Abstract

Daily snow data during 1961–2013 at the 105 meteorological stations in Xinjiang, China were used to investigate the spatiotemporal variations of several parameters, including starting and ending dates, duration, annual and monthly average and maximum snow depths. The modified Mann–Kendall test, empirical mode decomposition, empirical orthogonal function (EOF), and the inverse distance weight interpolation were applied. Snow lasted for 71 to 120 days. Snow depth decreased from north to south. Daily snow depth had periodical variations and were classified as four typical types, i.e., flat peak, multi-peak, sharp single-peak, and right-skewed. After daily snow depth was decomposed into 17 intrinsic mode functions (IMFs), IMF9, IMF10, and IMF11 over 189, 302, and 437 days of scales accounted for 79% of the total spatiotemporal variance in snow depth. Both annual starting and ending day numbers had decreasing trends, while the duration in days had an increasing trend. The average and maximum snow depth increased in most sites whether considering the seasonality in December, January, February, or annual values. EOF1 accounted for 70% of spatial variability and the temporal coefficient EC1 varied periodically. The spatiotemporal analysis of snow properties provides a basis for snowmelt understanding.

You do not currently have access to this content.