Numerically Simulated Electrification and Lightning of the 29 June 2000 STEPS
Supercell Storm
KRISTIN M. KUHLMAN
Cooperative Institute for Mesoscale Meteorological Studies, and School of Meteorology, University of Oklahoma, Norman, Oklahoma
CONRAD L. ZIEGLER
National Severe Storms Laboratory, Norman, Oklahoma
EDWARD R. MANSELL
Cooperative Institute for Mesoscale Meteorological Studies, and National Severe Storms Laboratory, Norman, Oklahoma
DONALD R. MACGORMAN
National Severe Storms Laboratory, and Cooperative Institute for Mesoscale Meteorological Studies, Norman, Oklahoma
JERRY M. STRAKA
School of Meteorology, University of Oklahoma, Norman, Oklahoma
(Manuscript received 30 June 2005, in final form 6 January 2006)
ABSTRACT
A three-dimensional dynamic cloud model incorporating airflow dynamics, microphysics, and thunder-
storm electrification mechanisms is used to simulate the first3hofthe29June 2000 supercell from the
Severe Thunderstorm Electrification and Precipitation Study (STEPS). The 29 June storm produced large
flash rates, predominately positive cloud-to-ground lightning, large hail, and an F1 tornado. Four different
simulations of the storm are made, each one using a different noninductive (NI) charging parameterization.
The charge structure, and thus lightning polarity, of the simulated storm is sensitive to the treatment of
cloud water dependence in the different NI charging schemes. The results from the simulations are com-
pared with observations from STEPS, including balloon-borne electric field meter soundings and flash
locations from the Lightning Mapping Array. For two of the parameterizations, the observed “inverted”
tripolar charge structure is well approximated by the model. The polarity of the ground flashes is opposite
that of the lowest charge region of the inverted tripole in both the observed storm and the simulations. Total
flash rate is well correlated with graupel volume, updraft volume, and updraft mass flux. However, there is
little correlation between total flash rate and maximum updraft speed. Based on the correlations found in
both the observed and simulated storm, the total flash rate appears to be most representative of overall
storm intensity.
1. Introduction
The Severe Thunderstorm Electrification and Pre-
cipitation Study (STEPS) took place during 2000 to
study severe storms in the high plains of the United
States. One of the main goals of the project was to
achieve a better understanding of the interactions
among storm kinematics, microphysics, and electrifica-
tion, especially in storms that produce predominately
positive cloud-to-ground (CG) lightning (Lang et al.
2004). Numerous storms—including supercells, short-
lived multicell storms, and large mesoscale convective
complexes—were observed and documented during
STEPS.
An in-depth study of storm processes requires a com-
bination of observations and numerical simulations.
Corresponding author address: Kristin M. Kuhlman, School of
Meteorology, University of Oklahoma, Norman, OK 73019.
E-mail: kkuhlman@ou.edu
2734 MONTHLY WEATHER REVIEW VOLUME 134
? 2006 American Meteorological Society
MWR3217

STEPS provided a comprehensive observational
dataset for detailed comparison with numerical simula-
tions of storm evolution. The present study focuses on
numerical simulations of the 29 June STEPS supercell
that produced an F1 tornado, large hail, and predomi-
nately positive ground flashes. The objectives are to
evaluate the simulated charge structure, lightning flash
rate, and polarity in comparison with the observed
storm and to determine the sensitivity of the modeled
storm to different electrification parameterizations.
The origins of positive (
) CG flashes and the relation-
ships between the modeled total flash rate and storm
characteristics are of particular interest.
a. Studies of inverted-polarity severe storms and
positive CG flashes
The traditional conceptual model of the gross elec-
trical structure of thunderstorms is that it can be de-
scribed as either dipolar or tripolar (Figs. 1a and 1b),
with the main charges being a middle-level negative
charge and an upper-level positive charge. A small, spo-
radic positive charge was sometimes found beneath the
negative charge to form the tripolar structure (e.g., Wil-
liams et al. 1989), and later it was found that a negative
screening-layer charge is often near the upper cloud
boundary. Several studies have suggested that dipole or
tripole models are not sufficient to describe how charge
is distributed in all thunderstorms. Rust and Marshall
(1996) argued that the current tripole models are too
simplistic to apply to all mature thunderstorms and me-
soscale convective systems. A more complex charge
structure consisting of four main charge regions near
the updraft and six charge regions outside the updraft
in the convective precipitation region was suggested by
Stolzenburg et al. (1998). Other researchers have sug-
gested that the tripole is adequate to describe the basic
thunderstorm charge structure, even arguing that aban-
donment of the tripole model would be “ill-advised”
(Williams 2001).
The existence of an inverted-polarity storm tripole
(Fig. 1d) was first noted by Marshall et al. (1995a)
based on an electric field sounding in a strong storm
near Dalhart, Texas. Marshall et al. suggested that

CG flashes in the Dalhart storm immediately follow-
ing the sounding may have been caused by the inverted
charge structure of the storm. Additional storms having
inverted-polarity electrical structure were observed in
the STEPS field program and were studied by Rust and
MacGorman (2002), Rust et al. (2005), MacGorman et
al. (2005), and Wiens et al. (2005). These last two stud-
ies include observations of the 29 June 2000 supercell
storm being considered in the present paper.
Various studies have examined correlations between

CG flashes and severe weather. The first documenta-
tion of severe storms that commonly produce
CG
flashes was provided by Rust et al. (1981), who con-
cluded that the occurrence of
CG flashes may indi-
cate a storm is severe. In a statistical study of Oklahoma
storms, Reap and MacGorman (1989) found that
storms that produced a larger number of
CG flashes
had a higher probability of producing severe weather.
Subsequently, Seimon (1993) noted
CG flashes pre-
ceding an F5 tornado, while MacGorman and Burgess
(1994) showed that damaging tornadoes occurred after
peaks in
CG flash rates in several storms. However,

CG flashes may prove ineffective as a severe weather
indicator unless correlations can be shown to be reli-
able (Branick and Doswell 1992; Perez et al. 1997).
Knapp (1994) and Carey et al. (2003b) have shown that

CG flashes are usually associated with severe storms
in the central United States, not in the eastern states or
coastal regions. Further study is needed to determine
the relation of
CG flashes (if any) to the severity of
the parent storm.
Several studies [including the above studies of
CG
flashes, as well as MacGorman et al. (2001), Gilmore
and Wicker (2002), MacGorman et al. (2005), and Wil-
liams (2001)] have examined the conditions needed to
produce
CG flashes. Charge configurations hypoth-
esized to be favorable for producing
CG flashes in-
clude tilting of the charge layers due to wind shear,
precipitation unshielding of positive charge (i.e., fallout
of a lower-level negative charge region revealing the
upper positive charge), feedback formation of a lower
FIG. 1. Conceptual model of charge structure of a thunderstorm.
(a) Normal dipole model, containing upper positive and lower
negative charge centers. (b) Normal tripole model, containing up-
per positive, main negative, and smaller lower positive charge
centers. (c) Inverted dipole model, with lower positive charge and
upper negative charge centers. (d) Inverted tripole model, main
positive with upper and lower negative charge centers.
OCTOBER 2006 KUHLMAN ET AL. 2735