【摘要】 利用1961-2005年45 a中国日平均温度的站点资料和累积频率的统计方法,确定了以90%(10%)累积频率为标准的日、旬、月和季四种不同时间尺度的极端高(低)温事件的阈值。分析表明,日尺度阈值(简称"日阈值")用于检测极端高(低)温事件最为合理。但在日阈值的季节变化中包含系统性的天气扰动,因此需对其进行低通滤波,滤去8日以下的波动,最终可得到较为合理的检测极端温度事件的日阈值。对用日阈值检测出的全国极端温度事件,定义频数(日平均温度等于或超过高阈值的次数)和平均强度(极端温度事件的日平均温度总和与频数的比值)两个参数进行分析,发现频数的线性趋势表现为:高温事件(大于等于高阈值)中,内蒙古和新疆北部均为大趋势区,而西南地区为小趋势区,低温事件(小于等于低阈值)则相反,趋势显著的地区为高温事件中的东北、华北和青藏高原一带。平均强度的线性趋势表现为:高温事件中,我国东北地区为负趋势的大值区,云南西北部和海南省大部分地区为正趋势的大值区。低温事件则相反,趋势显著区则为高、低温事件中25°N的以南地区。更多还原

【Abstract】 By use of the 1961-2005 daily mean temperature station data over China and the cumulative frequency distribution(CFD),the extremely high(low) temperature thresholds on four different time scales,day,pentad,month and season,are defined by the 90th(10th) percentiles of the CFD.It is found that the threshold on day scale(day threshold for short) contains systematic weather turbulences and,therefore,is not smooth enough.For this reason,the low-pass filtration is used to elide the wave less than eight days.Finally,a more reasonable day threshold is obtained to define temperature extremes.The results show that,for the warm(equal to and greater than the high temperature threshold) extremes,the trends of extreme days increase more remarkably in the south of 30°N than in the north with the smaller trend in the west of Inner Mongolia and the greater trend in the south of Yunnan and Hainan Province.Contrarily,for the cold(equal to and less than the low temperature threshold) extremes,the obvious regions with extreme days are in Northeast and North China and Tibetan Plateau.Both cold and warm extremes show increasing trends in average intensity in Southwest China and decreasing trends in Northeast China.The regions located in the south of 25°N are the obvious areas both in warm and cold extremes.更多还原