A rare long record of deep soil temperatures defines temporal temperature changes and an urban heat island

A long-term set of deep soil temperature data collected over a 64-year period beginning in 1889 in a rural Illinois area provide a rare opportunity to assess the natural shifts in temperatures in a pristine environment without any urban or instrument bias. Temperatures from 1901 to 1951 increased 0.4 degrees C, and this was 0.2 degrees C less than nearby values from two high quality surface temperature data sets that supposedly are without any influence of urban heat islands, shifts in station locations or instrumentation, or other changes with time. Comparison of the soil values with surface air temperatures from a nearby weather station in a growing university community revealed a heat island effect of 0.6 degrees C. This value is larger than the adjustment based on population that has been recommended to eliminate the urban bias in long-term temperature trends in the U.S. Collectively, the results suggest that additional efforts may be needed to eliminate the urban influence on air temperatures, beyond techniques that simply use population as the basis. Population is only an approximation of urban factors affecting surface temperatures, and the heat island influences inherent in the values from weather stations in smaller communities which have been used as control, or data assumed to be unaffected by their urban environment in the adjustment procedures, have not been adequately accounted for.