New research led by University of Hawai‘i (UH) at Mānoa atmospheric scientists found that the century-old yardstick used across many industries to estimate air-conditioning and refrigeration energy demand — called cooling degree days — is missing something fundamental. Their study, published recently in Nature Communications, reports a new, physics-based version of the metric — effective cooling degree days — that captures how temperature and humidity together impact the actual work a cooling system is required to do. Applying it across North America, they found that cooling efficiency has been quietly declining by 2–4% per decade since 1971, and that the old temperature-only metric misjudges how much cooling is required in different regions, overstating demand in some places and understating it in others.
“Estimates of cooling degree days are used everywhere,” said Jake Casselman, Atmospheric Sciences postdoctoral researcher in the UH Mānoa School of Ocean and Earth Science and Technology (SOEST). “Utility companies, grid operators, energy planners, and engineers use them to anticipate electricity demand, and financial markets trade futures on them to hedge against unusually hot summers. If that yardstick is biased in ways that depend on a region’s climate, then the planning decisions built on it are biased too. That can mean building the wrong amount of power generation in the wrong place, or misjudging where the grid is most at risk during a heat wave.”
Challenging an old assumption
The previous approach to estimating cooling requirements assumes every degree of heat takes the same amount of energy to cool, no matter the conditions. But real cooling systems don’t work that way: they get less efficient as it gets hotter, and humidity makes it worse, because the system has to spend energy wringing moisture out of the air, not just lowering the temperature.
Casselman and Christina Karamperidou, an atmospheric sciences professor in SOEST, combined climate science with refrigeration engineering to develop a new cooling demand metric that explicitly accounts for how cooling system efficiency changes with temperature and humidity. Their approach incorporates a simplified model of the refrigeration cycle—the same physics governing every air conditioner and refrigerator—to estimate how efficiently a cooling system can remove heat under different atmospheric conditions. They applied this new cooling demand metric to fifty years of high-resolution weather data (1971–2020) across North America to quantify how climate-driven changes in cooling efficiency have already reshaped cooling demand across regions. They then repeated the analysis using projections from 19 climate models under a high-emissions scenario to assess future changes. Finally, they mapped the results onto the U.S. electricity grid, accounting for where people live, to identify the regions and power systems most likely to experience the largest shifts in cooling demand.
“Our results show that the regions facing the steepest future increases in cooling demand are the Northwest, Great Lakes, and Mid-Atlantic, where some grid regions are projected to see cooling-related electricity demand more than double by mid-century in this high-emissions ‘worst-case” scenario’,” Karamperidou shared. “Getting cooling demand right isn’t an academic exercise; it directly affects how we plan, size, and operate future energy infrastructure as the climate changes.”
New insights from a more elegant approach
While the researchers were not surprised that cooling gets less efficient as temperatures climb, as that is basic thermodynamics, the impact of humidity introduced a real twist. They reported that in some regions, like the desert Southwest, they expected efficiency to be falling, but once they accounted for the air also getting drier, the decline at times vanished, and in a few spots efficiency actually held steady or even improved. Drier air is genuinely easier to cool, and that partly cancels the penalty from higher temperatures. The flip side was just as striking: in humid regions, heat and moisture compound each other, so the energy burden is worse than temperature alone would suggest.
“Seeing these two effects pull in opposite directions across the continent, what we ended up calling a ‘tug-of-war’, is something a temperature-only view would never reveal,” said Casselman.
This project connected fundamental climate science with real-world challenges. It was both motivated and made possible through the researchers’ interactions with thermal-management engineers at the National Science Foundation-funded Environmentally Applied Refrigerant Technology Hub (NSF ERC EARTH), of which UH is a key partner institution.
As cooling systems are genuinely complex, Casselman and Karamperidou had to incorporate a refrigeration model simple enough to apply everywhere, yet faithful enough to capture the real physics of how temperature and humidity change efficiency. Through their interdisciplinary collaborations with researchers working on next-generation cooling technologies at NSF ERC EARTH, they were able to strike the right balance: enough engineering detail to be meaningful and enough simplicity to be computable at climate scale.
This project is an example of UH Mānoa scientists addressing a problem that touches nearly everyone on the planet, how we stay cool as the climate warms, and solving it by bridging disciplines, from atmospheric physics to engineering.