On the Radar: CMU Professor creates state-of-the-art tornado simulation computer software


2_tornado

Since coming to Central Michigan University in 2003,  Leigh Orf has dedicated his time and resources toward Mesoscale Meteorology – or in simpler terms, thunderstorms. 

Since age seven, when lightning struck his childhood home, storms have fascinated Orf. Wanting to learn more about them, and their byproduct, tornadoes, Orf has dedicated the 10 years of his life towards creating a computer model which would be able to simulate the warning signs of a thunderstorm. The model would allow him to see if the storm was capable of producing an EF5 level tornado – one of the most deadly and destructive.

Taking a break from his work, Orf had time to answer a couple of questions about his new technology and how he plans to use it for the future.

Ten years, that’s quite a long time to be working on a simulation. What made the process take so long?

ORF: The (original) model was written by a scientist at another institution called the CM1 model. I have taken that model and modified it so that I can do all this analysis.

In order to do a simulation like this, this model can produce about 100 terabytes of data. When you’re dealing with that much data, you can easily be crippled by it. I first started thinking about this project about a decade ago. I started to think about how I was going to manage the amount of data future super-computers would produce in my research. And that was where I started.

So it was a lack of technology that hindered you?

One of the reasons why this wasn’t done earlier was because we just didn’t have the computing technology. Recently, a computer was built – which didn’t exist before – called Blue Waters, and it’s from the University of Illinois. It’s a computer that’s powerful enough such that this work was possible.

Why choose to work with meteorology, and more specifically, tornadoes?

Tornadoes are a real unsolved problem, and scientists like to find really interesting and timely and important areas of research that can have an effect on society. If any of this research can save any lives – if for instance, down the road, the weather service might take some of the results of my research and be able to identify features of (storms) using their technology on a storm that just began, then they might be able to say ‘Wow, this storm has all the Hallmarks of being an on track, EF5 tornado’.  They can (then) warn the public with a specific warning so that people take it seriously.

But that’s a long ways away because we don’t even know what those features are yet, or even if they can be detected on radar. I’m starting to see them in the simulation. You need to understand how something works before you can predict its behavior.

Can you how the computer model works?

I’ve conducted a simulation of a thunderstorm that produces a devastating tornado. The most uncommon type of thunderstorm produces the most devastating tornado, but these (thunderstorms) don’t happen all that often – the types of tornadoes that we’re trying to understand are very rare. They produce a disproportionate amount of devastation, damage and loss of life.

Up until this simulation, nobody had (created a simulation) like this. Nobody had a thunderstorm form that then produced the tornado. Right now, I’m trying to understand the storm from when the tornado first descends to the ground, it’s on the ground for two hours, and then it dissipates.

So you said you’re looking for an EF5 tornado, for people who aren’t as versed in meteorology as yourself, what does that mean?

Tornadoes are ranked on a scale from zero to five: zero being the weakest, five being the strongest and (that scale) is called the Enhanced Fujita Scale. I’m looking at the EF5 tornadoes that are on the ground for 60 to 70 miles, a really long track.

What do you hope to accomplish with this technology?

The ultimate goal of this research is to issue better (tornado) warnings and to communicate to the public more effectively when there is one of these devastating storms that’s knocking on their door.

What’s wrong with the way tornado warnings are currently issued?

Right now, we issue tornado warnings based upon somewhat crude indicators on weather radar and often times a tornado doesn’t even form, so we call that a false alarm. False alarms make the public disillusioned with the warnings, so they don’t respond and (it becomes) ‘The Boy Who Cried Wolf’ syndrome, so to speak.

So when you have that problem and there’s a real storm, a real wolf, then people might not respond.

How does the rest of the weather community regard your work? Are they excited about it?

The metrological community is very interested in this. It has made a big impression, and people are very interested in this simulation. I’m at the point where I’ve presented at conferences and I’m working on papers.

Although it’s still kind of in its infancy now, where do you hope to see this simulation in the future?

I think we will be able to, and this is the distant future, we will be able to run simulations like this in real time. For instance, meteorologists are very good at identifying the types of environmental conditions that are conducive to these types of storms – they’re very good at saying ‘Okay, this region of Oklahoma, we’re saying there’s going to be a high risk of storms occurring in this location.’ That’s where we’re at today.

Tomorrow, I’d like to be in a place where we can run simulations and see if the storms that are simulated in that environment are severe storms. And then if that’s true, we will be able to tell people within a mile or so that they’re going to be in the track of the tornado when it’s traveling, even if it hasn’t formed from the thunderstorm yet.

Share: 

About Jordyn Hermani

Troy senior Jordyn Hermani, Editor-in-Chief of Central Michigan Life, is a double major ...

View Posts by Jordyn Hermani →