In the spotlight: Mutlu Ozdogan draws climate clues from satellite imagery

July 30, 2013

With access to more than 40 years of satellite imagery and innovative computer models, Mutlu Ozdogan is zeroed in on how humans are changing Earth’s landscapes and to what effect.

Mutlu Ozdogan
Mutlu Ozdogan

An assistant professor of environmental studies and forest ecology, Ozdogan strayed from his original path in geology to pursue more immediate timescales. He now focuses on using remote sensing data to predict how land-use changes may affect natural resources like water that are important for the well-being of citizens around the world.

“There are a number of brand-new tools at the intersection of computer sciences and environmental sciences that can be used to improve things like water resource management or agriculture and help people today,” says Ozdogan, a faculty member of the Nelson Institute Center for Sustainability and the Global Environment. “I work with the computer science community to bring these tools into remote sensing.”

What is the focus of your research?

Ozdogan: One of the things I’m really interested in is how humans are changing a variety of landscapes across the planet: through agricultural expansion, deforestation, forest harvest, irrigation expansion and more. I monitor those changes using remote sensing data. We have a digital archive of images of Earth going back to 1972, all taken from satellites with cameras. We can look at any place around the planet and see if and how the landscape has changed.

I’m also interested in the physical consequences of those changes. For example, if we cut down half the forest in Northern Wisconsin, what does that mean for river flows? What does that mean for the carbon and water cycles? What does that mean for the weather, as land surface strongly regulates the atmosphere?

"We can look at any
place around the planet
and see if and how the
landscape has changed."

Those are the two pieces of my research. In the first part I use digital image processing tools to extract these archival and recent images. And for the second part I either use statistical tools or computer models to understand the consequences.

Why is this research important?

As an example, I’m currently assisting NASA Earth System Science’s Applied Sciences program with a large water resources project in the Middle East. The World Bank gave $5 million to five different countries – Morocco, Egypt, Tunisia, Lebanon and Jordan – to better manage their water resources. Our research team is developing tools and training for these countries. I’m hosting a group this October and November to come here and train with these new tools. If I can make even a 1 or 2 percent change in the way they manage their water, that’s huge.

What drew you to this area of study?

I grew up in a small city in Turkey right by the water, so I was always drawn to nature and environment. I then studied geology as an undergraduate at Istanbul University in Turkey. When I came to America, I did my master’s in geology at North Carolina State University. I was going to continue with geology when I went to Boston University for my Ph. D., but I realized that most geologists deal with long time scales rather than immediate concerns. I switched to geography and remote sensing.

You recently began work on a project to map the world’s crops. What impacts do you anticipate from this research?

The project is funded by NASA’s MEaSUREs (Making Earth System Data Records for Use in Research Environments) program to create large-scale data sets from space-based observations and computer models.

An Ethiopian farmer sows Teff
A farmer sows teff in the red soils
of Ethiopia. Ozdogan is part of a
new project mapping crops globally. 

The goal of the project is to not only map crop extent across the planet, but also crop type, irrigation status and perhaps yield – all of which is increasingly more difficult to measure. We can tell if a particular piece of land is agriculture or not, but telling if it is corn or soybeans is much more difficult.

This particular project is important because there are certain places around the planet where we have very little crop information. People are making management decisions based on the existing data, which is inaccurate – especially in Africa. There are a variety of reasons for that: few ground observations, neglect due to conflict and other issues, and limited interest in the research.

We’re trying to improve the state of knowledge by taking a step back and putting the basics on the ground. If you know what kind of crops are growing where, you might be able to find out if those crops are suitable or not and suggest alternatives.

Also, most climate change scenario data is generated using computer models that rely on land cover type. In Africa, most land cover types but especially the cultivated categories are wrong, meaning the predictions in those parts of the world may also be wrong.

What are some of the most important or interesting things you’ve learned in your research?

Personally, the most crucial thing I’ve learned is that my work matters. One of my advisors at Boston University told me that a graduate student has to realize that the work he or she is doing is important. Unfortunately, I realized that really late. For me it wasn’t a very fast transition from becoming a student to becoming a researcher – it took me awhile. Now I feel very comfortable. I’ve learned as I’m doing research and I’m also having a lot of fun.

In my research, one of the most interesting things I learned recently was from a paper where we teamed up with Rutgers University to simulate climate change scenarios. We used the same models that show what would happen with rises in carbon dioxide or methane, but simulated what would happen during a limited nuclear war between India and Pakistan.

The idea was that if all the big cities in India and Pakistan were attacked with nuclear arsenal there would be a lot of fires and those fires would lead to a lot of soot and black smoke in the atmosphere, inhibiting solar transmission. We took a very limited war, assimilated the smoke that would come out of that war, and then evolved the model to find out what would happen to temperature, precipitation and solar radiation across the planet. We applied the results to four locations in the midwestern United States and found that corn yields under the war scenario would go down by up to 30 percent, just from the climatic impacts of a war nearly 10,000 miles away.

When did you come to UW-Madison, and what attracted you to the Nelson Institute?

Mutlu Ozdogan at cotton field in China
Ozdogan, center, visits an experimental cotton field in
Xinjiang, China. He 
and his students are studying the
environmental effects of large-scale irrigation in the region.

I was always interested in the innovative research taking place at the Center for Sustainability and the Global Environment (SAGE) through the work of Jon Foley, Navin Ramankutty, and Chris Kucharik. SAGE made a name for itself as the first group to develop widely used data sets at fairly fine resolutions. But I also knew that SAGE didn’t do a lot of remote sensing at the time, so I thought I could contribute. I came to UW-Madison in 2007.

What is your favorite course to teach at the university and why?

My favorite course to teach is Environmental Studies 556: Digital Image Processing. It’s thrilling because I teach things that I turn around and use on a daily basis. I’ll teach students how to classify an image in the morning and in the afternoon I’ll use those same tools to do my own research and publish a paper. There’s immediate gratification.

At the end of the semester we do final projects and it’s very rewarding to watch students apply the tools they learned and then actually produce something. The course draws students from across campus – engineering, geography, the Nelson Institute, landscape architecture and more.