A Terabyte of a Twister
Donna Cox is a professor and the director of visualization and
experimental technologies at the National Center for
Supercomputing Applications at the University of Illinois at
Urbana-Champaign, where she oversees the animation, coloring and
design of various simulation projects. Her collaborators on this
project were Robert Patterson, Stuart Levy, Alex Betts and
Matthew Hall (visualization); and Robert Wilhelmson, Matthew
Gilmore, Louis Wicker, Glen Romine, Lee Cronce and Mark Straka (simulation).
F. F. The triangular shapes on the surface seem to indicate
a directionality. Am I reading that correctly?
D.C. A new "glyph" was developed for this project. The
triangular shapes on the surface of the ground are three-dimensional
cones—a visual representation adapted from conventional
vectors made with black and white line plots. The cones indicate
direction of the air at the ground. In addition, these glyphs
provide two more pieces of information: They use color to indicate
temperature, and the strength of the wind is indicated by their
tilt. This is analogous to seeing wheat grass blow in the wind: You
can see how hard the wind is blowing by how flat the grass is pushed
F. F. Can you explain the color choices? For example, I
notice that orange appears in the tubes and also in the
triangular surface forms. Was that intended to show a
correlation or was it purely an aesthetic choice?
D.C. "Cool" colors in the cyan-blue-purple range and
"warm" colors, red-yellow-brown, are used for indicating
temperature. In general, cool temperatures are indicated by bluish
colors and warm temperatures by reddish colors. Aesthetic decisions
control the levels, saturations, shades and intensities. The cool
colors of the cones and streamtubes are both representing cool
temperatures in the air.
The spherical particles are colored according to pressure. The
redder they are, the higher the pressure. The streamtubes are
colored by temperature, with a sharp divide between cool and warm.
Finally, the color of the cones also has a divide between cool and
warm air, but the cones have a slightly different saturation level
to stand out from the colors used on the streamtubes.
F. F. Did you test various color palettes before you came up
with this final version?
D.C. We tested many versions of the color-transfer functions that
are the "color maps" used here. We tested color intensity,
value, saturation, luminosity, translucency and color temperature.
These color maps were determined to help show a difference between
the surface and the air temperature. They were also developed while
working with the scientists.
F. F. I wonder if you see a benefit to reading this
simulation frame by frame, along with the dynamic version.
D.C. The ability to study a frame-by-frame image sequence as still
frames in addition to seeing the time-evolving animation is the best
combination for study and understanding of data. While the animation
provides a qualitative, dynamic motion study, I think that being
able to study a time sequence frame by frame provides a different
level of information and enhances the understanding of the overall interaction.
F. F. What do you find is the biggest challenge in
collaborating with scientists while making these representations?
D.C. The biggest challenge in collaborating with scientists is
overcoming differences in language and vocabulary. Visualization
experts have definitions for "features" that may or may
not always fit with a scientist's definition. Finding common ground
on how to explain concepts using language and pictures is the