Originally posted on 21 October 2011 at Skeptical Science
A paper submitted for peer review by the Berkeley Earth Surface Temperature study (BEST)finds that urban heating has an influence on global temperature trends that is “nearly negligible” and that what effect has been observed is even slightly negative, which is to say that temperature trends in urban areas are actually cooler than the trends measured at rural sites, and that the Earth’s land surface has warmed approximately 1°C on average since 1950.
The Urban Heat Island Effect
It has long been observed that temperatures in cities are higher than in the surrounding countryside, caused, in part, by human structures that reduce albedo and evapo-transpiration, as well as by the effects of waste heat emissions, McCarthy et al 2010. Even though most (99%) of the Earth’s surface is not urbanized, some 27% of the Monthly Global Historical Climatology Network (GHCN-M) temperature stations are located in cities having populations of more than 50,000. Since urbanization has grown dramatically over the past few centuries, it seems reasonable to ask how much of the observed rise in global temperatures is due to urbanization. For example, McKitrick and Michaels claimed in 2007 that about half of the recent warming over land is due to urban heat island effect, although this result was disputed bySchmidt in 2009.
Several studies have looked in depth at possible heat island contamination of land temperature records. For example, Hansen et al in 2010 used satellite measurements of night lights to remove possibly affected urban sites from the temperature records. Allowing for urban effects reduced the global temperature trend over the period 1900-2009 by 0.01°C. Menne et alperformed an analysis on US surface temperature records and found only very small effects due to poor siting of temperature stations. They reported that poorly-sited stations did show some small positive bias in recorded minimum temperatures but there was a small, but slightly larger, negative bias in recorded maximum temperatures, resulting in an overall small residual cool bias.