Most scientists agree that the world is getting warmer due to anthropogenic carbon emissions. Some areas, such as the Arctic, appear to be warming faster than others. The phenomenon is temporal as well as geographical: in general, night-time temperatures have increased more than daytime temperatures.

There have been explanations for this unevenness or asymmetry. For instance, the Arctic is thought to be warming faster because of ice-albedo feedback: as ice melts, it exposes a darker surface underneath, which absorbs more sunlight and so causes more ice to melt. But according to Richard Davy at the Nansen Environmental and Remote Sensing Centre in Bergen and colleagues, such explanations have been sidelining one crucial variable: the atmosphere's heat capacity.

Ignoring heat capacity, said Davy, has left a gap in climate scientists' understanding of global warming. This gap "has been used by the climate change skeptics who claimed that the temperature is not rising in some regions, or that the temperature rise is due to natural variability there," he added. "Our explanation is narrowing the gap for speculations."

In the atmosphere, heat capacity is dependent on the depth of the planetary boundary layer, which can be between 50 and 2,000 metres thick. Davy's group hypothesized that the observed asymmetry in global warming could be at least partially accounted for by variations in the boundary layer: regions with a shallow boundary layer would have a relatively low heat capacity, and would warm faster, while regions with a deep boundary layer would have a relatively high capacity, and warm more slowly.

The researchers tested their hypothesis by making warming predictions for different regions of the planet, based on the depth of the boundary layer. One of these predictions was that sub-tropical deserts, which have shallow boundary layers, should be experiencing relatively strong warming. This prediction – and others made by Davy and colleagues – seemed to be borne out in temperature data.

Davy believes that the influence of the boundary layer and heat capacity may have been overlooked because previous climate models have often used "over-simplified" boundary-layer processes in regions dominated by shallow, fast-warming boundary layers. "Only recently have we developed sufficiently accurate models to study the shallow boundary-layer processes with the required accuracy," he said.