EOF重构声速剖面对深水多波束的声速改正分析

声速误差是多波束水深地形测量主要误差源之一,通常采用现场声速剖面测量的方式加以改正,但在深远海多波束水深地形测量时,现场获取全深度的声速剖面并非易事。针对这一问题,利用东南印度洋海洋调查工作中采集到的17个站位的CTD数据,将所有站位声速剖面拓展到全深度,采用经验正交函数分析法(Empirical Orthogonal Functions,EOF)构建调查区声速剖面场,可获得声速剖面场内任意一点的声速值。然后通过EOF重构声速剖面场获得的声速值对测区内多波束水深地形数据进行改正,并与实测声速剖面对多波束水深地形数据的改正结果进行对比,结果表明,5000 m水深范围内2种声速改正结果相差很小,EOF重构法对深水多波束的声速改正满足水深测量的要求。     

XCTD与CTD的对比测试研究

海水声速剖面的准确获取对于利用多波束声呐系统进行水深测量至关重要,而传统的声速剖面获取方式都需要停船进行测量,导致海上调查作业效率较低。为了解决该问题,本文首先介绍了温盐深剖面测量仪(CTD)和抛弃式温盐深剖面测量仪(XCTD)间接测量声速剖面的原理,然后对"海洋地质六号"调查船在同一站位及时间利用CTD、XCTD和AML PLUS SV声速剖面仪测量得到的声速剖面进行了一个对比分析。研究结果表明,三者测量得到的声速剖面在相同水深处声速互差引起的水深差值最大为0.130 9 m。在多波束水深测量过程中,可考虑使用CTD和XCTD间接测量获得的声速剖面代替声速剖面仪直接测得的声速剖面,通过合理布设CTD站位以及使用XCTD来提高海上多波束水深调查的作业效率。     

海水声速直接测量和间接测量结果分析

文中使用了2001～2003年问的海上调查的CTD和声速仪SVP数据,通过直接测量的声速与常用的三种利用CTD计算声速的经验公式所计算的声速剖面进行对比,并采用了相关运算和回归分析,结果表明:在三个经验公式中,Chen-Miller公式与实测声速吻合得最好;表明历史上大量的CTD资料(河口区除外)可直接用来换算声速剖面,进而用于声场预报。     

深远海声速资料对比及校正方法研究

为准确获取深远海海洋声速资料,充分了解深水声速规律,选取了西太地区两个水深超过5000m的S1和S2站位的声速资料为研究对象,以SVP(声速剖面仪)实测资料为参考标准,通过对CTD资料利用Chen-Millero、Del-Grosso以及Wilson的3种经验公式计算的声速与SVP资料进行对比分析,得出Del-Grosso经验公式计算的声速误差最小。为进一步提高声速资料精度,对Del-Grosso公式进行修正,并利用另外3个站位数据进行验证,发现利用校正后的公式计算的声速资料精度明显提升,这为其他深远海区利用CTD或其他温盐深资料获取高精度声速资料提供参考。     

一种提高深远海全深度声速剖面重构精度的方法

开展多波束水深测量应同步进行声速剖面探测。因海上作业条件恶劣、作业时间受限及设备性能局限等影响,在深远海海域常获取不到全深度的实测声速剖面。尽管利用温盐场模型可将声速剖面直接延拓至实地水深的最大深度,但这种气候态平均声速剖面与实际的声速剖面间存在不可控的系统性偏差,会给声速改正及水深测量成果带来质量隐患。给出了一种提高深远海全深度声速剖面重构精度的方法,即利用有效探测深度附近的实测温度盐度值,对大于有效探测深度的各水层的模型温度盐度值施加程度不一的约束控制。结果表明,经优化后全深度声速剖面的重构精度得到明显提高,其中2个XCTD站点声速剖面的互差SSPD分别由-2.5～1.0 m/s优化为0.0～1.0 m/s、0.0～2.6 m/s优化为-1.5～0.0 m/s; 2个CTD站点声速剖面的互差SSPD分别由-0.5～1.7 m/s优化为-0.4～0.3 m/s、-2.15～0.8 m/s优化为-1.4～0.8 m/s。     

联合WOA2018温盐模型及实测温盐资料重构全深度声速剖面(四):声速剖面应用

在深远海海域开展多波束水深测量时,受海上苛刻作业条件等多种影响,获取全深度声速剖面往往比较困难。首先联合WOA2018温盐模型和多个站位CTD、XCTD实测温盐剖面资料开展了全深度声速剖面重构,进而使用三组来源不同的全深度声速剖面开展了多波束测深声速改正对比分析。从试验结果看,这几组声速剖面对多波束测深精度的影响基本一致。特别是当假定CTD站位采用XCTD设备并由此推算深度大于1099m的温盐及声速剖面时,多波束测深的声速改正结果也能满足海底地形成果的质量要求。     

深远海海域多波束水深测量声速剖面获取方法研究

声速是影响多波束勘测精度的重要的外部因素,它决定着声线跟踪的精度,并最终影响到测深精度。由于停船投放CTD时间成本比较高,探索经济高效的远海走航式多波束水深测量,特别是航渡测量期间的声速剖面获取方法成为现场测量人员急需解决的问题。在对HYCOM/WOA13数据与现场CTD数据进行了数据偏差分布、相关性等比对,验证HYCOM/WOA13数据适用性的基础上,提出了基于HYCOM模式数据、WOA13同化数据及单点历史CTD数据与现场XCTD/XBT多源组合的远海走航式多波束水深测量声速剖面获取方法。对比表明,该多源组合的声速剖面能较好反映施测位置的声速剖面情况,该方法对提高远海水深测量的精度和经济效益具有一定的借鉴意义。     

陆缘深水区声速剖面EOF延拓方法研究

针对陆缘深水区多波束测量声速剖面数据采集困难的问题,提出一种深海声速剖面EOF延拓方法,以测区少量全水深声速剖面采样数据为依据,采用EOF算法对未进行全水深采样的声速剖面进行声速全水深延拓。实例计算结果表明:采用该方法延拓得到的全水深声速剖面与实测的声速剖面具有较好的一致性,用延拓后的声速剖面进行声线追踪后得到的水深数据能够满足多波束水深测量精度要求。     

东太平洋秘鲁海盆海域夏季声速特点分析

通过对海洋六号船于2016年12月在东太平洋秘鲁海盆海域获取的10个全深度声速剖面进行分析研究,获取该海域的声速特征,最终认为区内声速剖面具有典型的三层结构特征—表面层、中间跃变层以及深海恒温层。参考Argo数据给出的3类剖面的平均声速结构,分析该海域的声速剖面属于热带型、亚热带南部型混合型。表面层平均厚度约50 m,50~800 m为跃变层,当水深大于800 m以后,声速与水深成线性相关,声速梯度稳定。通过与气象温度数据相对比,表层声速变化与温度成正比,符合表层声速受温度影响大的特点。     

关于测深仪声速改正问题的探讨

用累加法对声速剖面仪和CTD声速剖面仪测量声速剖面资料进行了类比计算，得出两者测量声速具有可替换性的结论，并对声速剖面仪测量声速剖面的时间间隔作了探讨。     

原位测量技术在黄海沉积声学调查中的应用

介绍了最新研制的基于液压驱动贯入的自容式海底沉积声学原位测量系统及其在南黄海中部海底沉积声学调查中的应用。该系统可以实现对海底沉积物声速和声衰减系数进行原位测量,通过液压驱动装置将四根声学探杆匀速贯入到海底沉积物中,减少了对沉积物的扰动,可按照预设的工作参数在海底全自动工作,无需甲板上人员实时控制,采集的声波信号自容式存储于存储单元。系统工作水深为500 m,测量深度为1 m,测量频率为30 kHz,采样频率为10 MHz。使用该系统在南黄海中部获得了40个站位不同类型沉积物的声学特性原位测量数据,并使用CTD剖面仪对该系统声速测量进行了标定,相对误差均小于0.5%,表明该系统测量数据准确、可靠。     

Deep CTD Casts in the Challenger Deep,Mariana Trench

On 1 December 1992, CTD (conductivity-temperature-depth profiler) casts were made at three stations in a north-south section of the Challenger Deep to examine temperature and salinity profiles. The station in the Challenger Deep was located at 11°22.78′ N and 142°34.95′ E, and the CTD cast was made down to 11197 db or 10877 m, 7 m above the bottom by reeling out titanium cable of 10980 m length. The southern station was located at 11° 14.19′ N and 142°34.79′ E, 16.1 km from the central station, where water depth is 9012 m. CTD was lowered to 7014 db or 6872 m. The northern station was located at 11°31.47′ N and 142° 35.30′ E, 15.9 km from the central station, and CTD was lowered to 8536 db or 8336 m, 10 m above the bottom. Below the thermocline, potential temperature decreased monotonously down to 7300–7500 db beyond a sill depth between 5500 m and 6000 m, or between 5597 db and 6112 db, of the trench. Potential temperature increased from 7500 db to the bottom at a constant rate of 0.9 m°C/1000 db. Salinity increased down to 6020–6320 db, and then stayed almost constant down to around 9000 db. From 9500 db to the bottom, salinity increased up to 34.703 psu at 11197 db. Potential density referred to 8000 db increased monotonously down to about 6200 db, and it was almost constant from 6500 db to 9500 db. Potential density increased from 9500 db in accordance with the salinity increase. Geostrophic flows were calculated from the CTD data at three stations. Below an adopted reference level of 3000 db, the flow was westward in the north of Challenger Deep and eastward in the south, which suggests a cyclonic circulation over the Challenger Deep. Sound speed in Challenger Deep was estimated from the CTD data, and a relation among readout depth of the sonic depth recorder, true depth, and pressure was examined.     

声速剖面EOF表示对多波束水深数据的影响研究

基于实测数据,将测区内声速剖面进行EOF表示,进而采用常梯度分层声线跟踪法对EOF表示的声速剖面和实测声速剖面进行比对,统计有效波束比。比对结果表明:采用EOF表示的声速剖面进行的水深数据改正能够满足多波束水深测量的精度指标要求,论证了采用EOF方法表示多波束勘测水深声速剖面场的有效性。     

An Improvement of Long Baseline System Using Particle Swarm Optimization to Optimize Effective Sound Speed

Abstract

In long baseline (LBL) positioning system, errors due to uncertain sound speed are the major facts to its positioning accuracy. In this study, the problem is solved by setting acoustic signal travels between the target and different hydrophones with different sound speed and using particle swarm optimization algorithm to solve the multi-parameter optimization problem to obtain the sound speeds. Presented simulation results show that the proposed algorithm can effectively improve the positioning accuracy of the LBL system compared to existing algorithms and its computational efficiency is high enough.     

基于正交匹配追踪的声层析方法

声速剖面的变化会对声传播产生较大的影响，经验正交函数模型经常用来实现对声速剖面数据的简化描述。然而在内波、湍流等海水不均匀性存在时，这种正则化操作会造成声速重构精度的大幅降低。本文利用字典学习生成声速剖面的非正交原子，在稀疏编码时采用正交匹配追踪（OMP，Orthogonal Matching Pursuit）算法，更新字典则使用KSVD （Kernel Singular Value Decomposition）的字典更新算法。由于字典学习不需要强制使用正交条件，对于训练数据更加灵活，从而可以使用少数的原子组合达到更高的重构精度。利用一次浅海声学实验多次测量的声速剖面研究了海水声速剖面的经验正交函数表示和字典学习，研究表明：相比于正交函数表示，学习字典可以利用少数原子（甚至一个原子）更好的解释声速剖面扰动。字典学习可以提高声速剖面的稀疏性，从而提高声速剖面的反演精度。     

Seasonal variations of sound speed in the Arabian Gulf

Sound-speed computations from CTD casts in the Arabian Gulf during 1992, reveal spatial and temporal variations in acoustic properties. Hydrographic conditions affecting sound speed propagation were seasonally investigated. A monotonic decrease in sound speed profiles with depth was commonly observed at almost all the stations in the Gulf. However, an exception occurred at Hormuz strait during winter. The water exchange pattern between the Gulf of Oman and the Arabian Gulf seems to influence the sound-speed structure, especially in the southern part of the latter. Winter profiles along the Gulf axis showed almost vertically homogenous sound speed. Maximum speeds are observed in summer, with a strong gradient associated with the development of the summer thermocline layer. Horizontal distributions in both winter and summer show a decreasing trend in sound speed from the Strait of Hormuz to the head of the Gulf. The resultant profiles provide a more comprehensive and reliable data set than any that have been reported in the literature. Shallowness and multiple refraction and reflection in the Arabian Gulf may cause sound speed energy to be trapped. No sound channel was detected inside the Gulf. A correlation analysis shows that sound speed is closely correlated with temperature throughout the Gulf, except in winter in the southern half where salinity effects, as a result of inversion and water exchange at the entrance, are found to be dominant.     

模糊ISODATA聚类算法在声速剖面自动分类中的应用

依据中国海浅海区30′按月历史统计声速剖面数据,通过归一化处理和Akima差值采样得到梯度剖面,建立起各方区按月归化后的声速剖面分层梯度样本集,并采用模糊ISODATA聚类算法对声速剖面进行聚类分析.通过对分类结果和类内总方差和的分析表明,聚类参数m值在1.1～2.1之间,并以最远邻系统聚类法结果为初始类中心的模糊分类效果较好.应用该方法对海洋中的声速剖面进行自动分类和区划对海洋环境的战术应用意义重大.