Because anthropogenic heat is generated close to where temperatures are measured in
both countries, we have not used a climate model to investigate the transport of such
released heat further afield. Early attempts to do this globally found temperature
variations of a similar order to the model’s natural fluctuations (Washington 1972), and
Flanner (2009) found no significant effect for the present day. Oleson (2012) used
CMIP5 simulations to model future changes in urban minus rural temperatures in
response to changing climate over the 21st century, rather than the effects of changing
energy consumption. More recently however, Zhang et al. (2013) despite including only
42% of world energy consumption in their model, found significant winter and autumn
temperature changes up to 1° C in mid- and high- latitudes, far from heat sources, that
correspond well to areas of previously unexplained differences between observed and
modelled temperatures. Chen et al. (2014), entering anthropogenic heat flux into a
refined model that included long wave radiation, found higher and more widespread
increases over standard models: 1°-2° C in mid- to high-latitude areas of Eurasia, North
America and parts of the southern hemisphere, and concluded that anthropogenic
heating is an important factor in global warming that should not be ignored. Our study is
the first of its kind that provides direct observational evidence of this.