The world is full of “huge, gnarly problems,” as Oak Ridge National Laboratory research scientist and musician Melissa Allen-Dumas puts it — no matter what line of work you’re in. That was certainly the case when she would wrestle with a tough piece of music.

“There are passages in music that are much harder than other passages, and that require much more practice,” said Allen-Dumas, who performed and taught the violin and viola professionally for much of her life. “It’s important to take those and break them down into easier bites for yourself, and to solve smaller problems.”

Chopping big problems into manageable pieces, as it turns out, is also the crux of her work now as a computational scientist in ORNL’s Computational Sciences and Engineering Division. There, she tackles global climate change by developing methodologies that scale that big, gnarly problem down to the level of cities — even specific neighborhoods. Pulling together different kinds of data — on weather, building size and shape, migration patterns and demographics, for example — she sheds light on the effects of climate change at the local scale, and on how communities can best prepare for them.

In a 2020 study using data on ORNL’s campus and a Chicago neighborhood, for example, Allen-Dumas and her coauthors used computer simulations to project how different proposed developments would affect the building site’s microclimate. They learned that, depending on variables such as building design, land use and population density, some scenarios would lead to notably warmer microclimates. What’s more, those effects extended beyond the development’s boundaries and into the surrounding area.

Buildings consume three-quarters of the electricity used in the United States, so being able to assess the effect of their shapes and sizes could help cities better mitigate climate change. Allen-Dumas hopes her methodologies can be developed into tools that officials addressing climate change can use at the local level.

Much of her research examines climate impacts on critical infrastructure and the unique challenges facing the southeastern U.S., which produces more carbon dioxide and has a faster-growing population than other regions of the country.

In one study examining data from Knoxville and Atlanta, Allen-Dumas and her coauthors demonstrated a methodology for planning for future land, water and energy needs in the face of climate change. The study presented numerous scenarios for those cities in detail, highlighting the advantages and trade-offs of each. Atlanta could reduce its carbon footprint by heavily investing in solar power, for example, but that path would require significant land resources. The findings could help urban planners make better decisions as they project long-term investments in infrastructure.

Allen-Dumas’s other work includes development of a method to pinpoint which electrical service areas in the southern U.S. will be most vulnerable as populations grow and temperatures rise, and an analysis of post-Katrina climate migration revealing where the quarter of New Orleans residents who left the city for good ended up.

In her research, Allen-Dumas also finds evidence of how disadvantaged communities suffer disproportionately from climate change, an issue the federal government has been seeking to address. In an upcoming study examining the city of Las Vegas, for example, she and her colleagues found that more vulnerable populations tend to live in older buildings located in the hottest part of the city. Currently, Allen-Dumas is working on models that predict what the cities of the future will look like, and how to factor these new urban morphologies into simulations on climate change.

ORNL’s world-leading supercomputers and massive data sets have been critical to her work, Allen-Dumas said. But more important, she said, is the ability to collaborate with lab colleagues who are recognized authorities in diverse areas of expertise.