Science Stories

Discover More about XSEDE-enabled science, programs, trainings, and more.

Key Points
See how XSEDE supports science
« Back

XSEDE researchers visualize massive Joplin, Missouri tornado

XSEDE researchers visualize massive Joplin, Missouri tornado

Using XSEDE services and resources and the power of NCSA's Data Analysis and Visualization group, a pair of Illinois researchers were able to accurately simulate an EF-5 tornado in Joplin, MO from 2011.

What if we could peer into the formation of a tornado in hopes of better understanding it? Better yet, what if we could look into one of the deadliest and most unpredictable storm cells in recent memory?

Thanks to XSEDE, visualization expert David Bock (National Center for Supercomputing Applications) and atmospheric science researcher Brian Jewett (University of Illinois) are doing just that, by working together to visualize a unique tornado that wreaked havoc on Joplin, Missouri in 2011.

The tornado, which lasted for 38 minutes on May 11, 2011, claimed 158 lives and resulted in over 16,000 insurance claims totalling more than $2 billion dollars, more than any other single event in the history of Missouri.

But how can we gain present-day insights from a storm that happened over seven years ago?

With the power of data visualization, data becomes explicitly visible, leading to new insights into atmospheric behaviors that weren't previously obtainable. So long as relevant data exists (radar imagery, temperature and humidity data, etc.), it can be molded into something visible.

Thanks to an allocation from the Extreme Science and Engineering Discovery Environment (XSEDE), which connects researchers with relevant resources and expertise, enabling accelerated discovery from scientific experts, Brian Jewett, a professor of Atmospheric Science at the University of Illinois, has been able to collaborate with NCSA visualization expert David Bock to help to reveal insights from merging storms, an ongoing mysterious atmospheric phenomenon.

"XSEDE makes it so easy. We can get to work on something immediately," said Bock. "We have at our disposal not just visualization capabilities, but computational help. It created a conduit for us to work together."

With the help of XSEDE's Extended Collaborative Support Services (ECSS) team, which provides the framework for researchers to find relevant collaborators, Jewett was able to connect with a dedicated visualization expert (Bock) to make his data come to life. Due to ECSS's robust collaboration structure, this project was relatively seamless.

"Formally, we're working through XSEDE's ECSS team," said Jewett. "I've worked with XSEDE's ECSS team before on the computational side, but I've never gotten to work on the visualization side. ECSS provides a formal procedure where I can put in a request to work with someone like Dave Bock, right down the street here at NCSA, and it can happen quickly."

This collaboration lets Bock and Jewett, two subject-matter experts in their respective fields, combine their skill sets to produce groundbreaking research - in this case, surrounding the formation of tornadoes.

"There are still questions about how a tornado starts, and numerical simulation is critical to addressing a lot of those questions," said Jewett. "Numerical models give you two things. First, they give you an incredibly detailed picture in time and space of what's going on, so as often as one chooses to save the data, you can establish continuity, a vital aspect of a weather simulation."

These models also give researchers like Jewett an extra tool in analyzing storms – they allow for variables to be changed to observe storms under different conditions.

"One can also go in and tweak things in numerical models," Jewett said. "For example, in the past if you wanted to study a warm case and a cold case, you'd have to dig through weather maps and find cases with identical conditions. With the model, we can control the amount of rain or sunshine or even the time of year, and see how a given storm reacts."

Now that numerical simulations can be run, the help of a visualization expert is needed to transform the data into something visible. That's where NCSA's David Bock comes in. With the help of Jewett, he takes massive amounts of data and turns them into something we can see, and in turn, learn about, even when the concepts are novel.

"In the Joplin storm, there were actually two storms that were merging. Before this, we had visualized one big storm, but we never had to consider different regions of the same data model," said Bock "How do you color these distinct storms differently to tell the difference between them? What happens when these storms merge? We're talking about how to do that, and it's fascinating to extend my software to do these new, complex things."

Both on the visualization and modeling ends, however, there is substantial work yet to be done.

"As you saw with the Joplin storm, looking at it from above, it's beautiful," said Bock. "But we don't have enough resolution to see the space where the tornado forms. We need more frames and we need more data, but to me, that's the exciting part. We figure out what combination of variables we want to look at, and then we can run another simulation to look closer."

While this may not provide us all of the answers, visualizing the previously-un-visualized leads to important scientific insights – ones that could help us better predict and prepare for deadly tornadic events like those in Joplin seven years ago.

"Decades ago, scientists at NCAR (National Center for Atmospheric Research) developed simulations that showed splitting storms. After their work was released, people starting looking at splitting storms much closer," said Jewett. "In turn, simulations can inform the community with things to look out for, which has led to strides made in forecasting and warning. It hasn't painted the whole picture, but these sort of simulations can inform the general public's understanding of atmospheric sciences."