M) = fM'
dt
The vertical component equation is just
dw
(6)
g(e-e)
--.
dt
8
g'
-e
If it is assumed that the parcel and environment are saturated, then'
and8 should be replaced by
' and 5 in (6).
- 201 -
SYMMETRIC INSTABILITY
The basic state is assumed to be in thermal wind
assuming
balance
which,
saturation, can be written (Durran and Klemp, 1982; Emanuel,
1983a) as
dM
dv
g fdln
dz
dz
f
e
dx
Therefore, the vertical gradient of M and the horizontal
gradient
in
eare connected through the thermal wind relation.
The
susceptibility
stability
can
of
the
atmosphere
figures on following page).
acceleration.
symmetrical
and G, of
the
parcel
(see
If M' = M - M > 0, then from (5) there is
If 8' =
a net eastward (outward) acceleration.
or
moist
then be assessed by displacing a parcel in a slantwise
direction and assuming conservation of M
processes,
to
8
) -
>
0
for
dry
e
> 0 for moist, then from (6) there is a net upward
If
the
net
forces
act
in
the
direction
of
the
displacement, for a given slantwise displacement Ss
as = dxi + dzj
then the environment is unstable for displacements in
that
slantwise
direction (see figure b on following page).
In all
appropriate
analyses,
in
it
was
assumed
that
moist
processes
were
the region of convection, and therefore stability was
assessed relative to ( surfaces.
- 202 -
SYMMETRIC INSTABILITY
--
0
Displacement
a. Stable
Configuration
Acceleration
x
\M
3
b. Unstable
Figure A.1:
Configuration
Sample symmetric instability cross-sections.
- 203 -
APPENDIX B
LIST OF UPPER AIR SOUNDING STATIONS
Station
Latitude
Location
Longitude
AHN
33.95
83.32
Athens, GA
AQQ
29.73
85.03
Apalachicola, FL
AVL
35.35
82.47
Asheville, NC
AYS
31.25
82.40
Waycross, GA
CKL
32.90
87.25
Centreville, AL
DAY
39.87
84.12
Dayton, OH
FNT
42.97
83.73
Flint, MI
GSO
36.08
79.95
Greensboro, NC
HTS
38.37
82.55
Huntington, WV
IAD
38.98
77.47
Washington/Dulles, DC
LSF
32.35
85.00
Lawson Field, GA
PIT
40.53
80.23
Pittsburgh, PA
SSC
33.85
80.32
Sumter, SC
TBW
27.70
82.40
Tampa Bay, FL
99A
36.96
81.35
Wytheville, VA
-
204 -
APPENDIX C
BIPOLE AND ENVIRONMENTAL CHARACTERISTICS
This appendix
environmental
consists
a
characteristics,
graphically presented
organized
of
by
grouping
in
figures
lightning
of
bipole
the
contents
of
in
section
4.4.
hour
in
which
The first set of
the
and
are
The
data
location and polarity information.
end of the bipole was determined by detecting less
per
lightning
also
table
is
together the three bipoles occurring over the
ocean and the three land bipoles.
from
table
bipole
area.
The
therefore, given to the nearest hour.
area
derived
The beginning and
than
duration
The
is
ten
flashes
each bipole is,
of
given
is
the
area
encompassing cloud-to-ground flash densities of greater than 3 flashes
per 10,000 square kilometers, rounded to
kilometers.
The
number
of
flashes
the
is
nearest
the
10,000
total
number
cloud-to-ground flashes produced in the bipole area during the
lifetime.
The
square
of
bipole
percent positive is the number of observed flashes in
the bipole area that were positive divided
by
flashes,
The geographic centers are
rounded
to the nearest percent.
the
total
number
determined by taking the mean latitude and longitude of all the
locations of a given polarity.
of
flash
The size is the distance between these
- 205 -
BIPOLE AND ENVIRONMENTAL CHARACTERISTICS
nearest
two geographic centers rounded to the
is
orientation
reported
to
for
orientation
observed
the
all
degrees
in
"d(orientation)" is the difference
The
five degrees, and the row
nearest
the
kilometers.
ten
observed
the
between
maximum
hours, and the minimum
observed orientation.
The second set of data is determined in the same
the
the
describing
information
is
the
which
was
Following this
the upper air sounding launch time were used.
environment
meteorological
The environment was
derived from the NWS real-time analyses.
assumed
be typified by the analysis coincident with the bipole location as
Heights were interpolated from the analysis
shown in figure 4.10.
the
as
except that only the flashes occurring within thirty minutes of
first
to
manner
to
center of the bipole, and the gradient in the height field around
this point was used to
The
wind.
geostrophic
calculate
row
"d(direction)"
direction in degrees minus the minimum wind
direction
is
and
speed
the
is
direction
the
direction.
orientation
(row 10).
in
The
height
between
the
the
wind
average
The
overall
The wind speed shear was calculated by
subtracting the wind speeds at two levels and
difference
maximum
just the arithmetic mean of the four observations.
row "ave dir-orient" is the average wind direction minus
bipole
of
the
dividing
two levels.
information is described in section 4.4.
- 206 -
this
by
the
The cloud structure
BIPOLE AND ENVIRONMENTAL CHARACTERISTICS
LAND
OCEAN
Start
1/23
1/30
3/6
2/5
2/21
3/10
time
17
0
18
17
10
22
duration (min)
600
720
480
480
840
360
area (10e5km2)
3.3
5.0
5.7
1.4
3.6
2.2
2094
1419
1655
879
2244
319
fl/min
3.5
2.0
3.4
1.8
2.7
0.9
10e-5fl/min/km2
1.1
0.4
0.6
1.3
0.8
0.4
%positive
13
34
28
12
11
44
size (km)
310
180
190
80
90
150
orientatn (deg)
235
235
220
250
255
230
d(orientation)
25
30
135
40
150
35
distance East
245
90
135
240
750
240
distance South
145
345
55
10
45
20
ave speed (km/hr)
30
30
20
35
60
50
Soundg time (UTC)
0
0
0
0
12
0
301
15
244
106
246
39
%positive
12
73
36
7
5
69
size (km)
380
90
210
10
100
140
orientatn (deg)
240
220
205
240
190
240
#flashes
num of flashes
-
207 -
BIPOLE AND ENVIRONMENTAL CHARACTERISTICS
850mb SURFACE AND GEOSTROPHIC WINDS
height
(m)
direction (deg)
speed (m/s)
1525
1470
1350
1500
1410
1400
240
250
255
250
235
235
5
25
20
10
15
30
700mb SURFACE AND GEOSTROPHIC WINDS
height
(m)
direction (deg)
speed (m/s)
3080
3020
2880
3100
2970
2980
225
235
235
260
255
235
10
30
25
15
20
35
500mb SURFACE AND GEOSTROPHIC WINDS
height
(m)
direction (deg)
speed (m/s)
5660
5560
5380
5760
5580
5500
230
220
230
255
255
240
15
45
40
45
30
40
250mb SURFACE AND GEOSTROPHIC WINDS
height
(m)
direction
(deg)
10440
10420
10200
10700
10360
10420
235
220
245
250
265
260
speed (m/s)
30
65
60
25
40
45
d(direction)
15
30
25
10
30
25
235
230
240
255
255
245
0
-5
20
5
0
15
ave direction
ave dir-orient
WIND SPEED SHEAR (m/s/km)
850-700
3
3
3
3
700-500
2
6
6
11
500-250
1
4
4
4
850-250
3
4
5
2
- 208 -
BIPOLE AND ENVIRONMENTAL CHARACTERISTICS
CLOUD BAND STRUCTURE
ave orientation
d(orientation)
cloud-bip orient
220
235
245
200
225
235
35
45
15
15
25
15
0
10
-20
-25
-20
-10
Explanation of some rows below, for more information see text section 4.4.
area = area of cloud-to-ground lightning flash locations
d(orientation) = maximum hourly orientation - minimum hourly orientation
(for bipoles only sizes > 25km were used)
d(direction) = maximum direction - minimum direction
ave direction = (sum of four observed directions ) / 4
ave dir-orient = laverage wind direction - overall bipole orientationJ
in degrees
- 209 -
Acknowledgements
The author would like to thank the following people
their assistance in this effort:
for
Bob Boldi
Steve Cohn
Dole
Prof.
Randall M.
Prof.
Kerry Emanuel
Speed Geotis
Stan Heckman
Prof.
Earle R.
Williams
This work was paid for by a grant
Science Foundation for the GALE project.
A special thanks
support and advice.
to
Larry
- 210 -
Engholm
from
the
National
for
his
constant
References
Beasley, W., Positive Cloud-to-Ground Lightning Obvservations,
Geophys. Res., 90, 1985, 6131-6138.
J.
Bennetts,D. and Hoskins,B., Conditional Symmetric Instability - a
Rainbands,
Frontal
for
Explanation
Possible
Quart.J.Roy.Meteor.Soc,105,19 7 9, pp945-962
Bosart,L., The President's Day Snowstorm of 18-19 February
a Subsynoptic-Scale Event, Mon.Wea.Rev.,109,1981,pp1542-1566
1979:
Presidents'
Bosart, L., and Lin, S., A Diagnostic Analysis of the
Day Storm of February 1979, MWR, 112, 1984, pp 2148-2177.
Krehbiel, The Electrical
R.
Nakano, and P.
Brook, M., M.
J. Geophys. Res., 87,
Thunderstorms,
Winter
Hokuriku
the
of
Structure
1982, 1207-1215.
Carbone, R., A Severe Frontal Rainband.
2 58 27 9 .
Hydrodynamic Structure, JAS, 39, 1982, pp
Stormwide
I:
Part
Few,
A.
Christian, H. J., M. Weber, M. F. Stewart, and A.
Thunderstorms,
Inside
Fields
Electric
of
Measurements
Balloon
Proceedings in Atmospheric Electricity, edited by L. H. Ruhnke and J.
1 92 1 9 4 .
Latham, A. Depak, Hampton, VA, 1983, pp
Colman, B. R., Deep Convection Occurring Above a Stable Planetary
Boundary Layer, ScD Thesis, MIT, 1984.
Durran, D. R., and J. B. Klemp, On the effects of
2152 2 158 .
the Brunt-Vaissala Frequency, JAS, 39, 1982, pp
Emanuel, K., The Langrangian Parcel Dynamics
Instability, JAS, 40, 1983, pp2368-2376.
of
Moisture
Moist
Emanuel, K., On Assessing Local Conditional Symmetric
from Atmospheric Soundings, MWR, 111, 1983, pp2016-2033.
Emanuel, K. Mesoscale Meteorology Models, edited by Lilly and Gal-chen, D.
pp217-229 and pp231-245.
Symmetric
Instability
Theories, Observations and
Reidel Publishing Co., 1983,
Emanuel,K., Frontal Circulation in the Presence
Symmetric Stability, J.Atmos.Sci,42,1985,pp1062-1071
of
M., Positive Cloud-to-Ground Lightning
Fuquay, D.
Thunderstorms, J. Geophys. Res., 87, 1982, 7131-7140.
- 211 -
on
Small
in
Moist
Summer
Houze, R. Jr., Hobbs, P., Biswas, K., and Davis, W., Mesoscale
Rainbands in Extratropical Cyclones, MWR, 104, 1976, pp868-878.
Electrification
J., Review of Thunderstorm
Illingworth, A.
by L. H.
edited
Electricity,
Atmospheric
in
Processes, Proceedings
pp149-152.
1983,
VA,
Hampton,
Deepak,
A.
Ruhnke and J. Latham,
Jacobson, E. A., and E. P.
Produced by Florida Lightning, J.
Krider, Electrostatic Field Changes
Atmos. Sci., 33, 1976, 103-117.
Ice
Saunders, Charge
R.
P.
and C.
Jayaratne, E. R.,
Crystals in Laboratory Clouds, JGR, 88, 1983, 5494-5496.
on
Kessler, E., The Earth's Electrical Environment, National
81 8 9
Press, Washington, D. C., 1986, pp - .
Academy
Krehbiel, P. R., The Earth's Electrical Environment,
90 1 1 3 .
Academy Press, Washington, D. C., 1986, pp -
National
The Earth's Electrical Environment,
P.,
Krider, E.
30 40
Academy Press, Washington, D. C., 1986, pp - .
National
Inside
Kuettner, J., The Electrical and Meteorological Conditions
Thunderclouds, J. Met., 7, 1950, 322-332.
Kuettner, J., The Development and Masking
Thunderclouds, J. Met., 13, 1956, 456-470.
of
Charges
in
Magono, C., T. Endoh, and T. Taniguchi, Charge Distribution in
Active Winter Clouds, Proceedings in Atmospheric Electricity, edited by
Deepak, Hampton, VA, 1983,
Latham, A.
Ruhnke and J.
H.
L.
pp242-245.
Winn, Measurement of Charged
P.
and W.
C.,
Marshall, T.
Lower Positive Charge
Thunderstorm:
Mexico
Precipitation in a New
7141-7152.
1982,
87,
Res.,
Centers, J. Geophys.
Bosart, An
Orville, R. E., R. W. Henderson, and L. F.
1029-1037.
1983,
64,
Bul.,
AMS
Network,
Coast Lightning Detection
East
Orville, R. E., R. A. Weisman, R. B. Pyle, R. W. Henderson,
Orville, Jr., Cloud-to-Ground Lightning Flash Characteristics
E.
R.
92, 1987,
Res.,
Geophys.
from June 1984 through May 1985, J.
5640-5644.
The Earth's Electrical Environment,
Orville, R. E.,
D. C., 1986, pp23-29.
Washington,
Academy Press,
National
Bosart, Bipole
F.
Orville, R. E., R. W. Henderson, and L.
Storm Systems,
Mesoscale
in
Locations
Patterns Revealed by Lightning
Geophys. Res. Let., 15, 1988, 129-132.
- 212 -
Orville, R. E. Jr., An Analytical Solution to Obtain the Optimum
Source Location Using Multiple Direction Finders on a Spherical
Surface, J. Geophys. Res., 92, 1987, 10877-10886.
T., Electrostatic Field-changes Due
Pierce, E.
Discharges, Q. J. Roy. Met. Soc., 81, 1955, 211-228.
Roy.
Pierce, E. T., The Development of Lightning
Met. Soc., 81, 1955, 229-240.
to
Discharges,
Lightning
Q.
J.
Reap, R. M., Evaluation of Cloud-to-Ground Lightning Data from
the Western United States for the 1983-84 Summer Season, J of Climate
785 799 .
and Appl Met, 25, 1986, pp
Compendium of Thunderstorm Electricity, New
Reynolds, S. E.,
and Technology, Research Development
Mining
of
Institute
Mexico
3-99-07-022, Signal Corps Project
Project
Army
the
of
Division, Dept.
17B, 1954.
The Earth's Electrical Environment,
Roble, R. and Tzur, I.
National Academy Press, Washington, D. C., 1986, chapter 15.
T.
MacGorman, and R.
R.
Taylor, D.
Rust, W. D., W. L.
in
Thunderstorms
Severe
of
Arnold, Research on Electrical Properties
1286-1293.
the Great Plains, AMS Bull., 62, 1981,
Rust, W. D., D. R. Macgorman, and R.
Cloud-to-Ground Lightning Flashes in Severe
Let., 8, 1981, 791-794.
Arnold, Positive
T.
Storms, Geophys. Res.
Rust, W. D., The Earth's Electrical Environment, National Academy
Press, Washington, D. C., 1986, pp41-45.
the
in
Structure
Bosart, Mesoscale
and L.
Sanders, F.
I:
Part
1983.
February
11-12
of
Snowstorm
Magalopolitan
1985,
42,
JAS,
Frontogenetical Forcing and Symmetric Instability,
ppl050-1061.
the
in
Structure
Bosart, Mesoscale
and L.
Sanders, F.
Doppler
Magalopolitan Snowstorm of 11-12 February 1983. Part II:
13 98 1407 .
Radar Study of the New England Snowband, JAS, 42, 1985, pp
J.
and F.
C.,
Simpson, G.
Electricity in Thunderclouds, Proc.
309-352.
Scrase, The
Soc.
Roy.
of
distribution
161, 1937,
A.,
Shands, A. L., Thunderstorm Probabilities in the U.
of the AMS, 29, 1948, 270-274.
- 213 -
S., Bulletin
Robinson, The Distribution of
D.
and G.
C.,
Simpson, G.
Roy. Soc. A., 177, 1941,
Electricity in Thunderclouds II, Proc.
281-329.
Takahashi, T., Riming Electrification as a Charge Generation
Mechanism in Thunderstorms, J. Atmos. Sci., 35, 1978, 1536-1548.
J.
Takahashi, T., Thunderstorm Electrification-- A
Atmos. Sci., 41, 1984, 2541-2558.
Numerical
Study,
L.
Ishikawa,
Nakano, H.
Israelsson, M.
Takeuti, T., S.
Lundquist, and E. Astrom, On Thunderstorms Producing Positive Ground
H.
Flashes, Proceedings in Atmospheric Electricity, edited by L.
24 6 248 .
Ruhnke and J. Latham, A. Deepak, Hampton, VA, 1983, pp
Takeuti, T., Z. Kawasaki, K. Funaki, N. Kitagawa, and J. Huse,
On the Thundercloud Producing the Positive Ground Flashes, J. Met.
Soc. Japan, 63, 1985, 354-357.
Thunderstorm Morphology and Dynamics, edited by E.
of OK Press, Norman and London, 1983.
Kessler, Univ.
Uman, M., The Lightning Discharge, International
Series, 39, Academic Press Inc., Orlando, FL, 1987.
Geophysics
Passarelli, Band
Wolfsberg, D. G., K. A. Emanuel, and R. E.
Rev., 1986, 114,
Wx.
Mon.
Storm,
Formation in a New England Winter
ppl552-1569.
-
214 -