PELLSTON, Mich. — Dr. Anshuman Swain likes searching for valuable, actionable science in overlooked places.
At Harvard, that meant helping uncover a rare 320-million-year-old fossil site in a Massachusetts parking lot.
At the University of Michigan Biological Station (UMBS), it’s using artificial intelligence, machine learning and large language models to help analyze more than a century of field research archives, some of which is digitized but much remains untouched in storage closets and filing cabinets.
“Many people assume most human knowledge is online, but it’s not. Most human knowledge is sitting in basements, unread,” Swain said. “That’s the thing about dark data. We don’t know what is out there. A lot of natural history stuff is buried in plain sight in places like UMBS and museums.”
Swain is a new assistant professor in the U-M Department of Ecology and Evolutionary Biology and a curator in the Museum of Paleontology in Ann Arbor.
He spent a week at UMBS in late May to scope out research possibilities in northern Michigan habitats with Resident Biologist Adam Schubel and Director Aimée Classen and test the AI waters on the field research station’s deep historical archives spanning 117 years with Data Manager Jason Tallant.
“The people here are infectious,” Swain said. “Not only the incredibly helpful staff, but the visiting research scientists and course instructors are energetic and excited about making science happen too. They’ve taken me to bogs, fens, meadows, and all kinds of forests. I have so many leads.”
Swain’s research philosophy involves understanding local nature, both living and dead.
For his ongoing study of plant-insect interactions, thermogenic flowers are one possible research direction.
Cutting through cold northern Michigan air, thermogenic flowers heat up to attract pollinators, instead of changing color or shape.
Skunk cabbage, he said, can get hot enough to push through and melt snow to lure early pollinators such as flies and beetles. And there are large patches of skunk cabbage around the more than 10,000-acre research and teaching campus.
It’s a local connection to evolutionary history, long before brilliantly colored flowers evolved to dazzle bees and butterflies.
“Before there was color, there was heat,” Swain said. “You can learn about evolution through ecology.”
In addition to skunk cabbage, Swain named pawpaw, some water lilies, and magnolias as examples of thermogenic flowers, what he called few of the oldest branch of flowering plants.
Another research direction is pollinator choices — how pollinators choose which flowers to visit based on nutritional or elemental needs.
The technical term is stoichiometry. It means the balance of elements such as carbon, nitrogen and phosphorous in organisms.
Swain is interested in whether pollinators are not just looking for pollen, but for the right recipe of nutrients and whether those choices shape plant-insect relationships over time.
“Through the spring and summer, flowers come and go. Pollinators have to continually learn what to adapt to,” Swain said. “In some bees, for example, the pollen they collect is correlated with their own carbon-nitrogen-phosphorous ratios in their body. How much does what makes up an organism play a role in shaping its interactions with plants? And how does this change as a function of space and time?”
Swain’s work also includes microbial ecology, fossils and planktic foraminifera. Known as forams, those are single-celled marine organisms that build shells and play a critical role in ocean carbon cycling.
Across all these systems, he studies how environmental conditions and biological interactions shape communities over ecological and evolutionary time.
Before coming to the University of Michigan last year, Swain was a junior fellow at Harvard’s Society of Fellows. He earned his Ph.D. in biological sciences from the University of Maryland, College Park, after beginning his academic path in physics. His research continues to reflect that interdisciplinary background, combining field biology, paleontology, quantitative methods and large-scale questions about ecological change.
Swain is interested in UMBS as a living laboratory that is uniquely positioned to answer a wide variety of important questions about our natural world, where past and present data meet.
He considers the 117-year history at UMBS a scientific gold mine.
And the timing of Swain’s arrival couldn’t be better.
As part of the field station’s five-year strategic plan, UMBS is in the process of preserving, analyzing and digitizing field data and reports that have been collecting dust in storage to the equivalent of a 120-foot-tall stack of paper encompassing research across different fields of study.
The goal is to unearth treasure troves of research to be publicly shared and used by researchers around the world because those files could include observations that are scientifically valuable today, especially for understanding long-term environmental change.
Swain wants to use AI tools to attack those records and transform the dark data into new science.
In fact, he began scratching the surface with already digitized UMBS data during his brief visit in May.
“This is a good use of AI,” Swain said. “New technology can help extract insights that lead to new investigations and discoveries much faster, and that’s thrilling. So many researchers have come and gone. It’s a major strength of UMBS. We can do more with the data they collected beyond their published papers. They have more to teach us.”
For example, a dataset gathered in the 1980s at Green Star Meadow — including insect observations that were never fully used — is already helping Swain explore how plant-pollinator networks shift through a growing season and how those patterns compare across decades.
“We are using data collected by someone in the 1980s and it’s still cutting edge,” Swain said. “One of the UMBS classes this summer may resample those same plots to compare.”
Another example of breathing analytical life into the historical archive is a dataset from the 1940s that looks at ecosystem types and how geology may help define vegetation patterns.
“The idea is to also get undergraduate students involved, and not just with grunt work,” Swain said. “AI opens new opportunities to help us access better questions faster and study long-term changes in forests and other ecosystems. The new tools can help us brainstorm undergraduate student research projects that are doable with data that’s sitting there waiting to be reinterpreted. The next generation of scientists would learn new research skills and contribute to real science.”
And those students wouldn’t have to start from scratch. They’d get a leg up from the century of researchers who came before them using first-hand data hidden in an archive.
“That’s the beauty of field stations,” said Director Aimée Classen. “They hold the long-term study of a place. No matter what you want to investigate, we have priceless datasets from which to build and discover. At the University of Michigan Biological Station, we’re happy to welcome Anshuman to our community of research and can’t wait to see the path he chooses to explore.”
The University of Michigan Biological Station serves as a gathering place to learn from the natural world, advance research and education, and inspire action. We leverage over a century of research and transformative experiences to drive discoveries and solutions to benefit Michigan and beyond.
Our vast campus engages all of the senses. Its remote, natural setting nurtures deep thought and scientific discovery.
Founded in 1909, UMBS supports long-term research and education through immersive, field-based courses and features state-of-the-art equipment and facilities for data collection and analysis to help any field researcher be productive. It is where students and scientists from across the globe live and work as a community to learn from the place.
Subscribe to the UMBS monthly e-newsletter and follow the field station on LinkedIn, Instagram, Facebook, YouTube and X (formerly Twitter).
