C32A-02:
Creating a United States Snowfall Regionalization Based on Frequency
Wednesday, 17 December 2014: 10:35 AM
Daria Kluver, Central Michigan University, Mount Pleasant, MI, United States
Abstract:
A snowfall regionalization is developed for the conterminous United States using variability of snowfall frequency. Principal components analysis and cluster analysis are used to group stations together over the time period 1930 to 2007. Seven snowfall regions are identified, which correspond to predominant storm tracks across the United States. These are: the southeast, the south central Plains and southwest, the Ohio River Valley and mid-Atlantic, the Pacific Northwest, and three sub-regions in the Upper-Midwest. Quantile regression is used to identify how the frequency distributions at each region change over the period of record. The Pacific Northwest and southeastern regions are experiencing significant declines in median and large frequency winters. The Northern part of the upper-Midwest region is experiencing significant increasing trends in all percentiles of snowfall frequencies.