Cities grow, and as they do, they heat up. The so-called Urban Heat Island effect is a big problem for measuring climate change because it contaminates the data and exaggerates the trend. And it's made worse by the fact that, historically, all our data comes from places people live, and the heat island starts building quickly as population grows. Every so often the alarmists try to wave the problem away and insist it doesn't matter, but the reality is otherwise. For example, a new study re-examining long term temperature data from China shows that half the measured warming since the 1940s is likely due to uncorrected urbanization bias.
The study points out the well-known fact that as cities grow, the daytime highs don't rise as much due to urbanization as the nighttime lows, which get biased upward because all the heat-absorbing surfaces slowly release their heat during the night. The daily mean is based on the average of the high and low, so if the low is trending up due to urbanization it pulls up the average too. But greenhouse warming should affect the daytime highs as well as the nighttime lows. So a good way to check what's driving the changes is to compare trends in the highs and lows.
The authors put together a detailed sample of Chinese cities with data back to the 1940s to do just that. They found not only that the nighttime lows were going up much faster than the daytime highs, but that the difference in trends was strongest in the more densely populated areas, pretty much confirming the cause. Comparing the observed trends to what would be expected if urbanization wasn't a factor, they concluded that about 50 percent of the warming in China since the 1940s is uncorrected urbanization bias, not climate change.
That's a pretty big contamination rate of the data. And others have found evidence of such contamination of the global surface temperature data as well. It's not surprising, in fact it would be very surprising if it weren't occurring. And even with urbanization biases, as our tour through weatherstats.ca shows, there has been remarkably little measured climate change in Canadian cities for the past half century or more. Now cut it in half, and you're probably getting closer to the actual rate of change.