Big PictureBy Brian Voerding
When Mingrui Zhang heard news of the BP oil spill, he immediately contacted his longtime friend and collaborator, professor Chuanmin Hu.
"My instinct told me he must have access to those satellite images. And he did."
That instinct begins the story of how one Winona State University professor is building technology that could fundamentally change how oil spills are identified, tracked, and managed — all while living about as far from any ocean as he can get.
Mingrui Zhang loves looking at both computer screens and the world's largest bodies of water: he has degrees in both computer science and oceanography. Though it isn't fair to say he's a strident environmentalist. Nor to say he's married to his computer. He's just a guy who seeks problems to solve with his particularly unique set of tools.
"I'm really intrigued by the challenges," he says. "We in computer science want to know how much we can help out to preserve the environments of coastal areas."
Zhang has capitalized on a groundbreaking shift in the intersection of environmentalism and technology; namely, the ability for the first time in human history to study nature from a perch much farther removed from, say, the view of a river from a bluff. He specializes in the area of remote sensing, which in his case means stitching together satellite photos to study large ecosystems and environments. That currently means photos of the oil spill in the Gulf of Mexico.
He collaborates most frequently with Hu, a professor at the University of South Florida, and has for over a decade. Their first work was in so-called red tides, large algae blooms in oceans that can wreak havoc on marine and coastal wildlife and ecosystems. It was during that research they created the technology they're now honing to identify oil spills.
Here's how it works:
Currently, researchers painstakingly study satellite photos of the ocean, manually outlining hot spots like red tides or oil spills. Satellite photos aren't received in real-time and they're only helpful when clouds aren't in the way. That's a problem, because it means the work is largely reactive, limiting the ability to predict and respond. Zhang's program uses artificial intelligence to scan photos and automatically outline areas, providing a quick turnaround that allows opportunities to track movements and coordinate responses.
Their program, when first applied to red tides, was so successful that it was adopted by agencies around the world, including the European Space Agency. Several current graduate students at South Florida studying red tides continue to use the program, and the U.S. Naval Research Laboratory later contracted Zhang and his collaborators to develop a program to study water quality along the coasts.
In 2008, when Zhang and Hu used the program to search for red tides in the Gulf of Mexico, they discovered several unidentified spots on the satellite photos. When they asked around, they learned the spots represented naturally occurring oil slicks.
"The discovery was purely by accident," Hu says. "I was looking for red tides but found oil."
He and Zhang began revising their program to identify oil on the ocean's surface, but eventually set it aside without completing it.
"At the time we just looked at some images and didn't think it could be a very big project," Zhang says. "But then all of a sudden (the oil spill) happened, and we started to consider it seriously."
Zhang took on the new project, using satellite photos that Hu provides on request. He works with Alex Corcoran, a Winona State undergraduate student, and Michelle Morgan, a graduate student in Jamaica who once studied with Zhang. With red tide, they searched for color. Here, they searched for reflectance: sunlight reflects surface oil much differently than water.
Zhang hopes to have the program ready by year's end. When completed, it will track surface oil spills in near real-time and be at least 80-percent accurate, meaning that while it will require human eyes to verify results, it's plenty reliable to turn a reactive technology into a proactive one.
Once it's finished, they'll turn it over to several major agencies that Zhang and Hu have worked with in the past, including the National Oceanographic and Atmospheric Administration.
Zhang expects that the program, much like the once that identifies red tides, will be refined and improved by other researchers once released to the global community of ocean researchers. To him, that's the definition of success: having researchers around the world recognize his work as something strong enough to use and build on.
He also sees future applications with the program for the general public, including integration with technologies like Google Maps. He envisions an interactive map of the ocean, where users could view historical and current images of oil spills, red tides, and other elements his programs detect.
Zhang never figured he'd be a leader in identifying and tracking oil spills. He just saw a new problem to tackle with his tools. He knows that, unfortunately, it's a problem likely still looming.
"It was a disaster. We don't want to see it happen. But it does happen. And in the future, it may happen again."
Zhang lives in the middle of the country but has never considered searching for a location closer to his research. Given today's technology, it doesn't much matter. He says that all things considered, he'd rather be in an office in Winona than standing on a beach in west Florida.
"The nice thing about this technology is that you see everything you need to see, you have the idea and the vision about what's going on over there," he says.
"If you go down to the Gulf of Mexico, if you get in a boat and go out to the ocean, you won't be able to see a thing."