【作者】 沈志超；

【导师】 任国玉；

【作者基本信息】 南京信息工程大学 ， 气候系统与全球变化， 2012， 硕士

【摘要】 利用中国东北地区90站1957—2010年逐日气温资料、74项环流指数和NCAR/NCEP再分析资料等,分析了东北冬季平均气温与极端气温事件的变化及受城市化的影响,与主要环流指数的关系等,探讨造成这种变化的局地人类活动影响和环流影响因子,并选取关键因子建立了预测模型。主要结论如下：(1)1957—2010年东北冬季增温明显,平均上升速率达到0.45℃／10a,北部增温更快。冬季增温主要是从20世纪80年代中期开始的,1980年代中期以前,气温在较小的范围内上下波动,而从80年代中期开始,气温呈不断上升趋势。冷冬年集中发生在20世纪60-70年代,暖冬年集中发生在90年代以后,且暖冬年发生强度和频率有显著增加趋势,冷冬年发生强度和频率则显著减小。(2)根据冬季平均气温距平REOF分解的前3个载荷向量的空间分布,将东北地区划分为分别以辽宁西南和黑龙江北部为中心的两个气候区。东北南、北两区的冷、暖冬年出现的时间有较大的不同,在1957~2010年间两者共同的冷冬年只有3年,暖冬年有7年。(3)采用6个极端气温指数,分析逐站极端气温指数时间序列及线性趋势,并对趋势的统计显著性进行检验,结果表明：东北地区暖夜日数、暖昼日数明显增加；冷夜日数、冷昼日数、冷日持续指数均呈显著减少趋势；日较差在逐年减小,北区变化更大。其中冷夜日数减少最明显,北区以3.3d／10a的速率下降,南区减少速率为1.9d／10a,区域平均以2.4d／10a的速率下降。各站及区域平均冷夜日数的变化趋势均通过0.001水平的显著性检验。(4)通过比较分析东北各站与乡村站的气温变化趋势表明：城市化对各种极端气温指数序列的趋势变化具有明显的影响。城市化对冷指数的影响比暖指数的影响更大。城市化对冷夜日数、冷昼日数减少的贡献率为35.8%和20.6%,暖昼日数增加的城市化影响贡献率为14.5%。(5)与东北冬季气温同期显著相关的环流因子是欧亚纬向环流指数,特别是南部相关更好。用前一年逐月的各种环流指数,分别计算了持续时间为1个月、2个月、3个月30种月季组合,共2220个备选因子数。选取超过0.001显著性水平的相关因子,得到了前期5个副高指数与2个极涡指数,相关最好的是东北北部。(6)考虑单因子显著水平和部分相关,使用向后去除变量选择方法,得到3个最优预测因子,分别是：8月东太平洋副高面积指数(175W—115W),10月亚洲区极涡面积指数(1区,60E-150E),8月北半球极涡面积指数(5区,0—360),建立了“最优”回归方程,检验表明模型有一定的预报技巧。更多还原

【Abstract】 Using a dataset of daily maximum and minimum temperature of1957-2010from90stations in Northeast China,74circulation indices provided by the National Climate Center of China, NCAR/NCEP reanalysis data, I analyze the variation of the Northeast China winter mean temperature and extreme temperature events and the effects of urbanization on the long-term trends of the extreme temperature indices, with an objective to reveal the temporal and spatial variation of winter temperatures in the region, and the in-phase and out-of-phase relationships between the temperature series and the circulation indices. I also develop a statistical prediction model for winter temperature anomalies by applying a few of key factors. The main conclusions of the paper are as follows:(1) Northeast China winter undergoes a significant warming, with an average rate of0.45℃/10a, and a more rapid rate since the mid-1980s. Winter mean temperature fluctuated without obvious linear trend before1980s, but it showed a significant rising trend since the mid-early1980s. Cold winters concentrated in the1960s and1970s, and the warm winters occurred more frequently in the1990s, with the intensity of warm winters significantly increased and the intensity of cold winters significantly reduced.(2) Using the method of REOF decomposition, The Northeast China region is divided into two sub-regions, with the centers located in southwestern Liaoning and northern Heilongjiang respectively. In the two sub-regions, cold and warm winters see a quite different occurrence. There were only3common cold winters and7common warm winters between the two sub-regions during the period1957-2010,(3) I analyze the characteristics of change of the6extreme temperature indices, and I find that warm nights and warm days all significantly increase, cold nights, cold days and continuous cold days all significantly decrease, and diurnal temperature range (DTR) gradually decrease in Northeast China. The decrease of cold nights is the largest, with the rates of-3.335days/10a in the North, and-1.85days/10a in the South, and the regional average rate of-2.431days/10a for the whole region. The decreasing trend of the regional average cold nights is significant at the0.001confidence level.Urbanization effects on the extreme temperatures indices series are investigated by comparing the regional average trends and the rural stations, and it is found that the effects are mostly significant, especially for the cold indices. (4) Northeast China winter temperature is significantly related to a few of circulation factors over the same period. The correlations with the Eurasian zonal circulation index are relatively good in the south. The winter mean temperature is also related to the previous abnormal circulation factors, including the previous January to November subtropical high and polar vortex area index. It is found that significant correlations exist for5Subtropical High Area indices and2Polar Vortex Area indices, with their relationship with winter temperature being better in the north of the region.(5) Cconsidering the confidence levels of the single-factor correlations and partial correlations and using the method of the backward selection, the3best predictors are obtained. They are the August eastern Pacific Subtropical High index (175W-115W) in August, the Asian Polar Vortex Area index (Area One,60E-150E) in October, and the Northern Hemisphere Polar Vortex Area index (Area Five,0-360) in August. A "best" regression equation is established, and it is verified, showing a good predicting skill for Northeast China winter temperature anomalies. It is also shown that the key predictors, the eastern Pacific Subtropical High and the Polar Vortex, are not independently affect the Northeast China temperature. Rather, they are interrelated and synergistically affect the variability of winter temperature of the study region. The underlying mechanism needs to be further investigated.更多还原