A new study by Roy Spencer focuses on the size of urbanization biases in temperature data from Canadian cities. Spencer and his colleague John Christy are using satellite measurements of land use change (“Landsat”) in various countries to measure how much of the observed warming in recent decades is due to increased urbanization. The usual way of doing so is to compare adjacent rural and urban stations. But Spencer notes that this is not very accurate if most of your data comes from urban locations to begin with so the rural sample is too small to give robust, reliable readings. So he and Christy are building a data set in which they look at temperature averages over recent decades, which coincides with Landsat surveys of land cover characteristics, to see how closely changing land cover correlates with changing temperature. Spencer presented data on Canadian cities with large airports, and then constructed detailed temperature histories with and without effects of urbanization. And it turns out, in what should be to no one’s surprise except perhaps the experts, that urbanization has played a big role in the apparent warming of Canadian cities.
Spencer notes that previous studies had shown urbanization tends to have a larger effect in cities at night, when the heat absorbed during the day is released and gets trapped by the stable air layer that forms overnight as the winds die down. Spencer and Christy find that for every 10 percent increase in urbanization in a region, nighttime temperatures increase by 0.35°C, which is quite the trend. Daytime highs increase by less, only 0.21°C. But suppose that the area covered by urbanization in a region increases by 50 percent over a few decades. According to their findings, the city will warm by about 1.8 degrees at night (because 0.35 x 5 = 1.75) and about 1 degree during the day (0.21 x 5). Which in many places is a big enough trend to explain all or most of the warming, especially since the total warming since 1850 according to climate orthodoxy is just over 1°C.
Then Spencer looks at 10 urban airports across Canada where they were able to get urbanization data over the entire 1978 to 2022 interval. Measured nighttime warming ranged from -0.11°C/decade in Regina, Saskatchewan to +0.48°C/decade in Abbotsford, British Columbia. But on average the urbanization effect explained half of it, so the 10-station average warming fell from +0.19°C/decade to a de-urbanized rate of +0.10°C/decade. Daytime highs warmed, on average, by +0.28°C/decade, but only +0.06°C/decade of this was attributed to urbanization, leaving a de-urbanized trend of +0.22°C/decade.
Spencer also looks at the data from the Climate Research Unit at the University of East Anglia, of Climategate emails fame, whose data is relied on by the IPCC. He finds that their estimated trend for the region encompassing Calgary and Edmonton, which they claim has been scrubbed clean of urbanization bias, matches the data he collected but only from the pre-de-urbanization step. So the data being used by the IPCC is still contaminated with an upward warming bias due to urban heat islands.
In my community our weather station is in the middle of a farm field. On weatherstats.ca there is no trend up or down over 50 years, yet the Enviro-nanny site has our area rising in annual temperature. When I ask them why they tell me the models are more accurate than the raw data, but don't provide any reasoning as to why.
Models more accurate than raw data? WTF?