海洋环流与波动重点实验室知识库

黑潮延伸体是北太平洋副热带环流圈的重要组成部分，其在年代际尺度上呈现稳定模态和不稳定模态之间的低频振荡现象。不同模态对海洋环境、大气环流乃至全球气候状况有重要影响。模态转变的内在机制和可预报性研究是物理海洋学领域的前沿问题。本文以两种模态之间的转变过程为研究对象，通过资料分析考察了其动力学机制，进而在此基础上探究了与之相关的第二类可预报性问题之一，即风应力不确定性对黑潮延伸体模态转变过程预报的影响。

首先，通过研究涡动能（Eddy Kinetic Energy：EKE）的快速变化，揭示了黑潮延伸体模态转变的局地动力学机制。在模态转变过程中，黑潮延伸体区域的中尺度涡活跃程度表现出明显的变化。在稳定模态向不稳定模态转变（Stable-to-unstable：SU）阶段，涡动能强度由低转高，急流路径受此影响变得复杂、弯曲。上、中游（140–150°E）水平流速垂向剪切引起的斜压不稳定是涡动能增长的主要原因。从不稳定模态向稳定模态的转变（Unstable-to-stable：US）过程中，黑潮延伸体中游（145–150°E）水平平流和下游（150–155°E）风应力耗散的协同作用导致了涡能量的衰退。当涡旋消耗殆尽，黑潮延伸体的路径完成了向稳定模态的转变。

涡动能研究表明，风应力在黑潮延伸体模态转变过程中起着一定的作用。为此，基于区域海洋模式（ROMS）并利用条件非线性最优边界扰动（Conditional Nonlinear Optimal Perturbation for Boundary Conditions：CNOP-B）方法进一步探究了风应力误差对模态转变预报的影响。ROMS模式成功再现了1990至2012年间的四次模态转变事件，即两个SU转变事件与两个US转变事件。针对不同的转变事件，考察了能够引起最大预报误差且时间依赖的最优风应力误差（最优边界误差；Optimally Boundary Error：OBE）。OBE空间上呈现多涡旋的结构，大值位于黑潮延伸体主轴附近。时间上，其强度随着接近预报时刻而逐渐衰减。OBE起初通过埃克曼抽吸引起小尺度海洋误差，这些误差随后在海洋内部过程的作用下不断增长，最终导致黑潮延伸体路径状态预报的显著偏差。在预报中，OBE总是抑制模态转变过程的发生。在SU转变预报中，与流速水平剪切相关的正压不稳定过程是误差增长的主要原因；在US转变预报中，误差的发展由风应力做功主导。

最后，开展了黑潮延伸体模态转变预报中关于风应力的适应性观测研究。对OBE的时空结构进行了敏感性试验。结果表明，扰乱OBE的时间或空间结构均会引起预报误差的大幅下降。OBE的时间和空间结构均对最大预报误差的产生至关重要。基于OBE的时间结构识别出了对预报效果影响较大的适应性观测敏感时段，即预报期的前7−10个月。根据OBE空间分布确定的观测敏感区位于黑潮延伸体主轴附近。在敏感区内去除风应力误差能够明显提高数值预报技巧，改善程度高达65%，其效果显著优于其他对比区域。

本文的研究加深了人们对黑潮延伸体年代际变率的理解和认识，并为黑潮延伸体区域的风场观测方案设计和改进路径状态预报提供了理论指导。

The Kuroshio Extension (KE) is an important component of the North Pacific Subtropical Gyre. The KE exhibits decadal low-frequency oscillation between stable and unstable states. Different states have significant impacts on marine environments, atmospheric circulation, and even global climate. The intrinsic mechanism and predictability of the KE state transitions are frontier issues in physical oceanography. This study, focusing on the transition process between the KE bimodality, investigates its dynamic mechanisms through data analysis, and further explores one of the related second kind predictability problems, namely the impact of wind stress uncertainty on the prediction of KE state transitions.

Firstly, by studying the rapid changes in Eddy Kinetic Energy (EKE), the local dynamic mechanisms are revealed for the KE state transitions. The activity of mesoscale eddies in the KE region varies significantly with the dynamic state of the KE. During the transition from the stable-to-unstable state (SU), the EKE intensity increases as the KE path becomes more complex and curved due to eddy disturbances. The baroclinic instability induced by vertical shear in the mid-upstream region (140–150°E) is the main cause of EKE growth. In the transition from the unstable-to-stable state (US), the combined effect of horizontal advection in the midstream (145–150°E) and wind stress dissipation in the downstream (150–155°E) results in the decay of eddy energy. When eddies are depleted, the KE path completes the transition to a stable state.

The EKE analysis indicates that wind stress does play a role in KE state transitions, prompting this study to further investigate the impact of wind stress errors on transition predictions. It employs the Conditional Nonlinear Optimal Perturbation for Boundary Conditions (CNOP-B) method and Regional Ocean Modeling System (ROMS) to conduct the predictability study. ROMS successfully reproduces four state transitional events from 1990 to 2012, namely two SU and two US processes. The time-dependent optimal wind stress errors (Optimal Boundary Error: OBE), which causes the maximum forecast errors, are investigated for four transitional events. OBE exhibits a local multi-eddies spatial structure with decreasing magnitude as the end time of prediction approaches. The OBE initially induces small oceanic errors through Ekman pumping. Subsequently, these errors grow in magnitude as oceanic internal processes take effect, which exerts significant influences on KE path predictions. OBE always inhibits the occurrence of state transitions in predictions. Barotropic instability associated with horizontal velocity shear is crucial for the error growth in the prediction of the SU transitions while work generated by wind stress error plays a more dominant role in the US transitions.

Eventually, this study investigates the adaptive observation of wind stress for the prediction of KE transitions. Sensitivity experiments are conducted on the temporal and spatial structure of OBE. The results indicate that perturbing either the temporal or spatial structure significantly reduces prediction errors. Both the temporal and spatial structures are crucial for generating maximum prediction errors. Based on the temporal structure of OBE, the sensitive periods for adaptive observations impacting prediction accuracy are identified within the first 7−10 months of the forecast period. The sensitive areas, determined by the spatial distribution of OBE, are located near the KE axis. Removing wind stress errors in sensitive areas can enhance numerical prediction skills, with improvements reaching up to 65%, outperforming other comparison regions.

This study deepens understanding of the decadal variability of the KE and provides theoretical guidance for wind observation schemes and forecast of the KE path state.

第1章 绪论 1

1.1 选题背景与研究意义   1

1.1.1 黑潮延伸体的基本特征    1

1.1.2 风应力与黑潮延伸体的相互作用    2

1.1.3 黑潮延伸体的模态转变    3

1.2 研究现状与研究进展   4

1.2.1 黑潮延伸体模态转变的机制研究    4

1.2.2 可预报性研究进展    6

1.3 科学问题、研究内容及章节安排      7

第2章 黑潮延伸体模态转变过程的局地能量学机制    9

2.1 引言       9

2.2 方法和数据   10

2.2.1 涡动能能量学收支计算    10

2.2.2 数据介绍    11

2.3 模态转变过程中的涡动能分布   12

2.4 模态转变的能量学机制      16

2.4.1 稳定向不稳定模态转变的机制 16

2.4.2 不稳定向稳定模态转变的机制 22

2.5 小结与讨论   27

第3章 数值模拟及CNOP-B优化系统构建   29

3.1 ROMS模式    29

3.1.1 模式介绍    29

3.1.2 模式动力框架    31

3.1.3 模式坐标系统    33

3.1.4 模式网格    36

3.2 模式设置      37

3.3 模式模拟效果      39

3.4 条件非线性最优边界扰动方法   43

3.5 物理约束下的非线性优化系统   45

3.6 小结       47

第4章 黑潮延伸体模态转变预报的最优风应力误差    49

4.1 引言       49

4.2 最优风应力误差的求解      50

4.3 最优风应力误差   52

4.3.1 最优风应力误差的结构    53

4.3.2 最优风应力误差对模态转变预报的影响 57

4.3.3 预报误差的产生原理 58

4.3.4 预报误差的增长机制 62

4.4 小结与讨论   71

第5章 风应力的适应性观测研究    73

5.1 引言       73

5.2 最优风应力误差的敏感性试验   74

5.3 适应性观测敏感区识别及验证   78

5.4 小结与讨论   82

第6章 结论与展望    85

6.1 本文主要结论与创新点      85

6.2 工作展望与讨论   86

参考文献      89

附录一 英文缩写一览表    107

附录二 涡能量收支方程的推导 111

致谢      113

作者简历及攻读学位期间发表的学术论文与其他相关学术成果   115