A temperature increase of at least 1.4 °C is needed before changes in regional precipitation can be distinguished from regular variability and attributed to global warming.

A team led by Irina Mahlstein of the National Oceanic and Atmospheric Administration in Boulder, Colorado, used a suite of general circulation models to analyse regional precipitation trends from 1900 to 2099. The analysis focused on wet seasons, for which models performed most accurately against historical trends.

By the end of this century, the study suggests, increases in wet-season precipitation will be apparent in many areas. However, the authors note that changes in extreme weather or annual precipitation might be detectable much earlier.

Geophys. Res. Lett. http://dx.doi.org/10.1029/2011GL050738 (2012)