Abstract:

Based on daily low temperature data which was below －20 ℃ from December to February during 1961 －2012 at 7 meteoro-
logical stations in Altay region of Xinjiang，the spatial － temporal variations of low temperature days were analyzed by using linear re-
gression，empirical orthogonal function，Mann － Kendall mutation detection and R/S analysis． The result is as follows:The monthly low
temperature days in winter decreased gradually from southeast to northwest with the maximum in January while the minimum in Februar-
y． The average winter low temperature days of Qinghe and Fuyun county in the eastern Altay were the most，which were 75． 0 d and
65． 9 d，respectively，while it was the least in Jimunai in the west of Altay，which was only 21． 6 d． The spatial distribution of extreme
value of winter low temperature days was in accordance with the average value，but the maximum mostly appeared in the middle 1960s
to the middle 1970s，while the minimum occurred consistently in 2006． Among them，the maximum and minimum value ranged from
64 d to 86 d and from 0 d to 46 d，respectively． The climatic tendency rate of winter low temperature days at all stations was from －
1． 71 to －4． 24 d/10 a in recent 51 years，which showed a trend of decrease． Though the monthly tendency rate of all stations was also
negative except for Jimunai in January，the value was smaller than that in winter． The first characteristic vector field displayed that win-
ter low temperature days existed good consistency on the space with the center of high value 0． 455 in Altay and Habahe． There was no
mutation in Habahe，Jimunai and Buerjin in the west and the central of Altay，while for Fuhai，Fuyun and Qinghe in the southern and
eastern Altay a mutation with downward trend occurred in the 1980s． H index of low temperature days in winter at all stations from R/
S analysis was more than 1/2，which meant that the trend of winter low temperature days in future would still decrease compared with
that in the past．

Key words: low － temperature days, climatic characteristics, winter, Altay  

摘要：

利用 1961 ～2012 年新疆阿勒泰地区 7 站 1 月、2 月和 12 月及冬季日最低气温低于 －20 ℃的日数资料，采用线性回归、经验正交函数分解、Mann － Kendall 突变检测、R/S 分析等方法，对低温日数的时空变化进行了分析。结果表明:(1)冬季及各月的低温日数具有自东南向西北递减的空间分布特征，且 1 月出现的低温日数最多，12 月出现的最少;(2)冬季低温日数的极值空间分布特征与平均值的一致。其中，年极大值出现的时间集中在 1960 年代中期至 1970 年代中期之间，年份差异较大，而年极小值均一致的出现在 2006 年;(3)各站冬季平均低温日数在近 51 a 内的倾向率为 －1． 71 ～ －4． 24 d/10 a，呈减少趋势，而各月的倾向率(吉木乃站 1 月除外)同冬季的一致，均为负值，但变化率均小于冬季的;(4)第 1 特征向量场的分布显示，冬季低温日数在空间上具有较好的一致性，高值区位于阿勒泰市、哈巴河县等高发区，中心值为 0． 455;(5)西部的哈巴河、吉木乃和布尔津 3 县及中部的阿勒泰市，均没有发生突变，而南部的福海和东部的富蕴和青河 2 县均在 1980 年代有下降趋势的突变发生;(6)各站冬季及各月的低温日数的 R/S 分析的 H 指数显示，未来的趋势与过去一致，仍呈减少趋势。

关键词: 低温日数, 气候特征, 冬季, 阿勒泰地区