基于自组织映射神经网络的东亚区大尺度环流异常及其与气候异常的关系研究

	基于自组织映射神经网络的东亚区大尺度环流异常及其与气候异常的关系研究
其他题名	Study on the relationship between anomalous large-scale atmospheric circulation patterns and climate anomalies over East Asia based on self-organizing map neural network
	杨叶
学位类型	硕士
导师	高猛
	2020-05-15
学位授予单位	中国科学院大学
学位授予地点	中科院烟台海岸带研究所
学位名称	理学硕士
学位专业	地图学与地理信息系统
关键词	东亚气候变化大气环流模态自组织映射神经网络 East Asia Climate Change Atmospheric Circulation Patterns Self-Organizing Map Neural Network
摘要	大尺度大气环流异常是导致区域极端气候的主要原因之一。对气候再分析资料进行数据挖掘和统计分析是认知大气环流特征和极端气候发生规律的重要研究方法。本文首先利用自组织映射神经网络SOM（self-organizing map）对东亚地区1979-2017年间的异常大气环流模态进行分类。在此基础上，建立各种模态和多种气候影响因子间的对应关系，并着重分析了我国气温异常的变化趋势及其动力学机制。其次，将异常大气环流模态与东亚地区极端气温及降水事件进行关联，研究了异常大气环流与季风系统共同作用下的气候异常特征。最后，为更加深入地探索东亚地区在近40年来季均气候变化的特点及其大尺度环流条件，本文利用合成分析生成了西太平洋遥相关型WP（West Pacific Pattern）、北极涛动AO（Arctic Oscillation）及厄尔尼诺-南方涛动ENSO（El Niño-Southern Oscillation）正负位相时期的异常大气环流模态和海表温度距平模态，并选择受遥相关影响较大的一些区域作为研究对象，从海-气耦合作用角度探讨了遥相关型对这些区域气候变化的影响。主要结论如下：（1）SOM方法可以有效表征东亚地区典型的大尺度异常大气环流模态。（2）冬季AO会显著影响我国尤其是东北地区的气温变化。在夏季，AO的效应大幅减弱。冬季拉尼娜（La Niña）事件时期，我国大多数地区气温异常偏暖，而在厄尔尼诺（El Niño）事件时期，北方地区气温异常偏冷。夏季，ENSO和我国气温异常并不存在明显的联系。（3）将数据划分为1979-1998和1999-2017年两个时期，研究分析了极端气温事件的变化情况及其成因。结果显示，极端冷事件的减少，极端暖事件的增加主要和热力学因素有关，相反，动力学因素对这一变化的贡献微小，在某些情况下甚至起到了抑制作用。结果进一步证实了全球变暖对局域气候变化的重要性。（4）SOM的分类结果表明，在冬季，东亚地区极端降水事件的发生取决于大气水平环流及垂直环流的共同作用。在夏季，降水主要同大气垂直环流相关，其与水平环流间的关系并不显著。（5）较强的东亚冬季风EAWM（East Asian Winter Monsoon）会造成东亚地区出现大范围的冷天气，其与降水事件间的联系依赖于地理位置。而SOM方法对于东亚夏季风EASM（East Asian Summer Monsoon）强弱位相时期的大气环流形势的表征结果并不令人满意。（6）活跃的WP和AO会引发冬季东亚地区大范围的气温异常，而ENSO的作用相对较小。此外，研究还表明遥相关型对月际气温变化的效应更强，而日际气温受到局域-天气尺度环流系统和遥相关型的共同控制。
其他摘要	Anomalous large-scale atmospheric circulation is considered as one of the main drivers causing the regional climate anomalies. The application of data mining and statistical analysis on the re-analysis data is very useful for identifying the atmospheric circulation characteristics and the rules of climate extremes. In this thesis, the self-organizing map (SOM) method was firstly used to classify the large-scale anomalous atmospheric circulation patterns over East Asia during 1979-2017. The relationships between atmospheric circulation patterns and surface air temperature variations then were established to explore the possible formation mechanism of climate changes in China. Secondly, the anomalous atmospheric circulation patterns were further related to the temperature anomalies and precipitation anomalies over the East Asia Monsoon region to reveal the climate extremes under the combing effect of atmospheric circulation and monsoon. Finally, in order to explore the characteristics of climate changes and their large-scale circulation background, the atmospheric circulation anomaly patterns and sea surface temperature anomaly patterns in the strong positive/negative phase of Western Pacific mode (WP), Arctic Oscillation (AO) and El Niño-Southern Oscillation (ENSO) were composited to examine the role of teleconnections on the climate fluctuations in local areas in terms of atmospheric and oceanic process. The main conclusions in this study were drawn as follows: (1) The large-scale anomalous atmospheric circulation patterns over East Asia could be identified based on SOM method. (2) AO severely contributed to the surface air temperature variations during the winter within China, especially in northeastern area, while its effect was weakened prominently in summer. A La Niña event in winter enhanced the anomalous warm throughout the whole region of China, whereas an El Niño event was linked to the anomalous cold over northern area. The temperature anomalies in summer within China were less correlated with ENSO. (3) Difference of pattern frequencies and extreme occurrence intensity (EOI) in winter and in summer for the two periods of 1979-1998 and 1999-2017 were calculated to investigate the possible mechanism of changes of temperature extremes. Our analysis implied that the changes of EOI were mainly favored by the thermodynamics factor, whereas dynamic factor has little contribution on temperature extreme changes, and sometimes even has a negative effect. (4) The precipitation anomalies and extremes associated with anomalous large-scale atmospheric circulation patterns over East Asia during the winter and summer were presented in the fourth chapter. The precipitation extremes in winter were mostly enhanced by the joint effect of the horizontal and vertical atmospheric circulations. By contrast, the precipitation anomalies in summer were merely explained by the upward atmospheric motion. (5) The strong East Asian Summer Monsoon (EASM) enhanced the cold anomalies and extremes over most of the region within East Asia. Nevertheless, the relationship between the East Asian Winter Monsoon (EAWM) and precipitation extremes depended on the geographical locations. Additionally, it was not satisfied with the ability of SOM method to capture the typical features of anomalous atmospheric circulation patterns in the strong/weak phase of EASM. (6) The WP and AO patterns played an essential part in the prevalence of prominent Tmin fluctuations over East Asia in winter, while the effect of ENSO was relatively weak. Besides, the daily cold extremes were primarily triggered by the local-synoptic atmospheric circulations embedded in the large-scale teleconnection modes, while the monthly cold extremes have a closer relationship with these low-frequency modes.
语种	中文
文献类型	学位论文
条目标识符	http://ir.yic.ac.cn/handle/133337/24209
专题	中国科学院烟台海岸带研究所知识产出_学位论文中国科学院烟台海岸带研究所知识产出
推荐引用方式 GB/T 7714	杨叶. 基于自组织映射神经网络的东亚区大尺度环流异常及其与气候异常的关系研究[D]. 中科院烟台海岸带研究所. 中国科学院大学,2020.

条目包含的文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
基于自组织映射神经网络的东亚区大尺度环流（8681KB）	学位论文		开放获取	CC BY-NC-SA	浏览请求全文