海洋环流与波动重点实验室知识库

南极绕极流（the Antarctic Circumpolar Current，ACC）作为南大洋中最显著的风生环流系统，具有宽流幅、高流速、深影响等特征，对跨海盆尺度的物质能量输运、局地海气相互作用、生态与天气系统、海洋碳汇过程等具有重要的调制作用。观测和模拟均发现ACC流量存在较大的不确定性，特别是在次季节尺度内的快速变化现象，又被称为ACC流量跃变现象。该现象为ACC流量的短期预报带来了巨大挑战。

ACC流量短期变率主要受到涡旋、锋面等中小尺度过程的影响。因此，本文从海洋内部变率的角度开展了德雷克海峡（the Drake Passage，DP）扇区ACC流量跃变的可预报性研究。具体地，基于区域海洋模式（Regional Ocean Modeling System，ROMS）和条件非线性最优扰动（Conditional Optimal Nonlinear Perturbation，CNOP）方法，本文分别计算了ACC流量跃变的最优前期征兆（optimal precursor，OPR）和最快增长初始误差（optimally growing initial error，OGIE），并在此基础上开展了ACC流量跃变的目标观测研究。主要研究结果如下：

（1）利用ROMS对南大洋ACC进行了高分辨率的数值模拟。结果表明：ROMS能够较为准确地模拟ACC气候态结构和多尺度时间变率特征；模拟的ACC流量为136.5±8.0 Sv（1 Sv=10⁶ m³ s^–1），与观测结果一致；特别地，在次季节尺度内ACC流量具有较强的时间变率，且能够表征30天内的跃变现象，为计算OPR和OGIE提供了研究基础。

（2）基于ROMS伴随模式建立的非线性优化框架，针对三个高流量个例和三个低流量个例，分别计算了导致ACC流量跃变事件发生的OPR。结果表明：OPR在不同个例中具有相似的结构特征，其大值区均位于DP中部（58°S-62°S，72°W-64°W）1000-3000米区域；通过在上层流场中激发出一对涡旋状偶极子扰动模态，OPR对ACC流量产生影响；在高流量（低流量）个例中，OPR倾向于触发流量由高向低（由低向高）跃变；在OPR的非线性发展过程中，受到斜压不稳定的主导作用，扰动动能具有更快的增长速度。

（3）基于OPR所触发的ACC流量跃变过程计算了导致预报误差发展最大的OGIE。结果表明，OGIE的结构特征、发展和演变机制与OPR类似。OGIE集中分布在DP中部（58°S-62°S，72°W-64°W）3000米区域；对于高流量向低流量（低流量向高流量）跃变过程，OGIE倾向于增大（减小）预报的流量；整体而言，OGIE能够减弱甚至消除背景场中流量跃变过程；该过程受到斜压不稳定的主导作用，表现为一对涡旋状偶极子误差模态的生成与发展；同时，涡度误差收支的诊断分析表明线性平流和非线性平流过程相互制约，共同维持了OGIE的局地性发展。

（4）利用OGIE的水平分布，确定了预报ACC流量跃变的CNOP型观测敏感区，并进行目标观测研究。敏感性试验表明，ACC流量对于CNOP型敏感区内（尤其2000 m以下）的初始“观测”误差具有更大的敏感性。进一步，利用观测系统模拟试验，验证了在CNOP型敏感区内（尤其2000 m以下） 去除初始“观测”误差能够较大程度改善（超过50%）ACC流量跃变的预报效果。

本文通过探究ACC流量跃变的OPR、OGIE以及目标观测问题，揭示了海洋内部斜压不稳定和非线性过程对ACC流量短期变异的重要作用。同时，验证了在识别的敏感区内增加观测和优化初始场将有助于提升ACC流量跃变的预报技巧，对于设计和优化南大洋德雷克扇区ACC的观测网络具有一定的指导意义。

As the most prominent wind-driven circulation system in the Southern Ocean, the Antarctic Circumpolar Current (ACC) exhibits wide-extending, high-flowing, and deep-reaching features, which has a critical influence on the transmission of mass and energy across the major oceans, the regional air-sea interaction, ecosystem, weather system, oceanic carbon sinks, etc. Previous observations and simulations have found that ACC transport displays strong uncertainty, like the rapid fluctuations at the subseasonal time scale particularly. This is also called the sudden shifts, which can bring giant challenges for the short-range prediction of ACC transport.

Since that the great effects of the mesoscale and submesoscale processes (such as the eddies and front’s meanders) on the short-range variations of ACC transport, this thesis conducts the predictability study of the sudden shifts in ACC transport across the Drake Passage (DP) sector in the view of oceanic internal processes. Specifically, the optimal precursor (OPR) and the optimally growing initial error (OGIE) of the sudden shifts in ACC transport are calculated based on the Regional Ocean Modeling System (ROMS) and the Conditional Optimal Nonlinear Perturbation (CNOP) approach. On this basis, the targeted observation is also conducted. Major research results are described as follows.

(1) ROMS is configured to simulate the Southern Ocean. It is shown that ROMS can reproduce the climatological circulation and interannual variability of ACC accurately. The simulated ACC transport of 136.5±8.0 Sv (1Sv≡ 10⁶ m³ s^–1) is consistent with the observation. Specifically, the energic subseasonal variability and the sudden shifts in transport during 30 days are also identified well, providing a solid basis to explore the OPR and the OGIE.

(2) Nonlinear optimization framework is established based on the ROMS adjoint component, and the OPRs triggering the sudden shifts in ACC transport are explored for three high-transport cases and three low-transport cases. The results indicate that the OPRs keep similar distributions for various cases. The large values of OPRs are mainly located in the middle DP of 58°S–62°S, 72°W–64°W in the depth of 1000–3000 m. OPRs disturb ACC transport by triggering an eddy-like dipole perturbation in the upper-layer circulation. For the high-transport (low-transport) cases, OPRs tend to trigger high-to-low (low-to-high) shifts. During the nonlinear evolutions of the OPRs, the perturbed kinetic energy displays a faster growth rate, which is dominated by the baroclinic instability.

(3) The OGIEs leading to the maximum prediction errors of ACC transport are explored relative to the updated reference states triggered by the OPRs. It is revealed that OGIEs exhibit similar features with OPRs including the distributions, the development, and the dynamic mechanisms. The major error region in the OGIEs is located in the middle DP of 58°S–62°S, 72°W–64°W at 3000 m. For the high-to-low (low-to-high) shifts, OGIEs tend to increase (decrease) the predicted transport. That is to say, the OGIEs can damp the sudden-shifting processes as a whole. This process displays the generation and development of an eddy-like dipole error mode, which is dominated by oceanic baroclinic instability. Meanwhile, linear and nonlinear advection processes with opposite effects are verified to maintain the localized evolution of vorticity error by diagnosing the corresponding budget terms.

(4) CNOP-type sensitive regions are determined according to the horizontal distributions of the OGIEs in the application to the targeted observation. The sensitivity experiments indicate that ACC transport is more sensitive to the initial “observation” errors in the CNOP-type sensitive regions (especially for the depth-layer lower 2000 m). Furthermore, by observing system simulation experiments, it is demonstrated that eliminating the initial “observations” errors in the CNOP-type sensitive regions (especially for the depth-layer lower 2000 m) can significantly improve the prediction accuracy of sudden shifts in ACC transport (exceeding 50%).

By exploring the OPR, the OGIE of the sudden shift in ACC transport, and the targeted observation, this study reveals that the baroclinic instability and nonlinear process have significant effects on the shout-range variations of ACC transport. It also demonstrated that optimizing initial conditions by additionally deploying observations in the identified regions can conduce to improve the prediction skill of the sudden shift in ACC transport. The results provide the theoretical guidance for designing the observation systems in the Southern Ocean.