Sean Turner draws on his international experience and the latest deep learning tools to better understand water resources that can reshape how the nation powers its future. His work with utilities has given Turner insight into how to enhance the efficiency and reliability of nuclear and hydropower operations, paving the way for energy abundance in an era of rapidly rising demand.
Turner, a senior engineer in the Water Resources Science and Engineering Group at the Department of Energy's Oak Ridge National Laboratory, specializes in building hydrology models that simulate the behavior of water in complex systems. He is creating national-scale models that can solve real-world problems, even when data are scarce. He is currently using large sample, deep learning neural network models to predict how rising water temperatures could affect the Tennessee Valley Authority's hydropower and nuclear operations, which are closely connected.
"Water temperatures are important to the operation of nuclear power plants in the region, which need an adequate supply of cool water from hydropower dams further upstream," Turner said. "If the water temperature is too high, this process becomes less effective."
Reliable river temperature data is surprisingly limited. Turner said that the United States has only about 300 comprehensive, long-duration records of river temperatures, and data for just two river reaches in Tennessee. In years past, this would have been an intractable problem, Turner said, but large sample, deep learning models now allow researchers to extract far more insight from small datasets. With this new technology and ORNL's supercomputing facilities, Turner and colleagues can simulate fluctuations in water temperature in any of the 2.7 million stream reaches in the lower 48 states using the modest dataset available.
And that's just the beginning.
"These large sample approaches are transformative for hydrology, and we haven't tested them to the limits yet," Turner said.
An engineering career begins
Turner grew up in Bridge of Weir, a small town in the Scottish Lowlands that owes its existence to waterpower provided by the River Gryffe, which drove the textile production and leather tanning industries that are central to its history.
He earned his undergraduate degree in engineering and environmental design from the University of Glasgow, then joined an environmental consulting firm in nearby Glengarnock. After about a year at the firm he decided to go to graduate school and enrolled at Newcastle University in northern England, where he completed a master's degree in hydrology.
In 2010, Turner entered an engineering doctorate program at Cranfield University near London.
"Instead of a typical university setting, you did your studies with an industry partner," Turner said. "You still had to do a dissertation, but you were physically based in a company and working on a problem relevant to them."
Turner joined United Utilities, a company that provides water to seven million people in northwestern England. There, he used company models to better understand drought risk and its effect on supply.
"That experience has always kept me grounded in real-world applications," Turner said. "I always ask: how is this useful for industry? Does this really matter, or am I going off in an academic bubble?"
Diving into hydrology and modeling
In 2016, Turner moved to the United States as a postdoctoral researcher with the Joint Global Change Research Institute, a collaboration between the University of Maryland and Pacific Northwest National Laboratory. He later became a full-time Earth systems scientist at PNNL, where he studied how extreme weather patterns could affect the West Coast's hydropower systems, a key electricity source for the region.
In 2023, Turner joined ORNL, where he applies his skills as a modeler and hydrologist to practical problems that arise where water and energy production intersect. His next goal is to integrate national-scale predictive hydrological models with models that can simulate the operations of the U.S. power grid.
"Enabling the exchange of data between models of the grid and rivers could support the scheduling of hydropower resources and increase power grid reliability," he said. "This begins with fundamental research into model coupling using custom-built river operations and power grid models. We're now developing those tools and designing experiments that will reveal the most efficient and optimal ways of integrating rivers into power generating models and decisions."
Turner said the river temperature and quality simulations enabled by large sample deep learning will be useful in a variety of ways.
"These models could also support siting decisions for new power plants, including small modular nuclear reactors and data centers, both of which rely on water resources for cooling," Turner said.
Sharing the wealth (of data)
Turner is an advocate for open research.
"It's important to conduct your science in a way that is reproducible," he said. "Every paper I publish includes input data and calculations so anyone can replicate the work."
He encourages early-career scientists to be receptive and broad-minded, not just when it comes to sharing data with peers, but also with respect to emerging technologies. Developing programming and AI fluency is a must.
"It's important to be able to use the latest tools to stay up to date," Turner said. "Getting into an abundance mindset with your science - proactively sharing your results, data, and new methods - gives you the power to generate new ideas."
UT-Battelle manages ORNL for the Department of Energy's Office of Science, the single largest supporter of basic research in the physical sciences in the United States. The Office of Science is working to address some of the most pressing challenges of our time. For more information, please visit energy.gov/science . - Clare Kennedy
