Declining U.S. regional and continental trends in intra‐annual and interannual extreme temperature swings

ABSTRACT The occurrence of regional temperature extremes on weekly to seasonal time scales has been a common climate impact in recent decades. Both instances of extreme warmth and extreme cold have been documented and analysed in the literature. While these events have most often been analysed independently, in this study, the transition between temperature extremes is examined using station data. Five measures of extreme temperature change are examined. At stations across the United States, there has been a significant decrease in the temperature difference between the warmest and coldest percentile observed within each year based on 7‐, 30‐ and 90‐day temperature averages during both the 1900–2017 and 1950–2017 periods. The maximum difference between percentiles associated with adjacent 7‐, 30‐ and 90‐day periods in each year has also declined significantly. At the same time, the interval between the highest and lowest annual percentile occurrence has lengthened. On a decadal basis, the frequency of shifts from the sub‐5th to over‐95th temperature percentiles has also declined through time, while the average time period between temperature occurrences in opposite tails of the distribution has increased. In general, these results are very consistent across the United States, although some regional‐ and duration‐dependent differences are noted. For many of the extreme temperature metrics, a high level of field significance is obtained in the Southwest, Great Plains and Midwest regions. There has been a significant decrease in the difference between the warmest and coldest percentile observed within each year across the United States. At the same time, the interval between the highest and lowest annual percentile occurrence has lengthened. On a decadal basis, the frequency of shifts from the sub‐fifth to over‐95th temperature percentiles has also declined through time, while the average time period between temperature occurrences in opposite tails of the distribution has increased.