Understanding the Air We Breathe
Associate Professor of Chemistry Karena McKinney teams with Davis Institute for AI on prestigious NSF grant
The way we breathe is so automatic that it takes almost no thought. Over the course of reading this paragraph, you’ll take anywhere from 12 to 18 breaths, each containing a mix of oxygen, nitrogen, and argon gasses. In, out, in, out.
That’s exactly what Associate Professor of Chemistry Karena McKinney thinks about every day in the lab. She studies biogenic volatile organic compounds (BVOCs), or in layman’s terms, the emissions from trees.
“As an atmospheric chemist, I’m looking at chemical processes related to air quality and climate. My research has been focused specifically on plant emissions of organic compounds, which affect atmospheric composition, and how they interact with other, human-made emissions,” said McKinney. “Forest emissions contribute to small particles that form in the atmosphere that generally have a cooling effect on climate. As climate changes, the trees change, which impacts the emissions, and it becomes a feedback loop. I want to understand how it’s changing.”
Chemistry for the climate
Her latest project, funded by a $260,000 grant from the National Science Foundation, will map how forest composition changes the delicate balance of gasses in the atmosphere. “I want to understand what compounds plants are emitting, under what conditions, and how it varies from other plant or forest types over time,” she said. “These emissions change all the time. If it’s warm and sunny, the plant emits different compounds than when it’s cold and cloudy. These questions are all important to understand how it affects atmospheric composition.”
Working with wildly different ecosystems—including samples from Colby’s 128-acre Perkins Arboretum and the Amazon Rainforest in Brazil—she hopes to build a model that highlights the impact of climate change on these ecosystems. The grant allows funding for two research trips with students to the Amazon to compare and contrast these data sets, providing opportunities for undergraduate students to become involved in innovative STEM research.
“Different trees emit different compounds, but the types of compounds we’re looking for fit into broad categories,” said McKinney. “Traveling to Brazil, we’re asking the same questions and doing the same measurements. How are emissions changing with time? This is an area that’s incredibly important for global climate. It’s one of the world’s largest forests in terms of atmospheric impact.”
Her project will leverage advances in imaging technology, artificial intelligence, and machine learning techniques. Combining cutting-edge technology with old-fashioned fieldwork opens up a new avenue of previously unexplored research. “In addition to individual trees, we also use drones to collect samples above the forest, which allows us to cover a larger expanse and reach places you can’t get to on the ground,” said McKinney.
The project has the potential to develop tools that could have a range of applications in environmental sensing, which could be of particular interest to the atmospheric chemistry and broader environmental science community. There are also plans for a summer institute on AI-driven sensing and environmental modeling.
Colby student-powered fieldwork
On campus, she works both with chemistry students and environmental studies majors like Silas Gramaglia ’23, who conducted summer research on BVOCs. “I was able to work on two different projects with Professor McKinney that really opened my eyes to the world of climate chemistry, which I didn’t know existed,” said Gramaglia, who plans to apply to graduate school in the discipline. “My sophomore summer, I worked in the arboretum sampling trees to measure the differences in emissions between coniferous and broadleaf trees, which was such great field experience.”
Gramaglia was so inspired by his summer research that he returned to climate chemistry for his senior capstone with partner Luca Villani ’23. “As deforestation becomes more prevalent around the globe, we wanted to understand the effect of the forest edge, where trees are exposed to more light and more heat, on emission composition. Twenty percent of tropical forests are within 100 miles of the edge, and that number is growing. What is that doing to the air?” he said. “[Professor McKinney] helped guide us through that research in the lab, and she went out of her way to make sure our research project came to fruition. She was a great mentor to me throughout my time at Colby.”
From the arboretum to around the globe
But McKinney knows that understanding emissions in specific forests is only one small part of what makes up global climate change trends. To build a deeper understanding of forest emissions around the world, McKinney is partnering with the Davis Institute for Artificial Intelligence. “Obviously, sampling every forest in the world isn’t feasible, but I want to be able to extrapolate our results so that others can use them. That’s where AI comes into play,” said McKinney.
Ashley Ren ’24, a statistics and computer science major, has spent her senior year thesis project fine-tuning the artificial intelligence model that will help identify changes in emissions over geographical space and time.
“The main goal of my research is to build a deep-learning model of BVOCs that is easy to use, using satellite imagery to detect tree location and species,” said Ren. “My role is not to identify emissions but to train the AI to understand which characteristics differentiate the trees so that we can combine the datasets in the future.”
‘Sampling every forest in the world isn’t feasible, but I want to be able to extrapolate our results so that others can use them. That’s where AI comes into play.’
Karena McKinney, associate professor of chemistry
For Ren, this is a chance to explore a new avenue of research. “I was really interested in this because I hadn’t had a chance to work on computer vision before, which is an exciting aspect of computer science. And working with Professor [Amanda] Stent in the Davis lab is such a supportive environment. Each week, we all take turns to present our progress and help each other when one of us gets stuck on something.”
McKinney hopes this AI model can become sophisticated enough to project hypothetical emissions from global forests over time. “If we can figure out these relationships between trees and emissions even in forests we haven’t sampled, we can better understand how these processes are changing over time as forest compositions change. How is climate affecting emissions, and how are the emissions affecting climate?”
