WMOD00311

• Title: Harold D. Orville Symposium - Forty Years of Modeling Clouds and Weather Modification
• Date: 4/26/2003
• Weather Modification - Doc Type: Report

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I

during a non-coalescing collision are functions of
temperature and liquid water content. However,
there are differences between the two sets of
laboratory results regarding the signs and
magnitudes for the same temperatures and liquid
water contents. While the underlying theory of
how the noninductive mechanism operates still
needs to be worked out, we used the 2D SEM to
compare the results of a thundercloud simulation
using the two sets of laboratory work. Helsdon et
a/. (2001) used the 19 July CCOPE cloud for the
comparison and found that the laboratory results
of Takahashi, as implemented by Randell et a/.
(1994), produced electrical structures that agreed
better with observations than those of Saunders et
a/. (1991). The results suggested that the
laboratory work of Takahashi might be a better
representation of the non inductive charging
process. However, more work needs to be done
with respect to refinement of laboratory results,
and model studies using the 3D SEM are required
before any final conclusion can be reached.
Saunders (1993), in his review, suggested
that model studies were still required to determine
whether the convective mechanism could produce
organized charge separation on timescales
appropriate to the electrification of thunderclouds.
Masuelli et a/. (1997) undertook such a study,
hoping to improve on the work of Chiu and Klett
(1976). Using a 2D kinematic model with
precipitation driven by dynamic (field-of-motion)
output from a 3D model, they simulated a 10-min
period of cloud growth and electrification by ion
attachment. In contrast with the results of Ruhnke
(1970) and Chiu and Klett (1976), they found the
development of an opposite-polarity charge
structure and orders-of-magnitude greater charges
and electric field strengths, although still not to
thunderstorm magnitudes. They suggested that
their results needed to be checked using a more
comprehensive, coupled model. Since these
results were significantly different from earlier,
simple simulations, we undertook to do a
comprehensive evaluation of the convective
hypothesis.
Helsdon et a/. (2002) used the 3D SEM with
an upgrade of the lightning scheme of Helsdon et
al. (1992) to the 3D geometry to do a
comprehensive comparative study of the
convective mechanism vs. the noninductive
mechanism for two storm situations - a weak
storm (the 19 July CCOPE storm) and a severe
hailstorm (the 1 July 1993 storm from the North
Dakota Tracer Experiment (NDTE}). These two
storms represent opposite ends of the dynamic
storm spectrum. Each storm was simulated using

only the convective (ion attachment) process for
electrification and, again, with the non inductive
chare separation mechanism. As a diagnostic
tool, we introduced a calculation of the conduction,
convection, and point discharge components of
the Maxwell current to help in analyzing the
results. Results showed that the convective
mechanism was capable of producing only weak
and generally disorganized electrification of both
storms, although the electrification of the NDTE
severe storm case was stronger and showed
characteristics of being a current generator for a
brief period. In both cases using the noninductive
scheme, strong electrification (and lightning in the
NDTE case) were produced, consistent with the
observed character of both these storms. Our
results were consistent with those of Ruhnke
(1970) and Chiu and Klett (1976) and did not
support the charges and fields obtained by
Masuelli et al. (1997). We concluded that the
convective hypothesis, as envisioned by
Vonnegut, was not capable of producing strong
electrification in any type of storm situation.
However, ion capture processes are important in
the formation of screening layers in simulations
and should be included in the model physics.
The Maxwell current analysis showed that
both of the storms, with non inductive charging,
acted as current generators. Figure 7 shows the
a} convection and b} conduction current vectors in
a south-north slice through the model domain at X
= 20 km for the NDTE simulation at 20 min, 2 min
after the onset of lightning. Note the scale vector
length represents a current of 20 nAlm2 for the
convection current, but only 2 nAlm2 for the
conduction current. The convection current - the
current carried by charged particles moving in the
flow field (Fig. 7a) - shows a strong generator
current (upward current flow as opposed to the
downwardly directed fair weather current)
throughout most of the cloud volume. There is
some horizontal turn to the current vectors in the
upper right portion indicating horizontal flow of
charged hydrometeors into the anvil. The
conduction current - the current carried by ions
moving in response to the electric field (Fig. 7b) -
also shows a strong current into the stratosphere
as well as from the surface up to the base of the
cloud, completing the connection between the
surface, the cloud, and the upper atmosphere as
part of the global circuit. The conduction current
also shows a solenoidal character for currents that
are off-axis from the core of the cloud. This is
consistent with theory. The calculation of Maxwell
current components has proven to be a useful tool
in analyzing storm electrical evolution and will be

41