Abstract: Using the daily snow depth observations at 45 meteorological stations in Jilin province from 1961 to 2016,the spatial and temporal characteristics of snow cover increment and the number of days were analyzed based on the empirical orthogonal function (EOF),ensemble empirical mode decomposition (EEMD) and linear fitting methods.The results show that there are significant differences in the spatial distributions of the snow cover increment and the number of days.The higher values of them mostly present around Changbai Mountain located in the eastern part of Jilin province.The lower ones present in the mid-west plain of Jilin province.The amount of snow cover tightly connects with the geographical location of observations.In the higher altitudes and lower latitudes regions,more of snow cover is observed;while less is observed in the lower altitudes and higher latitudes.From the 1960s to the end of 1980s,the amount of snow cover increases with time,but it decreases in the 1990s.After 2000,it increases rapidly.Three typical spatial patterns of snow cover increment and days are found.The first pattern is a consistently more or less than the normal one over the whole region.The second pattern is a reverse phase one be-tween the eastern and western parts,and the third pattern is a seesaw one between the central regions and other regions.Both of the traditional linear fitting and the EEMD non-linear trend methods illustrate an increasing tendency in the amounts snow cover increment and the number of the days,but the former amplifies the increasing rates.

Key words: Snow cover increment, Snow cover days, Empirical orthogonal function (EOF), Ensemble empirical mode decomposition (EEMD), Linear fitting method, Spatial pattern