| 
June 2004
Identifying storms that produce
tornadoes
One of the
perennial challenges in severe-weather forecasting
is picking out which thunderstorms might produce
a tornado. Most significant tornadoes emerge from
mesocyclones, areas of rotating wind a few miles
wide within a supercell thunderstorm. However,
only a small fraction of mesocyclones results in
twisters. Even some of the largest, most intense
mesocyclones—such as the one that produced
a record hailstone in Nebraska last June—may
be unable to spin up a tornado.
Huaqing Cai. (Photo by Carlye Calvin.) |
An ASP postdoctoral researcher, Huaqing Cai, is working
on a technique that could help forecasters identify
the cells most likely to produce tornadoes amid a
batch of severe thunderstorms. Huaqing, now in the
second year of his fellowship, has been collaborating
with Wen-Chau Lee (ATD) and Roger Wakimoto (University
of California,
Los Angeles).
To get a new perspective on the problem, Huaqing
turned to fractals. Introduced by Benoit Mandelbrot
in the 1970s, fractal geometry is a mathematical
representation of patterns in nature. For example,
while a mile-long ruler would give one measurement
of a coastline, it would miss smaller curves. Fractal
theory emphasizes how rulers of different lengths
give a variety of measurements that sometimes add
up to a more complete picture.
Huaqing applied this technique to radar data of five
mesocyclones, three of which spawned tornadoes. Much
like using different-sized rulers, he analyzed the
data with a variety of horizontal grid scales. The
resolution varied from 0.3 to 9.6 kilometers (0.19
to 6 miles).
For each storm, he examined how the maximum vorticity
(a property of air flow rotating around an axis)
changed with different scales.
As fractal theory would predict, vorticity and scale
are tightly correlated. Maximum vorticity declines
with larger grid spacing because the larger grids
fail to capture intense, fine-scale vortices (just
as a larger ruler might fail to capture a sharp indentation
in a coastline).
But Huaqing found that the
slope of that relationship varies in an interesting
way: it’s apparently
steeper for tornadic than for nontornadic mesocyclones.
The more sharply the mesocyclone’s maximum
vorticity declines as the grid spacing increases
(which is an indication of more energy fueling small-scale,
rather than large-scale, vortices), the greater the
likelihood of a tornado.
This chart shows
an analysis of maximum vorticity (vertical
axis) versus horizontal grid spacing (horizontal
axis) on a logarithmic scale for five supercell
thunderstorms with mesocyclones. The steepest
gradients are for the Kellerville and Garden
City mesocyclones,
both of which spun up tornadoes (rated F3 and
F1, respectively). The gradients for the San
Angelo (very weak tornadic, not rated), Hays
(nontornadic), and Superior (nontornadic) are
more gradual. (Courtesy Huaqing Cai.)
If this relationship is confirmed by more data, it
could eventually lead to on-the-fly radar analysis
to distinguish the most tornado-prone storms, if
the storms are not too far away from radar.
Huaqing says that more work
needs to be done to understand the physical basis
for the relationship he’s
discovered. “This isn’t a tornadogenesis
theory. It doesn’t tell you why a mesocyclone
produces a tornado; it just tells you which one might.” •Bob
Henson
Also in this issue...
A site to behold: www.ucar.edu
Random
Profile: Raisa Leifer
A
fast way to measure PAN
Commuter
of the Year
Newsroom | Events | Publications | Visuals | Help
Center |
|
