Geophys. Res. Lett. 35, LO1702 (2008)

Credit: EDWIN OLSON

Some climate models predict that the agricultural heartland of the US will warm as much as 7 to 8 °C by the twenty-second century and dry out drastically, whereas others predict much more modest changes. Researchers have now found an important cause of this variability in models: a mechanism known as snow albedo feedback, which occurs as the snowpack retreats and the exposed ground absorbs more solar radiation.

Alex Hall and colleagues at the University of California, Los Angeles compared sensitivity to snow albedo feedback across 18 climate forecast models that were used in the most recent assessment report of the Intergovernmental Panel on Climate Change. Models with the strongest feedback from the melting of snow in winter and spring yielded the most warming and greatest drying in summer. By comparing feedback strength from the models to independent estimates from satellite data, the researchers found that over half of the models had unreasonable estimates of snow albedo feedback.

The researchers say that variability in climate predictions for the interior US could be reduced by one-third to one-half by including accurate observations of this key parameter in the next generation of climate models.