A Reevaluation of United States Temperature Trends from 1935 to 1981

Southeastern Geographer Vol. 25, No. 1, May 1985, pp. 51-62 A REEVALUATION OF UNITED STATES TEMPERATURE TRENDS FROM 1935 to 1981 Brent R. Skeeter INTRODUCTION. Numerous studies of global and hemispheric temperature trends have been undertaken in recent decades. One of the earliest studies was done by Willett in 1950. (1) Willett analyzed temperature records of fifty years or longer for 129 stations and additional records of between twenty and fifty years for 54 stations. He found that global temperatures increased 1°F (0.6°C) between 1885 and 1940. Winter temperatures during this same period were found to increase by 2.2°F (1.2°C). The most pronounced warming was in the high latitudes ofthe Northern Hemisphere. This warming trend conformed with the temperature trend that climatologists had expected to exist due to the increasing levels of carbon dioxide in the atmosphere. However, even though the amount of carbon dioxide in the atmosphere has continued to rise, cooling temperatures have been reported since about 1940. (2) The first study to report this reversal in temperature trends was done by Mitchell in 1961. (3) Mitchell updated Willett's work through the 1950s and found that temperatures had fallen 0.20C by the late 1950s from a peak in the early 1940s. In 1970, Mitchell stated that by the late 1960s global temperatures had fallen 0.3°C from the peak in the 1940s, approximately one-half of the prior rise. (4) Similarly, Budyko reported that temperatures in the Northern Hemisphere fell 0.3-0.40C between 1940 and 1976. (5) Summaries by Schneider and Dickenson, Kalnicky, Robuck, Roberts, and Agee all report Northern Hemisphere temperature declines of at least 0.50C since the 1940s. (6) In summary, Gribbin states "In worldwide terms, we are in a situation where the earth is cooling more quickly than it warmed up earlier this century." (7) From the above it is clear that the general concensus in the recent literature is that there has been a cooling in the Northern Hemisphere since the early 1940s (Fig. 1). Despite this general concensus, there is * The author gratefully acknowledges the guidance and support of Dr. Albert J. Parker. Mr. Skeeter is a Ph.D. student in the Department of Geography at the University of Nebraska in Lincoln, NE 68588. 52SoUTHEASTERN GEOGRAPHER reason to reevaluate the timing and spatial expression of the reported temperature trends. Mitchell reported on the various sources of error in instrumentally observed secular temperature trends, listing such factors as city growth, instrument relocations, instrument aging, and building construction. (8) Among these sources of inaccuracy in reported temperature readings, instrument relocations have the potential of introducing considerable artificial decline in reported temperatures. This is because during the period of the reported temperature declines many temperature recording stations moved from urban locations (where they are subject to urban heat island effects) to suburban airport sites which are normally cooler. (9) The Air Commerce Act of 1926 gave the Weather Bureau responsibility for providing meteorological support for the young commercial aviation industry, initiating thereafter an unprecedented migration of weather stations from downtown locations to suburban airport sites. Ott reported that 88.6 percent of these stations experienced significant thermometer relocations between 1935 and 1974, that 77.8 percent of these relocations were from city to airport, and that 84.7 percent of the city to airport relocations occurred between 1935 and 1950. (10) In 1940 alone, twenty-two stations moved to airport sites. This urban to suburban shift of temperature recording stations may make a significant artificial contribution to the reported cooling of the Northern Hemisphere. Moreover, the spatial patterning and timing of the reported cooling could be different than previously thought due to the greater number ofstation relocations in certain years, particularly the 1940s. In order to document temperature patterns in the coterminous United States during the last fifty years, while minimizing the influence of thermometer relocation from urban heat islands, data for the period from 1935 to 1981 from 150 rural sites across the coterminous United States were collected. Rural sites were chosen because thermometer relocations within rural areas are less likely to cause artificial declines in recorded temperatures. While there may...