IOCAS-IR
海洋潜沉率拉格朗日计算方法的误差分析及其相关应用
刘凯
学位类型博士
导师王凡
2022-11
学位授予单位中国科学院大学
学位授予地点中国科学院海洋研究所
学位名称理学博士
关键词潜沉率 拉格朗日方法 亚南极模态水 Tasman溢流 示踪物逆时回溯方法
摘要

海洋的潜沉率是通风温跃层理论中的一个重要的物理量,表示的是表层的水团穿过混合层底进入了大洋深层的速率。对潜沉率进行准确的估计有助于我们更好地量化海洋深层水团的生成和输运过程。本文围绕潜沉率进行了三个方面的研究。

首先是对潜沉率的传统拉格朗日计算方法的误差分析与改进。我们指出了该方法中可能存在两种误差:第一种是在潜沉源地的水量较小的情况下,在某些海域中,由于垂向泵压的强度较大,同时混合层的起伏很小,在使用该方法计算时,就会出现计算的潜沉水柱没于第二年局地混合层之下的情况,即计算量大于源头最大可能的潜沉水量,由此会出现对潜沉率的高估。使用Argo观测数据对南半球大洋的潜沉率计算结果分析发现,该误差主要发生在太平洋和大西洋的低纬度海区,对应于热带水团的生成位置。第二种误差则是来自于方法中的Stommel’s Demon假设,即只有深冬季节才能产生有效潜沉,通过对每个月的混合层底的粒子进行追踪,计算了一年中其他月份中产生的有效潜沉,结果显示该误差同样分布在纬度较低的海域,但是互不重合。综合来看,两种误差在南半球大洋的低纬度海域的主要潜沉区存在着对潜沉率高估的情况,而在中纬度,对应于亚南极模态水的形成区域,传统方法计算得到的潜沉率误差则存在低估。在此基础上,我们给出了该方法的改进公式。另外,由于该传统方法使用的垂直速度(w)不够精确,同样会导致潜沉率的计算结果出现偏差。以在亚南极模态水的潜沉区为例,在这里使用垂向泵压得到的w与模式结果相差很大,从而直接带来潜沉率的误差。而使用QG-Omega方程求解得到的w则和模式非常一致,由于观测数据不能直接得到w,该结果表明用QG-Omega方程计算w可以提高潜沉率计算的精度。

然后,我们使用了上述改进的潜沉率计算方法,以长时间序列高分辨率的模式数据研究了亚南极模态水潜沉率在不同位置处的长期变化趋势。结果显示,在1958~2016年间,SAMW(Sub-Antarctic Mode Water)的潜沉量在南太平洋和南印度洋在长时间段上存在着相反的长期变化趋势,即在南太平洋增大,在南印度洋减少,这与已有研究结果相符。但进一步的分析发现,SAMW潜沉量的空间分布存在着明显的差异。在南印度洋,其北部潜沉区的潜沉率有很微弱的上升趋势,而位于南部潜沉区的的潜沉率则有明显的下降趋势。与此同时,在南太平洋中,其西部潜沉区的潜沉率趋势非常小,而东部潜沉区的潜沉率则有明显上升趋势。总体规律是,密度较大的SAMW潜沉水团比密度较小的潜沉水团表现出更显著的长期变化的趋势。其中,南部变化趋势明显的大密度潜沉水量大概占总潜沉水量的60%,可见SAMW的总体趋势更多地来自其南部密度更大的潜沉区的贡献。进一步的分析表明,SAMW潜沉区的混合层深度(MLD)的长期变化趋势与潜沉率的长期变化趋势之间存在较为一致的空间分布。其中,在南太平洋东侧潜沉区的MLD长期增大趋势,主要由于风应力增大的作用,而其西侧潜沉区的MLD长期减小趋势,则主要受到海表浮力强迫的控制;在南印度洋,南侧潜沉区的潜沉率长期减小趋势更多的是受到浮力强迫的影响,而西北部的潜沉率长期增加趋势则主要由风应力增强导致的。

如所周知,潜沉率的长期变化趋势反应了大洋水团和气候变化的相关信息,其信号可通过中深层的翻转环流(大洋传送带)进行传递。例如南太平洋的水团性质可以通过Tasman溢流将信号传递到印度洋,使用拉格朗日质点回溯追踪Tasman溢流水团来源的结果显示,大部分Tasman溢流的水质点可以回溯到南太平洋SAMW潜沉区的混合层,且初始释放位置越深,其跨越海盆到达的经度越靠东,也更难返回混合层内,即使追溯到80年前,仍会有大部分水质点在南太平洋中深层的亚热带环流圈中运动。相应的浓度示踪物回溯实验则显示在回溯80年后,有62%的水回溯进入混合层,且有两个潜沉源地,其中在日界线以西的部分占46.4%,而当追溯起点改到Tasman溢流未转向时的上游之后,追溯到日界线以西的局地海域混合层的比例下降到了26.7%,且大部分位于靠近新西兰岛大陆的南部海域,也属于SAMW的范围之内。该结果确认了Tasman溢流水大部分是由亚南极模态水潜沉的结论。

其他摘要

The subduction rate is an important physical quantity in the ventilated thermocline theory and indicates the volume at which water crosses the bottom of mixed layer and enters the ocean interior. A more accurate estimate of the subduction rate would help us to better quantify the formation of certain water masses. In this paper, three aspects of subduction rates are investigated.

The first is the improvement of the traditional Lagrangian approach to the calculation of the subduction rate, and two factors are pointed out that may lead to deviations: first, there is a limit to the amount of water at the source of subduction, and in some areas of the ocean, due to the increased intensity of vertical pumping and the small fluctuations of the mixed layer, it is easy to occur where the subducted water column is below the local mixed layer in the second year, so that the deviation arises. By using more reliable observations for the Southern Hemisphere oceans, the deviation was confirmed to occur mainly in the lower latitudes of the Pacific and Atlantic, corresponding to the areas of tropical water mass formation. Second, the deviation is from the Stommel's assumption that effective subduction is only occurred in the deep winter season, therefore the effective subduction contributed by the other months was calculated in one year by tracking particles at the bottom of the mixed layer for each month, results show that the deviation is also distributed in the lower latitudes of the ocean. Taken together, the two deviations indicated that there is an overestimation of the subduction rate in the main subduction zone of the southern hemisphere oceans at low latitudes, while at mid-latitudes, the deviation in the subduction rate calculated by the traditional method is smaller. Based on the results, we give an improved formulation of the method. In addition, in the subduction area of subantarctic mode water in mid-latitude waters, the vertical velocity used by the conventional method is not accurate enough, which also leads to deviations in the calculation of the subduction rate, while the vertical velocity obtained by solving the QG omega equation maintains a high degree of similarity to the model output, and since the vertical velocity is often not available from the observed data, the use of this method for the calculation of the vertical velocity improves the accuracy of the subduction rate calculation.

In the following, we investigates the long-term trend of subantarctic mode water subduction rate at different locations with model data of long time coverage and high resolution using the improved method, and the results show that the subduction of SAMW(Sub-Antarctic Mode Water) in the South Pacific and the South Indian Ocean indicate an opposite trend in the time period during 1958-2016, i.e., it increases in the South Pacific and decreases in the South Indian Ocean, which is consistent with the results of existing studies. However, further analysis revealed that there were significant differences in the spatial distribution of SAMW subduction. In the South Indian Ocean, the subduction rate in the northern part of the subduction zone only shows a very weak increasing trend, while the subduction rate in the southern part of the subduction zone shows a significant decreasing trend. Meanwhile, in the South Pacific, the trend of subduction rate in the western subduction zone is very weak, while the water subduction in the eastern subduction zone has a significant increasing long-term trend. Overall, the denser SAMW subducted water masses show more significant long-term trends than the lighter subducted water masses. The subduction water with significant southern variability accounts for roughly 60% of the total subduction water, thus suggesting that the overall trend in SAMW comes more from the contribution of its denser southern subduction zone. Further analysis shows that there is a more consistent spatial distribution between the long-term trends in the mixed layer depth (MLD) of the SAMW subduction zone and the long-term trends in the subduction rate. In the South Pacific, the long-term increasing trend of the MLD in the eastern subduction zone is mainly due to increased wind stress, while the long-term decreasing trend of the MLD in the western subduction zone is mainly controlled by the buoyancy forcing at the sea surface; in the South Indian Ocean, the long-term decreasing trend of the subduction rate in the southern subduction zone is more influenced by buoyancy forcing, while the long-term increasing trend of the subduction rate in the northwest is mainly caused by increased wind stress in the south Indian Ocean.

It is known from the previous analysis that long-term trends in subduction rates reflect information about oceanic water masses and climate change, and their signals can be transmitted through the mid-deep overturning circulation (the oceanic conveyor belt). The nature of water in the South Pacific can be signaled to the Indian Ocean by inter-oceanic water exchange, and the Tasman leakage is an important way. The results of Lagrangian tracing experiment show that most of the particles released at the Tasman leakage location will be captured by the mixed layer in the SAMW subduction zone in the South Pacific, and the deeper the initial release location, the more easterly its longitude across the basin to reach and less likely to enter within the mixed layer, and still move in the deep subtropical circulation in the South Pacific even after the tracing experiment. While the JPL adjoint tracing experiments show two subduction source areas, where 62% of the water is captured by the mixed layer after 95 years of backtracking, with 46.4% of the fraction west of the dateline, while the fraction of the mixed layer traced to local area west of the dateline decreases to 26.7% when the starting point of backtracking is changed to the upstream of the Tasman leakage when it is not turned, and most of it is located near the New Zealand Island This confirms the conclusion that most of the Tasman leakage water is from the subduction of subantarctic mode water.

学科门类理学
语种中文
目录

1 绪论... 1

1.1 研究背景及意义... 1

1.2 围绕潜沉率研究的热点和现状... 5

1.3 目前研究存在的问题... 11

1.4 提出科学问题与主要研究内容... 11

1.4.1 科学问题... 11

1.4.2 主要研究内容... 12

2 数据和方法介绍... 14

2.1 数据资料介绍... 14

2.1.1 ArgoNCEP再分析资料... 14

2.1.2 ECCO模式数据... 14

2.1.3 OFES模式数据... 15

2.2 方法介绍... 15

2.2.1 JPL示踪物浓度追踪方法... 15

2.2.2 潜沉率的传统计算方法... 16

2.2.3 示踪物逆向追踪方法计算潜沉率... 18

2.2.4 计算垂直速度的方法... 18

3 传统方法计算潜沉率的误差... 22

3.1 拉格朗日方法计算的Argo潜沉率的分布... 22

3.1.1 混合层深度的空间分布差异和季节变化... 22

3.1.2 拉格朗日方法计算得到的南半球年潜沉率... 23

3.2 南半球传统潜沉率计算的误差分析... 24

3.2.1 第一种误差:源头水量不足引起的误差... 24

3.2.2 第二种误差:非深冬季节的潜沉过程对潜沉率的补充... 28

3.2.3 两种误差的综合效应... 30

3.3 对两种误差的初步验证... 31

3.4 对两种误差的讨论... 32

3.5 两种误差在北半球中低纬度分布情况的初步分析... 34

3.6 关于垂直速度的讨论... 38

3.7 数据分辨率对潜沉率计算的影响... 43

3.8 本章小结... 46

4 亚南极模态水潜沉率的长期变化趋势... 48

4.1 OFES模式对于潜沉率的模拟... 48

4.2 SAMW潜沉率的长期变化趋势... 51

4.3 对于SAMW潜沉率长期趋势分布的影响因素... 54

4.4 本章小结... 57

5 Tasman溢流水的潜沉区溯源... 57

5.1 Tasman溢流在ECCO模式中的特点... 58

5.2 Tasman溢流潜沉源区的拉格朗日追踪试验... 60

5.3 JPL示踪物逆时追踪方法确定Tasman溢流的潜沉源地... 64

5.4 Tasman溢流的上游回溯试验... 69

5.5 本章小结... 74

6 全文展望与总结... 74

6.1 论文创新点与总结... 74

6.2 未来展望... 76

参考文献... 77

  ... 84

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    85

文献类型学位论文
条目标识符http://ir.qdio.ac.cn/handle/337002/180103
专题中国科学院海洋研究所
推荐引用方式
GB/T 7714
刘凯. 海洋潜沉率拉格朗日计算方法的误差分析及其相关应用[D]. 中国科学院海洋研究所. 中国科学院大学,2022.
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
海洋潜沉率拉格朗日计算方法的误差分析及其(8821KB)学位论文 限制开放CC BY-NC-SA浏览
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[刘凯]的文章
百度学术
百度学术中相似的文章
[刘凯]的文章
必应学术
必应学术中相似的文章
[刘凯]的文章
相关权益政策
暂无数据
收藏/分享
文件名: 海洋潜沉率拉格朗日计算方法的误差分析及其相关应用.pdf
格式: Adobe PDF
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。