【作者】 李大炜；

【导师】 李建成；

【作者基本信息】 武汉大学 ， 大地测量学与测量工程， 2013， 博士

【摘要】 海洋潮汐作为时变海面的重要组成部分,是最常见的海洋自然现象之一,与人类的生存环境密切相关,作为一个自然科学问题,已有千百年的研究认识历史。作为一个发展海洋经济的现实问题,一切海洋活动都需要海潮涨落的精确信息。随着空间对地观测技术的发展,海洋潮汐和海洋、大气、陆地水及两极冰盖等负荷效应对固定在地球表面的任何测量都有影响,为了达到各自观测的科学目标,必须建立精密的海洋潮汐模型来有效修正这些影响。过去的二十年里,卫星测高技术的发展和日趋成熟,使我们对全球海洋有了更近一步的认识,它革命性地改变了传统潮汐测量模式,极大地丰富了海洋的观测数据,特别是开阔海域的水位观测数据,使得全球海洋潮汐模型有了重大发展。潮汐分析可以直接从测高海面高观测时间序列中提取潮汐的调和常数或者相应的权参数,然而对于执行精密重复轨道任务的测高卫星来说,周期远远大于作为潮汐主体的半日分潮和主要分潮,势必会导致混叠现象,如何克服混叠效应及其导致的分潮间相关现象是利用卫星测高数据建立海潮模型的首要问题。每颗测高卫星都有自己的优势和局限性,虽然单一卫星或者单一系列卫星所能获取的潮汐信息精度尚可,但是难以满足各研究领域对海潮模型高时空分辨率的需求。随着多源多代卫星卫星测高数据的积累,为了使海潮模型有更高的时空分辨率,联合多种卫星测高数据是研究海洋潮汐的必然趋势。由于受复杂的海洋环境和陆地反射的影响,卫星测高数据在近海海域数据质量普遍较差,在联合多种测高数据的同时,应用数据同化等方法改善全球海洋潮汐模型的近海海域部分也是当前潮汐研究的热点和难点问题。本文联合多种测高数据,从潮汐的混叠效应着手,分析多种不同重复周期的测高卫星中存在的混叠现象,研究不同采样规律对消除和削弱混叠现象的作用,详细分析和比较不同提取方案和潮汐分析方法的优劣性,解决从卫星测高数据提取潮汐信息的若干关键问题,探讨潮汐信号与非潮汐时变信号之间的相关性,在总结确定海潮模型的具体理论和方法的基础上,联合多种测高数据确定全球海洋潮汐经验模型,并利用代表函数展开法同化测高和验潮站数据确定中国海及其邻近海域局部精化海潮模型。具体工作和主要成果如下：(1)从潮汐基本理论的历史发展脉络出发,评述了海洋潮汐研究的现实意义和科学意义。从现代潮汐测量技术和卫星测高技术两个方面评述了全球海洋潮汐模型的发展现状及其应用,进而根据相关地球科学领域的科学目标对潮汐模型精度和分辨率的要求,阐述发展海洋潮汐模型的必要性。(2)阐述了建立海洋潮汐模型所涉及的基础理论,包括平衡潮理论和潮汐动力学理论,以及引潮位的展开和潮汐分析方法,为进一步研究基于卫星测高数据解算潮汐调和常数提供理论依据。(3)从卫星测高的基本原理出发,重点分析了卫星测高重复采样条件下的潮汐混叠现象及其导致的分潮相关性,利用数字分析证明了利用卫星测高地面邻近平行轨迹和交叉轨迹上的相位进动差可以判断卫星分潮间的相关性,特别是对于ERS-1/2, Envisat-1和GFO等卫星,虽然不如T/P的轨道设计具有良好的分辨分潮的能力,但是根据其交叉轨迹相位进动差提供的有效信息,可以判断数据中是否包含了有用的潮汐信息,为联合多种测高数据确定海潮模型提供了理论上的先决条件。(4)详细讨论了不同测高卫星海而高联合数据处理的技术与方法,研究不同测高卫星海面高参考框架的统一技术。在统一处理多种测高数据以及统一参考框架的基础上,构建不同系列卫星的海平面观测时间序列。利用沿轨提取方案对T/P, Jason-1/2系列变轨前20年的观测序列和变轨后6年的观测序列分别进行调和分析和正交响应分析。结果表明,对于较短的观测时段,正交响应分析的结果略优于调和分析。在观测时段满足分离任意两个分潮的会合周期时,两种方法的分析结果具有良好的一致性,8个主要分潮在交叉点位置的振幅内符合精度均优于1cm。同时对交叉点分析和沿轨分析结果进行比较,结果表明,在相同观测时段不同采样规律的条件下,交叉点分析结果明显优于沿轨分析,验证了不规则采样具有良好的去相关作用,求解的潮汐调和常数精度更高。(5) ERS-1/2, Envisat-1卫星能够覆盖±81.5°以内的海域,是对T/P系列卫星很好的补充,但是鉴于太阳同步轨道的卫星测高数据中S2、K1、P1,分潮的不可分辨性,如何扬长避短从其十几年的观测资料中提取高精度的潮汐信息是当前亟需研究的问题。在调和分析方式不适用的条件F,采用调和分析与正交响应分析相结合的方式提取了ERS-2、Envisat-1系列卫星在中国海及邻近海域的潮汐调和常数。结果表明从ERS系列卫星观测时间序列中完全可以提取出与T/P卫星精度相当的太阴分潮,并把K1、P1,从Sa分离出来,把K2从Ssa中分离出来,其中P,与T/P的符合精度优于3cm。与验潮站的比较结果略优于T/P与之比较的结果。(6)结合国内外海潮模型的研究进展和现状,详细地总结和归纳了建立海潮模型的基本理论理论和方法,其中主要包括流体动力学数值解法,联合实际观测资料(验潮站、卫星测高等)与动力学模式的数据同化方法以及经验分析方法,并分析了不同方法的适用性和优缺点。详细研究了变分同化中的伴随法和代表函数展开法,以及经验分析的具体计算步骤。(7)基于精化的多源多代卫星测高数据,利用主要分潮调和常数的空间相关性,建立了15’×15’全球海洋海潮模型WHU12。在开阔海域,其与几个具有代表性的全球海洋潮汐模型符合较好。在浅海海域,精度降低一个量级,但是WHU12在中国海域的表现略好于其他模型。在此基础上,利用代表函数展开法同化T/P,Jason-1/2系列卫星变轨前的20年连续观测资料以及近岸验潮站调和常数建立了7.5’×7.5’的局部模型WHUCS12,更准确地反映了中国近海的潮汐结构。更多还原

【Abstract】 Ocean tides, resulting mainly from the gravitational attractions of the Moon and the Sun, are one of the most fascinating natural events in the world with a significant importance for commerce and science over hundreds of years. In oceanography and geophysics, tides have strong influence on modeling of coastal or continental shelf circulations, play a significant role in climate due to its complex interactions between ocean, atmosphere, and sea ice, dissipate their energy in the ocean and solid Earth, and decelerate the Moon’s mean motion, there is a clear need for more accurate ocean tide predictions. With the rapid development of the technologies of the Earth observation from the space, almost all high accuracy measurements of Geo-science contaminated with the deformations and loading effects due to ocean tides, also require to be de-tided using global ocean tide models in order to reach the scientific goals, respectively. The advent of altimetry has been totally revolutionary; it overcomes the limitation of traditional tide gauge pattern and provides the globally sampled record of sea surface height. Advances in satellite altimetry in the past two decades, especially the launch of Topex/Poseidon satellite and its follow on mission Jason-1/2, can construct a continuous long term sea level change time series and enable numerous scientific studies or discoveries, including improved global ocean tide modeling. The approach (harmonic constants or orthoweights) to tidal analysis based on satellite altimeter data is a direct, efficient and convenient way for ocean tides modeling. However, the altimeter satellites fly over the same location on the ground track once every repeat period. Because the repeated periods are of the order of ten days to a month, far larger than the most energetic tides, such as semi-diurnal and diurnal constituents with frequencies near2and1cycle per day, therefore, this kind of short-period variability clearly cannot be resolved in altimeter observations and consequently aliases to the lower frequencies. The aliasing and the question of how to separate two tides that alias to nearly the same time are the major problems to estimate the ocean tide from satellite altimetry. Owing to the distinct orbit design, the coverage and resolution of each altimeter decided by inclination and exact repeat period (ERP) is too sparse to form a global ocean tide model. All these could be largely improved by combining several kinds of altimeter data, and it is helpful to improve the ability of monitoring sea surface and investigate the water driving dynamic mechanism in shallow water.Based on the status of the present researches on ocean tide modeling, the aim of this thesis is mainly to develop a global ocean tide model with an accuracy of better than2cm in the deep oceans from multi-satellite altimeter data and a more accurate regional tide solution to representer method by means of assimilation altimeter data and tide gauge observations. The main research word and contributions of this dissertation include:1. Starting with the historical context of classical tidal theories, we summarize and review the realistic significance of commerce and science on study of tides. Reviewing the history and present status of satellite altimetry and traditional tidal measure pattern, as well as the global ocean tides modeling from the aspects of the methodology. The necessity of global ocean tides modeling according to principle, the present situation and applications of the existing ocean tide models are also summarized. At last, pointing out the author’s opinions.2. Research on the fundamental theories of tides, including hydrodynamics, hydrostatic equilibrium, expansion of the tide generating potential, and empirical tidal analysis.3. Altimeter satellites are designed to monitor the sea level variation at same ground track every ERP. It is necessary to accomplish an ERP for investigating the time-variant signal, such as tidal current, ocean circulation and any other seasonal cycles. With these repeat orbits, diurnal and semi-diurnal ocean tides are aliased into periods much longer than a day or half a day. For three major satellite altimetry missions until now. i.e. T/P, ERS-2and GFO, the alias periods as well as the Rayleigh periods over which the aliased tides decorrelate have been identified. And prove these correlation problems can largely be solved by the differences of the tidal phase advances on crossing satellite ground tracks.4. Introducing the harmonization in multi-satellite altimeter data pro-processing and the fundamental principle to transform sea surface height from different missions in a certain reference frame and ellipsoids. The response method with orthotide formulation as well as harmonic method has been used to analyze the along track T/P observations, including20years initial orbit data and6years interleaved orbit data. The results show that the most accurate tidal solutions are obtained with the response method, which offers the possibility to infer a number of smaller tides from the dominant tides. The harmonic constants derived from the unevenly sampling pattern at crossover point are better than these from along track.5. The most complete altimeter coverage of the polar oceans is provided by the ERS-1/2and Envisat-1series of satellites. Unfortunately, their sun-synchronous orbits severely limit their usefulness for measuring solar tides. Nonetheless, with a decade-long time series from these satellites now in hand, and with other sun-synchronous satellite altimeters planned for the future, there is a need for detailed studies addressing what can be learned about tides from such data. The present work examines an empirical type analysis of ERS data which combine harmonic method and response method with orthotide formulation, validate the accuracy with the T/P response solution at dual-satellite crossover points and in-situ observation. The result show that the solution derived from the ERS series keep the same level as T/P response solution except the phase lag of solar tides. When we use multi-satellite altimeter data to develop a global ocean tide model, the ability of crossing ground tracks from other satellites could mitigate the decorrelation and aliasing problem.6. Investigating the methodology of ocean tide modeling in detail. Generally, ocean tide models can be classified into three groups, including hydrodynamic model by numerical simulation, assimilation model by data assimilation method and empirical model by direct tidal analysis on altimetry observations. After the launch of T/P, almost all ocean tide models are determined through the data assimilation or via the use of altimeter data in an empirical modeling solution. However, ocean tide model accuracy is still much worse, up to an order of magnitude, in the coastal regions or over partially or permanently sea-ice or ice-shelf covered polar ocean, than that of models in the deep ocean.7. A purely empirical ocean tide models with0.25°×0.25°spatial resolution has been determined using improved multi-satellite altimetry data from TOPEX, Jason-1/2, ERS-2, Envisat, and GFO, based on a Gauss function for weighting inverse proportional to the spherical grid node distance. Compared with contemporary ocean tide models using assessment from tidal constants of tide gauges and from variance reduction studies using crossover discrepancies. Meanwhile, a region ocean tide model with0.125°×0.125°is determined by means of representer method through assimilation of altimeter along track data and tide gauge data. The results show that:the ocean tides model is improved, particularly M2constituents in East China Sea and K1tide in South China Sea, the values of the other constituents are close to the uncertainty expected from the comparison of the global ocean tide models. It is indicated that assessing further improvements in tidal model accuracy will require development of a higher quality validation data set.更多还原