基于统计方法的SST年际和年代际可预报性研究

发布时间：2019-02-15 03:27

【摘要】：气候系统是一个复杂的非线性系统,(不)可预报性是其固有的属性。研究表明,气候系统年际到年代际的可预报性主要来自海洋。海洋作为气候系统的重要组成部分,其海表面温度(SST)是衡量气候平均和变率的一个重要因子。因此,研究全球海表面温度SST的年际和年代际可预报性具有重要意义,可以为预测未来的气候变化提供依据。气候变量的可预报性定义为可预报分量的方差与总方差的比值。与经验正交函数(EOF)分解类似,根据可预报性最大的原则,可以将气候变量分解为可预报成分和空间结构的线性组合。本文分别利用NOAA扩展重建的ERSST资料和GFDL模式CM3工业化革命前试验模拟结果研究SST的年际、年代际可预报性和可预报成分,从而寻找海洋中存在年际和年代际可预报性的主要区域。通过对观测的月平均SST进行分析可知,月平均全球SST的可预报性为3个月,第一可预报成分的可预报性为2年以上,空间上表现为北太平洋和北大西洋的异常增暖,表征了与AMO相似的SST气候态的长期波动特征,第二三可预报成分的可预报性为6个月左右。年际可预报性主要集中在热带太平洋,热带太平洋SST的可预报性为4个月,可预报成分具有与ENSO类似的结构,均呈现热带中东太平洋的异常增暖,其中第二可预报成分与Nino3指数相关较高。因此,热带太平洋SST的可预报性来自ENSO.通过对CM3模式模拟的工业化革命以前的年平均SST进行分析可知,全球SST在前置时间为1年时,预报技巧为0.55。SST的年代际可预报性主要集中在中高纬度。北太平洋、北大西洋SST的可预报成分具有5年以上的可预报性,并呈现明显的年代际变率,北太平洋SST第二可预报成分与太平洋年代际振荡PDO有一定的相关,北大西洋SST第二可预报成分与大西洋多年代际振荡AMO相关较好。综上所述,SST的年际可预报性主要在热带,并且与ENSO有一定的联系,而SST的年代际可预报性主要在中高纬度,如北太平洋、北大西洋,年代际可预报性与太平洋年代际振荡PDO以及大西洋多年代际振荡AMO有一定的相关。
[Abstract]:Climate system is a complex nonlinear system, and (not) predictability is its inherent attribute. Studies have shown that the interannual to Interdecadal predictability of the climate system comes mainly from the oceans. As an important part of the climate system, the sea surface temperature (SST) is an important factor to measure the climate average and variability. Therefore, it is of great significance to study the interannual and Interdecadal predictability of global sea surface temperature (SST), which can provide a basis for predicting future climate change. Predictability of climate variables is defined as the ratio of variance to total variance of predictable components. Similar to the empirical orthogonal function (EOF) decomposition, climate variables can be decomposed into linear combinations of predictable components and spatial structures according to the principle of maximum predictability. In this paper, the interannual, interdecadal predictability and predictable components of SST are studied by using the ERSST data reconstructed by NOAA and the simulated results of the GFDL model CM3 industrial pre-revolution test, respectively. Thus, the main areas of the ocean with interannual and interdecadal predictability are found. Through the analysis of the observed monthly average SST, it can be seen that the predictability of the monthly average global SST is 3 months, the predictability of the first predictable component is more than 2 years, and the spatial warming of the North Pacific and North Atlantic is abnormal. The long-term fluctuation characteristics of SST climate state similar to AMO are characterized. The predictability of the second and third predictable components is about 6 months. The interannual predictability is mainly concentrated in the tropical Pacific, and the predictability of SST in the tropical Pacific is 4 months. The predictable component has a structure similar to that of ENSO, and all of them show abnormal warming in the tropical Middle East Pacific. The second predictable component is highly correlated with Nino3 index. Therefore, the predictability of SST in the tropical Pacific comes from ENSO. Through the analysis of the average annual SST before the industrial revolution simulated by the CM3 model, it can be seen that the interdecadal predictability of the global SST is mainly concentrated in the middle and high latitudes when the leading time is one year. The predictable components of North Pacific and North Atlantic SST are predictable for more than 5 years and show obvious Interdecadal variability. The second predictable component of SST in the North Pacific is related to the Pacific Interdecadal oscillatory PDO to some extent. The second predictable component of the North Atlantic SST is well correlated with the AMO of the Atlantic Ocean intergenerational oscillation. To sum up, the interannual predictability of SST is mainly in the tropics and is related to ENSO, while the interdecadal predictability of SST is mainly in the middle and high latitudes, such as the North Pacific Ocean, the North Atlantic, the North Atlantic, Interdecadal predictability is associated with Pacific decadal oscillation (PDO) and Atlantic intergenerational oscillation (AMO).
【学位授予单位】：中国科学院研究生院(海洋研究所)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P732.4

【相似文献】