Introduction - Severe Convective Storms and Tornadoes: Observations and Dynamics - page 23

of 1m, it is not yet possible to include the entire parent storm at that fine a reso-
lution. Dave and Steve Lewellen and collaborators at West Virginia University
have been pioneers in this arena. Rich Rotunno at NCAR, building on earlier
work by Roger Smith, Lance Leslie, Bruce Morton, A. Szillinsky, Lou Berkofsky,
F. Wippermann, Francis Harlow, and Leland Stein, simulated tornado-like vor-
tices in the late 1970s and did some seminal nested grid simulations of tornado-
like vortices in supercells in the mid-1980s. The way subgrid-scale parameteriza-
tions are handled is very important.
More recently, it has been possible to use radar observations to fit numerical
models to the real atmosphere state and make predictions, a vision first put
forward by Doug Lilly at OU in the 1980s and implemented with work by Kelvin
Droegemeier, Ming Xue, and his colleagues at OU/CAPS. Since model equations
are highly nonlinear, the forecasts produced by numerically integrating the equa-
tions in time are highly sensitive to both the initial conditions and to the type of
model physics used to represent microphysical and turbulent processes. Ensemble
forecasts are being produced at CAPS that make use of numerical integrations
under many different slightly perturbed initial conditions and subject to different
parameterizations of microphysics and subgrid-scale turbulence. The mean of all
these integrations is thought to be a good guess at what will happen, while the
spread from all the different integrations is thought to be a good measure of the
forecast uncertainty.
1.3 METHODS TO BE EMPLOYED
In this text I will emphasize a number of aspects of severe convective storms and
tornadoes: recent (beginning in the mid-20th century) historical context, basic
descriptions of convective phenomena including photographs and radar imagery
(with a bias toward the U. S., where data are most easily available to the author),
Doppler radar wind analyses, polarimetric data, in situ measurements, idealized
analytical modeling, similarity theory, linear stability analyses, and three-
dimensional, non-hydrostatic cloud modeling results will all be considered. In
particular, analysis of real data and model data, employing Lagrangian parcel
analysis of vorticity and circulation, and analyses of horizontal vorticity to
represent vertical circulations in convective storms will be discussed.
Historical context is important because it gives credit to the pioneers who led
us to where we are today and to understand why certain avenues of research have
been pursued. Again, the reader is reminded of the non-exhaustive group of old
and recent photographs of some of the major figures and instrumentation involved
in research in severe convective storms and tornadoes in Figures 1.1 -1.10. In
recent years we have been able to obtain unprecedented high spatial and temporal
resolution observations and high spatial resolution numerical simulations of torna-
does and convective storms. Understanding the basic dynamics of the observed
phenomena, however, requires our using multiple approaches from our ''bag of
tricks'', which includes idealized modeling (cutting to the basic mechanisms, even