Geophys. Res. Lett. http://doi.org/xhh (2014)

Surface temperature is typically reported as a global average when considering climate change. The use of a global average allows us to see the overall trend in temperature change, but results in the loss of important spatial information.

To investigate trends in warm and cold temperature anomalies and their spatial pattern, Scott Robeson of Indiana University, USA, and co-workers apply a spatial percentile approach to a gridded temperature dataset on a monthly basis. Anomalies are calculated by comparison with the 1961 to 1990 period and analysis was performed individually on both hemispheres.

The analysis finds, for the period 1881 to 2013, cold anomalies warmed by a greater magnitude than warm anomalies across the globe year round. Winter cold anomalies in the Northern Hemisphere underwent some of the fastest change, and globally there was a reduction in the spatial extent of anomalies for the period. However, when considering a shorter, more recent period (1984 to 2013), the authors report that warm anomalies were warming faster in both hemispheres, causing an increase in the spatial distribution of warmer temperatures.

These findings highlight that extreme Northern Hemisphere anomalies are the most variable on decadal timescales and could be used as indicators of global temperature variability.