An appreciable number of nonurban stations in the United States and Canada have been identified with statistically significant (at the 90% level) decreasing trends in the monthly mean diurnal temperature range between 1941–80. The percentage of stations in the network showing the decrease is higher than expected due to chance throughout the year, with a maximum reached during late summer and early autumn and a minimum in December. Monte Carlo tests indicate that during five months the field significance of the decreasing range is above the 99% level, and in 12 months above the 95% level. There is a negligible probability that such a result is due to chance. In contrast, trends of increasing or decreasing monthly mean maximum or minimum temperatures have at most only two months with field significance at or above the 90% level. This is related to the tendency toward increasing temperature in the western portions of North America and decreasing temperature in the east.
The physical mechanism responsible for the observed decrease in the diurnal range is not known. Possible explanations include greenhouse effects such as changes in cloudiness, aerosol loading, atmospheric water vapor content, or carbon dioxide. Change in circulation is also a possibility, but it will be difficult to isolate since the patterns of the decreased diurnal temperature range have high field significance throughout much of the year, relatively low spatial coherence, and occur at many stations where individual trends in the maximum and minimum temperature are not statistically significant. Our data show that the trends in the maximum and minimum temperatures may differ considerably from trends in the mean.