气候系统与气候变化
Climate System and Climate Change

气候系统与气候变化研究进展

2012年，气候系统研究所以“东亚季风变异机理及其对中国气候影响”研究为核心，围绕东亚季风变异机理和预测、海-陆-气相互作用对东亚气候影响、青藏高原热力和动力过程及其影响、气候变化机理与预测理论4个主要方向开展研究，在东亚季风长期变化机理等方面取得了创新性成果。

1 东亚季风变异机理和预测研究

研究了中国东部地区盛夏降水与温度变化的关系，大气季节内振荡、ENSO和AO不同位相配置对中国气候灾害的影响，东亚冬季风预报因子的确定，并对NCEP气候预报系统（CFS）模式进行了评估。

1.1 中国中东部地区降水特征的年代际变化

以降水的强度结构为基础，通过e指数拟合和确定e折倍点对我国东部地区盛夏逐时降水进行分类，利用e折倍强度划分出弱降水和强降水，超过2倍e折倍强度的降水被视为极端降水。发现在过去几十年间，弱降水和强降水比例发生了显著变化，且弱降水百分比变化的空间型与气温变化的空间型相似。根据各站点的气温变化特征，划分了华北地区（增温）、中东部地区（冷却）和东南沿海地区（增温）。在增温区，弱降水比例明显减少；而冷却区弱降水比例呈增加趋势。各地区弱（强）降水比例与气温变化均存在负（正）相关，而极端降水与气温的关系却具有显著的区域差异（图1）。

1.2 MJO、ENSO和AO对中国极端干旱气候影响

研究发现，大气季节内振荡（MJO）和北极涛动（AO）的持续性异常对东亚地区旱涝有重要影响。2009—2010年我国云南发生的秋、冬、春3季连旱与不活跃的MJO和弱AO的长期维持密切相关；持续的MJO不活跃也是2011年云南主汛期期间夏季极端干旱的重要成因之一；秋冬季AO的负位相有利于乌拉尔山阻塞高压维持和发展，而贝加尔湖上空出现负位势高度异常，导致东亚中高纬度经向环流加强和冷空气向南侵袭，进而间接导致华北地区的干冷气候。

1.3 东亚冬季风预报因子

研究发现，前期秋季中纬度东太平洋区的海温异常、北极喀拉海海冰密集度和东亚高空气温具有很好的持续性，可以将异常信号从秋季传递到冬季，进而影响冬季风的强度。利用上述研究成果建立了东亚冬季风预报模型，并应用到2012/2013年冬季风预报中。

1.4 气候预报系统（CFS）模式结果评估

通过对多组NCEP 气候预报系统（CFS）60天后报结果的分析，比较了3种分辨率（T62，T126 和T254）的大气模式对NCEP CFS 进行季风季节进程预测的影响。结果表明，对于陆地上的夏季风建立及其推进，高分辨率模式整体表现比低分辨率模式好，且在南亚夏季风区比东亚夏季风区更明显。高分辨率模式最大的改进是更加真实地再现了大地形附近的降水分布，特别是在青藏高原附近。然而，提高大气模式分辨率不能有效提高海洋上空季风活动的预测效果，随着大气模式的分辨率从T62提升至T254，模拟的季风降水在南海和菲律宾海反而更差。（李建）

2 海-陆-气相互作用对东亚气候影响研究

揭示了北极海冰持续减少的动力成因，研究了欧亚大陆积雪对东亚夏季风异常的影响、两类ENSO与太平洋次表层海温的关系，以及北太平洋副热带和副极地流涡的变化等。

2.1 北极地区地表风变化模态研究

利用1979—2010年NCEP/NCAR月平均地表风，采用复向量经验正交函数（CVEOF）分析了北极地区地表风变化的2个主要模态。第1个空间模态包含2个子模态，分别为拉普帖夫海（NLS）模态和北极偶极子（AD）模态；第2个模态包括北极喀拉海（NKS）模态和北极中部（CA）模态。在过去的20年里，这2个模态的联合动力强迫是北极地区9月海冰面积不断缩小并降至记录最低值的重要原因。9月海冰面积最低值主要与夏季（7—9月）AD模态的负位相和CA模态的正位相有关，这2个位相都显示出北冰洋上为一个异常的反气旋控制。风场通过频率和强度影响9月海冰面积。过去20年9月海冰面积的下降趋势与海冰融化季节（4—9月）CA模态的增多增强趋势有关。因此，不能简单地认为是AD模态异常的结果。CA模态在1990年代后期表现出年代际变化特征，1997年以前北冰洋上为异常气旋环流而以后被异常反气旋所代替，与9月海冰面积下降的趋势一致。

2.2 欧亚大陆积雪对华南降水影响数值模拟

基于欧亚大陆积雪深度资料和中国台站降水资料的奇异值分析（SVD）结果，使用大气环流模式（CAM3.1）分别进行3组集合试验，来研究欧亚大陆积雪的反照率效应和水文效应对2010年5—6月华南降水的影响。第1组试验综合考虑积雪的2种物理效应，既有反照率效应又有水文效应；第2组试验仅考虑积雪反照率效应，忽略水文效应；第3组试验只考虑积雪水文效应，忽略反照率效应。试验结果表明，积雪的2种物理效应都会对后期华南降水产生影响，但是3组试验中积雪不同物理效应所导致的异常幅度和范围存在较大差异，其中积雪水文效应比反照率效应引起的变化幅度大。当2种效应共同作用时产生的异常与统计分析结果最接近，变化幅度也最大，但是并不等于单纯反照率效应和单纯水文效应作用之和。

2.3 春季欧亚积雪与东亚夏季风关联

研究了春季欧亚大陆融雪量与东亚夏季风的关系，并初步讨论了其可能联系机制。研究表明，春季融雪量EOF第1模态表现出年代际变化特征，这与东亚夏季风和中国夏季降水的年代际转型具有非常好的一致性。而EOF 第2模态与东亚夏季风在年际尺度上具有同位相变化关系，当春季融雪量在东西伯利亚和巴尔喀什湖附近异常偏多时，后期在东亚地区容易出现由高纬度至低纬度的“负-正-负”经向波列结构。融雪量异常偏少时，情况则相反。初步分析了春季融雪量异常与后期夏季东亚地区大气环流出现经向波列结构的可能联系机制，东西伯利亚以及巴尔喀什湖附近异常偏多的春季融雪量能够在该地区促使位势高度场表现为正异常，随着时间的演变，巴尔喀什湖附近地区的高压向东移动发展，东西伯利亚地区的高压一部分向低纬移动，可能造成夏季东亚地区的经向波列结构，进而对东亚的天气和气候产生影响。

2.4 表层和次表层海温耦合模态和2类ENSO关系

对2类ENSO事件时空变化研究发现，东部和中部型ENSO分别表现出显著的年际尺度（27年）和年代际尺度（10～15年）变化特征。在热带太平洋海表温度（SST）和次表层温度（SOT）变化存在显著的耦合关系：第1个SVD模态对应的SST和SOT异常场上均表现出纬向偶极型分布，即热带中东太平洋海温异常增暖，热带西太平洋海温异常偏冷；第2个SVD模态对应的SST和SOT异常场上均呈现出长期线性趋势的变化；而第3个SVD模态所对应的SST和SOT异常场上均呈现出纬向3极型分布，即海温异常偏高出现于热带中太平洋，热带西太平洋和东太平洋海温异常偏冷。此外，20世纪80年代后SOT在年代际时间尺度上振幅明显增强这一事实说明，中部型ENSO事件的频率和强度均呈现出上升趋势，且在过去的30年里中部型ENSO事件明显加强了对全球气候的影响（图2）。

2.5 北太平洋副热带和副极地流涡年际变率

北太平洋副热带和副极地流涡应对风应力变化的调整会产生不同尺度变化，在气候变化中有重要作用。基于SODA（Simple Ocean Data Assimilation）和GODAS（Global Ocean Data Assimilation System）资料，利用三维海洋环流诊断方法分析了副热带和副极地流涡的环流强度和中心位置的变化，建立了副热带和副极地流涡的强度、经向位置和深度位置的3类指数序列。结果表明，所建指数能够很好地表达副热带和副极地流涡的季节、年际和年代际变化。在夏季副热带流涡强度要远小于冬季，副极地流涡冬季最强，秋季最弱。副热带流涡从2—3月开始北移，到10月左右开始南移，1月左右位置最靠南。而副极地流涡是春季北移，夏季南移，秋季又北移，到冬季南移到极值点。2个流涡的共同特征是1976—1977年前偏弱偏北，之后偏强加深偏南。对于年代际尺度的变化，副极地流涡的贡献极其重要。副极地流涡的强度指数显示流涡在1976—1977年存在明显的由弱到强的跃变过程。强度指数与太平洋年代际振荡（PDO）指数的相关系数为0.45，要远好于副热带环流的强度指数与PDO指数的相关。检验证明，受中小尺度涡动的影响，大尺度流涡强度量级对资料分辨率有很强的依赖性，但指数所表达的流涡年际和年代际大尺度变化受模式分辨率影响较小。（祝从文）

3 青藏高原热力、动力过程及其影响研究

利用卫星观测资料，重点讨论了青藏高原东南地区降水的日变化特征，以及高原低涡和中国南方地区云的日变化特征。

3.1 青藏高原东南地区降水日变化特征

青藏高原东南地区地形复杂，河谷纵横，其夏季降水表现为傍晚和午夜降水双峰值并存的特征。卫星降水数据显示出一致的午后降水峰值，而常规台站观测降水则为午夜峰值。基于色齐拉山加密观测试验数据的分析表明，高原上降水的日变化特征与地形密切相关，山坡的降水峰值常出现在傍晚，而山谷的降水峰值常出现在午夜。卫星反演降水的日变化特征与色齐拉山的山坡站一致；而常规台站观测降水日变化与山谷站一致。卫星与常规台站的降水的日变化差异与常规台站位于山谷中有关，位于山谷中的常规台站观测降水数据难以全面反映高原降水的区域特征。同时，卫星降水对夜雨的低估对二者的差异亦有贡献。上述分析表明，在分析复杂地形区的降水性质时，充分利用多种观测资料将有助于全面揭示区域的降水特征，研究结果也为高原地区台站的后续建设提供了参考（图3）。

3.2 高原低涡和中国南方地区云日变化特征

利用美国国家环境预报中心（NECP）提供的每日4次再分析资料（水平分辨率为1°×1°）、我国12 h一次的常规探空资料和逐时地面气候要素观测资料，以及中日合作JICA项目提供的大气可降水量（PWV）资料和改则站每日4次的加密探空资料，统计了2006—2008年5—9月青藏高原低涡在各时次的生成频数，并对各时段大尺度环流场、相关要素场和加热场进行了合成分析。研究结果表明，青藏高原低涡的生成具有日变化特征，其在傍晚到午夜的生成频数最高，在清晨到正午产生频数最低。高原低涡的产生是大尺度环流场和高原地区加热场共同作用的结果，降水的凝结潜热加热对高原低涡的产生有直接的促进作用。傍晚，高原中西部500 hPa气流有较强的辐合，200 hPa高空西风急流和南亚高压的强度较强，并且高原主体上水汽收入最多，层结较不稳定，有利于降水在此时发生。降水的凝结潜热加热有利于形成上暖下冷的热力结构，使500 hPa的气旋性扰动得到发展，导致大量低涡形成于傍晚后到午夜这个时间段，清晨到正午反之。

3.3 夏季亮温揭示的云变化特征

利用风云静止气象卫星红外亮温数据，分析了夏季亮温日变化的气候特征。根据云分类产品剔除晴空数据后，按照云顶亮温不同的日变化特征，将云分为3类，即冷云（云顶亮温低于-30 °C）、中层云（云顶亮温介于-30～0 °C）和暖云（云顶亮温高于0 °C），着重关注中国南方地区3类云发生频率的日变化特征及其与降水日变化区域差异的关系。分析发现，中国南方地区夏季3类云的日变化存在共性特征，即多数地区的冷云在傍晚发生最为频繁，在高原下游亦存在夜间的峰值。中层云发生频率峰值位相多在夜间，西南地区的峰值多出现在下半夜而东南地区多在上半夜，在高原东部存在傍晚峰值。暖云在中午前后发生更频繁，且区域差异相对较小，只在南部沿海地区存在夜间峰值。同时，不同云的日变化亦存在鲜明的区域差异，特别是冷云与中层云的发生频率在西南和东南均表现出不同的日变化特征。冷云在西南（东南）地区的清晨（傍晚）发生更为频繁。中层云发生频率在西南地区为午夜至凌晨的峰值，而在东南地区为上半夜的峰值。此外，冷云和中层云频率的日变化还存在明显的季节变化，而暖云频次日变化季节变化小。冷云频率的傍晚峰值几乎全年存在，在暖季主要为单峰结构，而在冷季存在夜间至凌晨的次峰值。中层云频率峰值在夏季出现在上半夜，而在其他季节多出现在午夜至凌晨。暖云峰值位相在全年均出现在中午前后，其季节变化主要体现在日变化振幅上，在暖季较小而其他季节较大。研究结果表明，充分利用高分辨率的静止卫星亮温有助于增进对区域云雨日变化特征的理解。（陈昊明）

4 气候变化机理与预测理论研究

结合观测资料和数值模拟，讨论了东亚季风年代际变化和过去千年变化的可能成因。

4.1 亚洲大陆地表温度变化对东亚季风的影响

采用观测资料和美国国家环境预报中心CFS模式输出的大气分量，揭示了近半个世纪以来在全球变暖背景下亚洲大陆地表增温幅度小于其他地区的特征，尽管亚洲表面温度增加，但是由于世界其他地方的增温更大，亚洲大陆相对而言仍然是一个热沉降区。与亚洲之外的区域不同，近10年整层积分的亚洲对流层温度比前几十年偏冷。由此提出了亚洲大陆的这种“相对变冷”可能是导致亚洲季风变弱的重要原因之一。

4.2 全球变暖对东亚夏季风的影响

利用观测资料和数值模拟，研究了东亚夏季风减弱的可能成因。发现东亚夏季风的持续性减弱与环贝加尔湖地区的变暖存在密切联系。该地区的变暖可以在东亚地区上空激发出对流层异常反气旋，从而抑制了夏季西南季风向北输送，导致亚洲北部干旱和南方降水偏多现象。采用CAM3数值模式考虑温室气体、海表温度、太阳辐照度和火山活动的模拟结果可以再现贝加尔湖变暖和相关的大气环流，然而不包含温室气体的模拟不能再现贝加尔湖地区1970年代的变暖。这意味着全球变暖可能导致了贝加尔湖地区的局地变暖，进而导致了东亚夏季风近几十年来的减弱。

4.3 中国夏季气候长期变化模式模拟研究

利用国家气候中心大气环流模式（BCC_AGCM），初步诊断分析并评估了1955—2000年东亚夏季风年代际变化的整体结构。东亚夏季风在观测海温强迫下，合理再现了降水、气温和环流的年代际变化，模式再现了中国降水南涝北旱的主要特征，进一步增加了对南涝北旱成因的认识。同样，对流层中上层变冷、高层西风急流南移和低层西南季风环流减弱以及他们和降水变化的关系也被模式再现。模式模拟的一个最主要的缺陷是对流层变冷和大尺度环流变化的幅度比再分析资料的小，他们和降水变化的关系亦是如此。还对模式中对流层中上层变冷的中心和上层西风急流轴西移减弱的现象和观测进行了比较。总体而言，尽管BCC_AGCM模式在模拟东亚气候的年代际变化存在一些问题，尤其是在变化中心和幅度上，但是该模式能够合理地再现观测中降水变化的配置及其相关的气温和环流变化。因此，该模式可以用来进一步探讨东亚年代际变化的机制。同时，利用SST强迫的数值模式合理再现降水变化的配置及其相关大气环流结果表明，东亚气候的年代际变化可能是对全球气候变化的区域响应。

4.4 千年气候变化数值模拟研究

采用中等复杂程度的UVic地球系统气候模式模拟了气候强迫因子(太阳辐射、火山灰、太阳轨道、陆表植被变化、温室气体和人为排放的硫酸盐气溶胶)对中国东部地区过去千年气候变化的贡献。结果表明，考虑所有气候强迫因子的数值试验可以较好地再现北半球和中国东部地区中世纪暖期、小冰期和20世纪暖期这3个特征时期，与重建的气温在百年尺度上也具有较好的一致性。模拟结果很好地反映了中国东部气温异常在中世纪暖期和20世纪上半叶比全球气温异常偏高，以及小冰期气温异常比全球偏低的事实。根据中国东部气温的冷暖程度和气候强迫因子的相对贡献大小，将过去千年中国东部地区气温分为8个阶段：中世纪暖期3个、小冰期4个和20世纪暖期1个，并揭示了气候强迫因子对这些子阶段维持及其转换的贡献。结果表明，中国东部地区中世纪暖期的主要贡献来自于太阳辐射，火山灰次之；小冰期各个子阶段的转换过程中，主要取决于温室气体、火山灰和太阳辐射的相对贡献大小。温室气体和火山灰分别为小冰期的最后2个阶段中最主要的贡献因子。本研究发现了不同自然气候强迫因子之间和不同人为气候强迫因子之间的非线性响应，太阳轨道变化和火山灰气溶胶强迫的非线性响应与温室气体和陆表植被(或者硫酸盐气溶胶)强迫的非线性响应对中国东部地区20世纪末的增温贡献都达到了约0.2 °C，而自然气候强迫因子和人为气候强迫因子之间不存在明显的非线性响应。自然气候强迫因子之间和人为气候强迫因子之间的非线性响应对中国东部地区20世纪变暖大约分别贡献了0.09 °C和0.18 °C，二者之和约占中国东部地区20世纪末增暖的50%（图4，红线减去灰线），其余的增温来自于气候强迫因子的线性贡献(即线性响应)。气候强迫因子之间的非线性响应为认识气候变化提供了一个新视角（图4）。（肖栋）

图1 不同强度降水百分比变化(1986—2005年减1966—1985年)(实线、虚线和点线分别代表华北、中东部和华南沿海地区，粗线为7点平滑曲线)Fig1 The 20-yr mean changes (1986—2005 minus 1966—1985) in the percentage of rainfall amount as a function of rainfall intensity. (The solid, dashed, and dotted lines are for North China (sienna), central eastern China (green), and southeastern coastal area of China (blue), respectively. Thick lines denote the seven point smoothing series)

图2 1958—2007年热带太平洋地区海表面温度(SST)(a)与赤道(5°S-5°N 平均)次表层海温(SOT)的SVD 第3模态(b)及相对应的时间系数分布(c)(绿色粗体曲线是气候平均的20 °C等温线, SST时间序列(蓝色实线)与SOT 时间序列(红色实线)的相关系数为0.76）Fig2 The third SVD mode of the SST (a) and the SOT anomalous felds (b) and their expansion coeffcients (c) during 1958—2007. The expansion coeffcient for SST and SOT is indicated by the blue solid line and the red solid line, respectively. In (a) and (b), units are arbitrary, and the green bold curve in (b) is the climatological mean of the 20 °C isotherm. Variance fraction in the SST and SOT is expressed as a percentage value on the upper right corner above each panel, and the correlation coeffcient between the expansion coeffcients is on the top right corner above (c)

图3 台站观测(a)与TRMM 3B42反演降水(b)得到的1998—2004年暖季降水峰值时间分布(箭头指向表示当地时间，时钟如(a)左上角所示，填色为地形高度，(c)为台站观测(黑虚线)与TRMM反演(灰实线)降水(单位：mm/h)平均的日变化曲线)Fig3 Precipitation from station observation (a) and TRMM 3B42 (b) in warm seasons during 1998—2004. Vectors denote the local solar time (LST) when the rainfall reaches the diurnal maxima (The clock is shown in the upper-left of the plot. The shading denotes the topography. (c) The diurnal curves of average rainfall amount (unit∶ mm/h) from station observations (black dashed line) and TRMM 3B42 (grey solid line)

图4 气候强迫因子对中国东部表面气温变化的贡献(气候强迫因子包括温室气体(CO2和附加温室气体)、太阳入射辐射变化、太阳轨道变化、异常火山灰气溶胶、陆表植被变化、人为排放的硫酸盐气溶胶和所有气候强迫因子, 灰色表示气候强迫因子的单独贡献之和)Fig4 Contributions of climate forcing factors to the temperature variation of East China (transient run minus equilibrium run). The climate forcing factors are greenhouse gases (CO2and additional greenhouse gases), solar insolation variability, solar orbit change, anomalous volcano aerosols, land cover changes, and sulfate aerosols. The red and grey lines represent the contributions of all climate forcing factors and the sum of contributions of individual climate forcing factor, respectively

Progress in Climate and Climate Change Research

In 2012, the Institute of Climate System of CAMS focused their researches on variability and mechanisms of the East Asian monsoon (EAM) and its effects on the climate of China, explored (1) variability, mechanisms, and prediction of the EAM, (2) influences of the air-sea-land interactions on the East Asian climate, (3) atmospheric thermodynamic and dynamic processes over the Tibetan Plateau and their impacts, and (4) mechanisms and projection theories of climate change. Some new observational phenomena have been unveiled, and innovative research results have been achieved.

1 Variability, mechanisms and prediction of the East Asian monsoon

Researches in this feld include the relation between late summer precipitation and temperature in eastern China; infuences of MJO, ENSO, and AO on the climatic disasters in China; identifcation of the predictors for the East Asian winter monsoon; and evaluation of the NCEP Climate Forecast System.

1.1 Decadal variability of climate characteristics in China

The late summer rainfall over eastern contiguous China is classifed into moderate, intense, and extreme precipitation according to hourly rainfall intensity structure. The e-folding decay intensity derived from the exponential distribution of rainfall amount is defned as the threshold that partitions rainfall into moderate and intense rainfall, and the double e-folding decay intensity is used as the threshold to pick out extreme cases. In the last several decades, the ratio between moderate and intense rainfall has experienced signifcant changes. Moreover, the spatial pattern of changes in the percentage of moderate rainfall presents a direct relation with that of the surface air temperature. Based on temperature changes, three regimes, i.e. northern China (warming), central eastern China (cooling), and southeastern coastal area of China (warming), are defned. In the warming regimes, the percentage of moderate rainfall exhibits a decreasing trend. In central eastern China, where the temperature has fallen, the percentage of moderate rainfall increased prominently. In all the three regimes, signifcant negative (positive) correlations between the percentage of moderate (intense) rainfall and the change of temperature are found. The relation between the extreme rainfall and the surface air temperature is different in different regions (Fig1).

1.2 Impacts of MJO, ENSO, and AO on the climatic disasters in China

The persistent anomalies of Madden-Julian Oscillation (MJO) and Arctic Oscillation (AO) have important implications for the drought and food in eastern Asia. The three-season drought in autumn, winter, and spring during the period of 2009—2010 in Yunnan Province was caused by the extreme anomalies of MJO and AO∶MJO was inactive while AO was unusually weak. The persistent inactive MJO was also responsible for the extreme summer drought in Yunnan during the main food season in 2011. During the negative phase of the AO, geopotential height at 500 hPa decreases around the Lake Baikal, the Ural blocking high develops, and meridional circulation anomalies prevail over East Asia, indirectly causing a cold and dry climate in North China.

1.3 Identifcation of the predictors for the East Asian winter monsoon

The SST anomalies at the midlatitude eastern Pacifc, the sea ice concentration in the Kara Sea, and the upper-level temperature over East Asia are revealed to be persistent. These abnormal signals can maintain from autumn to winter and infuence the intensity of the winter monsoon. A predictive model for the East Asian winter monsoon has been set up based on this fnding and has been applied in prediction of the 2012/2013 winter monsoon.

1.4 Evaluation of the NCEP Climate Forecast System

A series of 60-day hindcasts by the Climate Forecast System of the US National Centers for Environmental Prediction is analyzed to understand the impacts of atmospheric model resolutions (T62, T126, and T254) on predictions of the Asian summer monsoon. It is found that, in predicting the magnitude and timing of monsoon rainfall over lands, high model resolutions overall perform better than lower model resolutions. The increase in prediction skills with model resolution is more apparent over South Asia than over Southeast Asia. The largest improvement is seen over the Tibetan Plateau, at least for precipitation. However, the increase in model resolution does not enhance the skill of the predictions over oceans. Over the South China Sea and the Philippine Sea, simulations of monsoon rainfall even become worse from T62 to T254. (Li Jian)

2 Infuences of the air-sea-land interactions on the East Asian climate

The dynamic cause of the declining Arctic sea ice extent are revealed. The influences of the Eurasian snow on the East Asian monsoon, the relationship between two types of ENSO events and Pacifc subsurface SST, and the variations of the North Pacifc subtropical and subpolar gyres are discussed.

2.1 Dominant modes in the Arctic wind felds

Monthly mean surface wind data from the National Centers for Environmental Prediction/National Centers for Atmospheric Research (NCEP/NCAR) reanalysis dataset during the period 1979—2010 are used to describe the frst two patterns of Arctic surface wind variability by means of the complex vector empirical orthogonal function (CVEOF) analysis. The leading pattern consists of two sub-patterns∶ the northern Laptev Sea (NLS) pattern and the Arctic dipole (AD) pattern. The second pattern contains the northern Kara Sea (NKS) pattern and the central Arctic (CA) pattern. Over the past two decades, the combined dynamical forcing of the frst two patterns has contributed to the Arctic September sea ice extent (SIE) minima and its declining trend. The September SIE minima are mainly associated with the negative phase of the AD pattern and the positive phase of the CA pattern during the summer season (July to September), and both phases coherently show an anomalous anticyclone over the Arctic Ocean. Wind patterns affect the September SIE through their frequency and intensity. The negative trend in the September SIE over the past two decades is associated with increased frequency and enhanced intensity of the CA pattern during the melting season from April to September; thus, it cannot be simply attributed to the AD anomaly characterized by the second empirical orthogonal function mode of sea level pressure north of 70°N. The CA pattern exhibits interdecadal variability in the late 1990s, and an anomalous cyclone prevails before 1997 and is then replaced by an anomalous anticyclone over the Arctic Ocean that is consistent with the rapid decline trend in the September SIE.

2.2 Simulation of the infuences of the Eurasian snow on the East Asian monsoon

The snow albedo and hydrological effects on the precipitation in South China in 2010 was investigated based on the singular value decomposition (SVD) analysis of the Eurasian snow depth dataset and the 160-station rainfall dataset in China. Three ensemble simulations were conducted by using the community atmosphere model 3.1 (CAM3.1). The frst ensemble simulation includes both the snow albedo and the snow hydrological effect. The second ensemble simulation considers only the snow albedo effect but ignores the hydrological effect. The third ensemble simulation considers only the snow hydrological effect but ignores the albedo effect. The results indicate that both effects could have impacts on the South China rainfall. However, there are great differences in amplitudes and ranges of abnormities induced by these two effects. The magnitude of abnormities caused by the snow hydrological effect is larger than that by the albedo effect. When those two effects work together, the result agrees well with observations and the magnitude is the largest. However, the magnitude is not equal to the sum of that caused by the snow albedo and hydrological effects.

2.3 Impacts of spring Eurasian snowmelt on the East Asian summer monsoon

The effects of spring Eurasian snowmelt on the East Asian summer monsoon (EASM) and summer precipitation in China, in addition to a possibly related physical mechanism, were investigated by using observation data. The leading mode of spring Eurasian snowmelt shows a decadal variation, which agrees well with the EASM interdecadal transition and the China summer rainfall. In addition, the second mode of the snowmelt variability correlates positively to the EASM variability. When the snowmelt in East Siberia near the Balkhash Lake increases in spring, the EASM tends to appear in a negative–positive–negative meridional wave train from high to low latitudes in summer. When the snowmelt decreases, opposite conditions appear. The link mechanism of the spring snowmelt and summer atmospheric circulation in East Asia may be attributed to a positive feedback between abnormal snow and atmospheric thickness anomalies in the same region during the same period, which promotes two high-pressure systems. Subsequently, the Balkhash Lake high-pressure system develops eastward, and part of the East Siberian high-pressure system moves to lower latitudes. As a result, the atmospheric circulations over East Asia may form the meridional wave train in summer, which may lead to an anomalous change in the East Asian summer weather and climate.

2.4 Linkage between two types of ENSO events and Pacifc subsurface SST

The variations of subsurface ocean temperature (SOT) were investigated to disclose their linkage to the eastern and central Pacifc ENSO (EP and CP-ENSO) events. The wavelet analyses suggest that the variation of the EP (CP-ENSO) events shows a 2–7- (10–15-) yr oscillation in the tropical sea surface temperature (SST), and is coupled with a zonal dipole mode and a tripole mode in the SOT anomalous feld revealed by the singular value decomposition (SVD) analysis. During the mature phase of CP-ENSO, the positive center of SOT at the subsurface layer locates in the west of dateline, which results in the increase of SOT in the Niño4 region and causes the CP-ENSO event. Statistical analysis implies that, the eastern and central Pacifc subsurface indices defned by the expansion coeffcients of the frst and third SVD mode for SOT have shown their capabilities in distinguishing the EP and CP-ENSO events. In addition, corresponding to the increase of the SOT amplitude on the 10–15-yr timescale, we fnd that the frequency and intensity of CP-El Niño events have exhibited an upward trend after the 1980s, which suggests that the CP-ENSO event had an enhanced impact on the global climate in the past decades (Fig2).

2.5 Variations of the North Pacifc subtropical and subpolar gyres

The adjustment of the North Pacific Subtropical and Subpolar Gyres towards changes in wind stress leads to variability on different timescales, which plays a significant role in climate changes. Based on the Simple Ocean Data Assimilation (SODA) and Global Ocean Data Assimilation System (GODAS) datasets, we diagnose variations of the Subtropical and Subpolar Gyres by applying the “three-dimension Ocean Circulation Diagnostic Method”, and establish three types of index series consisting of the strength, meridional and depth centers of the Subtropical and Subpolar Gyres. The results show that the indices can well present the seasonal, interannual, and interdecadal variability of the Subtropical and Subpolar Gyres. The Gyres are both strongest in winter, but the Subtropical Gyre is weakest in summer and the Subpolar Gyre is weakest in autumn. The subtropical gyre starts moving northward from February to March, southward in October, and to the southernmost around January. In contrast, the Subpolar Gyre moves northward in spring, southward in summer, northward again in autumn, and reaches the extreme southern point in winter. The common feature in their interannual and interdecadal variability is that the two gyres are weaker and more northward before 1976—1977, while stronger and more southward after 1976—1977. The Subpolar Gyre has made a paramount contribution to variability on interdecadal scales. As is indicated by the Subpolar Gyre strength index, there is an important shift from weak to strong around 1976—1977, and its correlation coeffcient with the North Pacifc Decadal Oscillation (PDO) is 0.45, which is far better than that between the subtropical gyre strength and PDO. Experimental results show that infuenced by small and mesoscale eddies, the magnitude of largescale gyres is strongly dependent on data resolution, but seasonal, interannual, and interdecadal large scale variability of the two gyres as manifested by the indices is less affected by model resolution. (Zhu Congwen)

3 Atmospheric thermodynamic and dynamic processes over the Tibetan Plateau and their impacts

Based on the satellite observation data, the diurnal variation of rainfall over the southeastern Tibetan Plateau, the Tibetan Plateau vortices, and the diurnal variation of clouds over southern China are studied.

3.1 Diurnal variation of rainfall over the southeastern Tibetan Plateau

The terrain of the Tibetan Plateau is complex, and it buckles into a series of ridges with a succession of gorges that carve their way through the mountains. The complex topography, as well as the substantial inhomogeneity in landscape and landcover on the Tibetan Plateau, may therefore produce some unique regional diurnal rainfall characteristics. The results show that the diurnal variation of summer rainfall over the Tibetan Plateau has evident double peaks. The prevailing nocturnal rainfall peak in observations at routine stations can be largely attributed to the relatively lower location of the stations, which are mostly situated in valleys. The records from a 3-yr intensifed observational experiment at eight stations along the hillside of Seqilashan over the southeastern Tibetan Plateau revealed an evident afternoon peak of warm season rainfall, similar to that indicated by the TRMM data. The different diurnal phases between valley and hillside station precipitations are closely related to the orographically induced regional circulations. The results of this study indicate that the prevailing nocturnal rainfall associated with the relatively lower location of routine observation stations can partially explain the diurnal rainfall variations between observation station records and TRMM data (Fig3).

3.2 Tibetan Plateau vortices and the diurnal variation of clouds over southern China

Using the NCEP global fnal analysis data (FNL), the observational radiosonde data, the hourly surface climatic elements dataset, the atmosphere precipitable water vapor (PWV) and 6-h radiosonde data at Gerze station, the generation frequency of the Tibetan Plateau vortices (TPVs) in 4 periods of a day (00∶00—06∶00 UTC, 06∶00—12∶00 UTC, 12∶00—18∶00 UTC, and 18∶00—00∶00 UTC), during May—September of 2006—2008, is analyzed. The effects of the large-scale circulation, the related meteorological elements, and the heating fields on the diurnal variation of the TPVs’ formation frequency are discussed. The generation frequency of the TPVs shows a robust diurnal variation, depending on both the large-scale circulation and the latent heat. The peak of the generation frequency of the TPVs tends to reach the maximum during evening to midnight (18∶00—00∶00 LT), and the minimum during early morning to noon (06∶00—12∶00 LT). The condensation latent heat induced by precipitation exerts a direct promotion effect on the generation of the TPVs. In the evening (at 18∶00 LT), there is notable convergence at 500 hPa over the central-western plateau, the jet stream and the South Asian high at 200 hPa are strong, the water vapor flows to the main body of the Tibetan Plateau, and the stratification is unstable at this time. All of the above conditions are helpful for triggering precipitation, which induces the condensation latent heat release. The upper-level heating is favorable for the development of the cyclonic disturbance at 500 hPa during 18∶00—00 ∶00 LT. The reverse situation occurs druing 06∶00—12∶00 LT.

3.3 FY-2C derived diurnal features of clouds in the southern contiguous China

Hourly infrared (IR) brightness temperature (BT) derived from China’s first operational geostationary meteorological satellite Feng-Yun (FY)-2C is analyzed over the southern contiguous China (20°–33°N, 100°–122°E) during 2005—2008. The focus is to investigate the diurnal variation of clouds and compare their different features between the southwestern and southeastern China. According to the diurnal features in summer, the clouds are frst divided into three categories by the cloud top temperature (CTT, derived from IR BT)∶ cold cloud (CC, defined as CTT lower than -30 °C), middle cloud (MC, defined as CTT between-30 °C and 0 °C), and warm cloud (WC, defned as CTT warmer than 0 °C). The CC occurs most frequently in the late afternoon in most regions over southern China. The summer mean frequency of CC, MC, and WC shows different diurnal variations in the southern contiguous China. The CC exhibits a large diurnal amplitude in frequency and occurs most frequently in the late afternoon in most regions. The frequency of MC shows a relatively weaker diurnal amplitude and presents a dominant nocturnal maximum during the day. The WC occurs more (less) frequently in the daytime (nocturnal hours) and reaches the peak around noon. Despite the similarity, the diurnal features of summer mean CC and MC frequency in the southwestern and southeastern China present remarkable regional differences, whereas the diurnal variation of WC shows no obvious zonal contrasts over most of southern China. The CC occurs more frequently in the early morning (late afternoon) in the west (east) region. In the west region, the frequency of CC also has a weak secondary diurnal peak. The frequency of MC shows midnight to early morning maxima in the west region and evening to midnight maxima in the east region. The diurnal variations of CC and MC frequency exhibit evident seasonal changes, whereas the frequency of WC reaches the diurnal maxima around noon all over the year. The late afternoon peak of CC frequency appears in almost all months. In the warmer season, the late afternoon diurnal maximum dominates; whereas in the cold seasons, there is also a midnight to late evening secondary peak. The MC frequency reaches the diurnal peak in the early evening in summer and in the late evening in other seasons. The seasonal changes of WC frequency only present in its diurnal amplitude, which is smaller in warmer seasons and larger in cold seasons. It is noted that the distribution of diurnal variation of the CC (MC) shows great similarity with that of convective (stratiform) rainfall detected by TRMM PR, especially the diurnal variation of precipitation profles. The similarity indicates that the BT from FY-2 series is potentially useful for the study of the diurnal variation of clouds and rainfall in the southern contiguous China, which may further extend our knowledge of the evolution of the regional cloud and rainfall, as the geostationary orbit satellites can provide much more information than the polar orbit satellites such as TRMM PR. (Chen Haoming)

4 Mechanisms and projection theories of climate change

Using the observed data and numerical simulation, possible causes of the decadal variations and past millennial variations of the East Asian monsoon are discussed. The main results are as follows.

4.1 Impacts of the change of surface air temperature over the Asian continent on the East Asian summer monsoon

Using observed data and output from the atmospheric component of the NCEP Climate Forecast System, we fnd a relatively smaller warming in Asia compared to the surrounding regions. Although the surface air temperature over Asia has increased, the landmass has become a relative ‘heat sink’ because of the larger warming in other regions of the world. Indeed, over Asia, the vertically integrated tropospheric temperature in the most recent decade is colder than that in the earlier decades, a feature different from the characteristics outside Asia. Therefore, the “heat sink” over the Asian landmass may be a plausible reason for the weakening of the Asian monsoon.

4.2 Possible impacts of global warming on the East Asian summer monsoon

The possible causes of the weakening of the East Asian summer monsoon (EASM) were investigated. We found that the decreased intensity of the EASM is signifcantly correlated with the increase of the surface air temperature (SAT) averaged over the Lake Baikal region defned as SATI. Corresponding to the increasing SATI, an anomalous low-level anticyclone occurs with northeasterly prevailing over northern East Asia, resulting in a weakened southwesterly monsoon and drier climate in this region. Numerical experiments with the community atmosphere model version 3 (CAM3) show that the joint forcing induced by greenhouse gases (GHG), sea surface temperature (SST), solar radiance (SR), and volcano activity (VC) can replicate the observed trend of SATI and its related circulation anomalies, but without GHG forcing the model failed tosimulate the warming trend of SATI after the 1970s. This implies that the global warming is likely responsible for the local warming around the Lake Baikal, which in turn weakens the EASM in recent decades.

4.3 Modeling of the climate change in China by BCC_AGCM

Primary diagnostic metrics are proposed to evaluate the integrated structure of interdecadal changes of the East Asian climate in midsummer (July—August) over the recent half-century (1955—2000) in the BCC_ AGCM models. When forced by historical sea surface temperatures (SST), the ensemble simulation with the BCC_AGCM reasonably has reproduced the coherent interdecadal changes of rainfall, temperature, and circulation. The main feature of the ‘southern-fooding-and-northern-drought’ in rainfall change is captured by the model. Correspondingly, the tropospheric cooling in the upper and middle troposphere, the southward shift of the upper level westerly jet, and the weakening of the low-level southwesterly monsoon fow are also reproduced, as well as their relationships with rainfall changes. One of the main defciencies of the simulation is that the amplitudes of the changes of the tropospheric cooling and large-scale circulation are both much weaker than those in reanalysis, and they are consistent with the rainfall defciency. Also, the upper and middle troposphere cooling center and the decreasing of upper-level westerly jet axis shift westward in the model simulations compared with the observations. Overall, although BCC_AGCM shows problems in simulating the interdecadal changes of the East Asian climate, especially the amplitude and locations of change centers, it reasonably represents the observed confguration of rainfall variation and the associated coherent temperature and circulation changes. Therefore, it could be used to further discuss the mechanisms of the interdecadal variations in East Asia. Meanwhile, the reasonably reproduced confguration of rainfall and its associated largescale circulation by SST-forced runs indicate that the interdecadal variations in East Asia could mostly arise from the regional response to the global climate change.

4.4 Modeling of global climate change in the past millennium

The UVic Earth System Climate Model (UVic Model), an earth system model of intermediate complexity, was employed to investigate the contributions of climate forcings (e.g., solar insolation variability, anomalous volcanic aerosols, greenhouse gas, solar orbital change, land cover changes, and anthropogenic sulfate aerosols) to surface air temperature over East China in the past millennium. The simulation of the UVic Model could reproduce the three main characteristic periods (e.g., the Medieval Warm Period (MWP), the Little Ice Age (LIA), and the 20th Century Warming Period (20CWP)) of the Northern Hemisphere and East China, which were consistent with the corresponding reconstructed air temperatures on century scales. The simulation results reflected that the air temperature anomalies of East China were larger than those of the global air temperature during the MWP and the frst half of 20CWP and were lower than those during the LIA. According to the surface air temperature of East China, the past millennium has been further divided into three periods in the MWP, four in the LIA, and one in the 20CWP. The MWP of East China was caused primarily by solar insolation and secondarily by volcanic aerosols. The variation of the LIA was dominated by the relative contributions of solar insolation variability, greenhouse gas, and volcano aerosols. Greenhouse gas and volcano aerosols were the main forcing of the third and fourth periods of the LIA, respectively. We examined the nonlinear responses among the natural and anthropogenic forcings in terms of surface air temperature over East China. The nonlinear responses between the solar orbit change and anomalous volcano aerosols and those between the greenhouse gases and land cover change (or anthropogenic sulfate aerosols) all contributed approximately 0.2 °C by the end of the 20th century. However, the output of the energy-moisture balance atmospheric model from UVic showed no obvious nonlinear responses between anthropogenic and natural forcings. The nonlinear responses among all the climate forcings (both anthropogenic and natural) contributed to a temperature increase of approximately 0.27 °C at the end of the 20th century, accounting for approximately half of the warming during this period (Fig4, red line minus grey line); the remainder was due to the linear response of the climate forcings themselves. The nonlinear responses of the climate forcing factors offer a new way to understand the climate change (Fig4). (Xiao Dong)