Cities only occupy about 3% of the Earth’s total land surface, but they bear the burden of the human-perceived effects of global climate change. Yet, current global climate models are set up mainly for big-picture analysis, leaving urban areas poorly represented.

In a new study co-authored by Princeton University, researchers take a closer look at how climate change affects cities by using data-driven statistical models combined with traditional process-driven physical climate models.

Their results predict that by 2100, average warming across global cities will increase by 1.9 degrees Celsius with intermediate emissions and 4.4 C with high emissions. They also show a near-universal decrease in relative humidity in cities, making surface evaporation more efficient and implying that adaptation strategies like enhanced urban vegetation could be useful.

“With warming approaching the limits targeted in the Paris Agreement, emissions reduction alone, no matter how stringent, is unlikely to avert dangerous outcomes for many urban areas, where adaptation is especially critical because dense populations intersect with high exposure and vulnerability to climate risks. This new approach will offer policymakers a global view of the necessary objectives of urban adaptation policies and measures,” said Michael Oppenheimer, Albert G. Milbank Professor of Geosciences and International Affairs and the High Meadows Environmental Institute and director of the Center for Policy Research on Energy and the Environment.