2. 9 UHIs & Climate Change

Figure 2.9: Using clusters of weather stations in the top left plot, a station’s temperature and population difference from its cluster average shows a weak correlation. Another analysis of these clusters, in the top right plot, shows a small urban heat island for stations with at least 20,000 people (filled circles). These results demonstrate, using weather station measurements, that an urban heat island exists. Regarding global temperatures, the bottom plot separates urban and rural stations at 30,000 people within 6 km in the year 2000. The curves show little difference (dark sits on top of gray), meaning the UHI didn’t cause the measured temperature increase over the last half-century (after Peterson and Owen 2005).

Despite the seemingly clear results of Figure 2.7, the urban heat island’s existence has been questioned[28] based on arguments that the UHI arises from measurement biases. These biases aren’t scientific misdeeds; rather, complications abound in measuring temperature. Once these many problems are corrected, the argument goes, there’s no difference between urban and rural areas.

While we consider these monitoring station studies, temporarily disregard the satellite images showing urban heat islands.[29] Clusters of measurement stations are scattered all around the United States. For each station of each cluster, the number of people living within 6 km can be counted, and its temperature and population difference from its cluster’s average calculated. The top left curve in Figure 2.9 plots these pairs of differences. An urban heat island exists, however weak, with at most 0.5C temperature differences. Another approach defines a threshold number of people living within 6 km of a weather station differentiating urban and rural stations. Using this threshold, data support the existence of an UHI for populations of 14,000 to 50,000 people in a circle of 6 km radius, but again, the heat island effect was small relative to Figure 2.7 UHI values.

Other studies compared UHIs in different places. In one such example, an examination found strong UHIs in the eastern United States but not in the west.[30] Similarly, coastal differences could cause a nationwide study to wash out important effects in one area due to no effect in other areas.

Let’s reconsider the Los Angeles temperature rise, which is seemingly consistent with the growing UHI of a growing population. Do those results mean that, perhaps, there is no global warming, and the observed global temperature rise might just reflect the effects of the urban heat island, an effect that has gotten worse because our hugely increased population (see Figure 1.1) resides mostly in cities? Perhaps easily fooled climate-change scientists have simply been sloppy with their science.[31] Using this collection of weather stations with differing populations within 6 km, the bottom plot compares two average temperatures over the last 70 years: that of all stations, and only those “rural” stations having populations less than 30,000 people in the year 2000.[32]

About 84% of the weather stations were thus classified as rural. Essentially no difference was found between the two averages, but there is a temperature rise of about 1C beginning in about 1970 in both averages.

The nationwide (and presumably global) temperature rise over the last few decades, clearly present in Figure 2.7, isn’t because of measuring increasing city temperatures as cities grew and thinking that increase was a global phenomenon. Increasing global temperatures are a global phenomenon.


[28]Peterson (2003) argues that the UHI is due mostly to measurement biases, but recently concedes very small UHI temperature increases (Peterson and Owen 2005).

[29]Tom Peterson suspects, but hasn’t shown yet, that the difference between thermal satellite images and weather stations is that weather stations are more likely to be placed in parks than in parking lots, meaning the difference between stations and satellites is revealing underlying variation in temperatures.

[30]UHIs were examined between the west and east coasts of the United States by Sun et al. (2006).

[31]Parker (2006) provides a nice summary of the literature pursuing this urban heating possibility for the global warming trends, considering calm and windy days as a test. There appears to be very little evidence to support the UHI-induced alternative idea. If true, calm days should have large UHIs as urban air temperatures rise, but on windy days, just as with a car’s radiator, cool rural air should blow away warmed city air, leaving small UHIs. Looking at data for those days separately demonstrated increasing trends in both cases, on a global scale, similar to the pictured L.A. trend. The results falsify the idea that the UHI plays a major role in the measurements of global warming, because data taken in the absence of an UHI still shows the global trend.

[32]For an older study of paired urban and rural stations that concluded urban heat islands exist, see Karl et al. (1988).