Recorded presentation15A.3
Trends in extreme snowfall years in the contiguous U.S
Kenneth E. Kunkel, ISWS, Champaign, IL; and M. A. Palecki, L. Ensor, D. A. Robinson, K. G. Hubbard, D. R. Easterling, and K. T. Redmond
A quality assessment has been undertaken of all U.S. long-term snowfall stations in which the authors have made an expert judgment on the quality of each station. Through this process, we have identified a set of stations we believe to be suitable for analysis of trends. We have analyzed snowfall variations back to 1900. Extreme snowfall years were identified for each station as years with total snowfall less than the 10th percentile threshold or greater than the 90th percentile threshold; these thresholds were calculated from the station's entire period of record. The percentages of stations with high or low extremes were determined for each year to create two time series, one for the coverage of high snowfall extremes and the other for the coverage of low snowfall extremes. Prior to about 1920, the percentage of stations with high snowfall extremes was quite high. This was followed by a period of generally low percentages in the 1920s, 1930s, and early 1940s. The period from the late 1940s to about 1980 was characterized by moderately high percentages. The period since the mid 1980s has seen a number of years with very low percentages. The percentage of stations with extreme low snowfall exhibits behavior approximately, but not exactly, inverse to the extreme high snowfall time series. There are high percentages in the 1920s and 1930s, low percentages from the 1940s to about 1980, and moderate values from the 1980s onward, approximately opposite to the results for extreme high snowfall. However, there were high percentages of extreme low snowfall years prior to 1920, when there also were high percentages of high snowfall years. A substantial portion of the interannual variability is related to temperature variability. There is a negative correlation between the percentage of stations with high snowfall and winter (December-February) temperature and a positive correlation between percentage of extreme low snowfall years and winter temperature. Winter temperature explains 34% of the variance for extreme high snowfall and 18% for extreme low snowfall. Recorded presentation
Session 15A, Detection and attribution of climate change: Part IV
Thursday, 24 January 2008, 1:30 PM-3:00 PM, 215-216
Previous paper Next paper