Evaluating changes in season length, onset, and end dates across the United States (1948–2012)

Discussed in a global context, the issue of scale is important to consider as it relates to climate change. As climatological studies utilize temporally and spatially static definitions of seasonal definitions, the understanding of seasonal change may be limited by the way in which seasons are defined. This research serves as a challenge to this perspective by examining the changes in seasons by incorporating variability into the way in which seasons are defined. Temporally and spatially variable seasons were defined based on thresholds of apparent temperature and upper‐level circulation patterns for the period 1948–2012. Using 60 US surface stations, two apparent air temperature metrics were used to define seasons that vary spatially and temporally. Stations represented major cities found in the contiguous United States. A large‐scale circulation approach utilized synoptic categorizations based on 250‐mb geopotential heights from the NCEP‐NCAR reanalysis ( NNR ) as a metric of seasonal delineation. In this method, seasons varied temporally but not spatially. Despite the different methodological approaches, consistent trends were found. Since 1948, late starts of autumn and winter have been observed while earlier onsets of spring and summer have taken place. For the individual stations, the largest shifts occurred along coasts and in larger, more urbanized environments. Individual locations also showed increased variability in start date, and significant changes were found for all four seasons in the circulation approach. Seasons have been shown to be important for a variety of processes including phenological responses and human adaptation to extreme temperature environments; therefore, the consideration of season variability may be appropriate for future climatological research.