【雷达与对抗】【2014.06】荷兰人工育滩工程Sand Motor的X波段雷达深度反演模型研制

本文为荷兰代尔夫特理工大学（作者：J. Friedman）的硕士论文，共97页。

一种被称为“Sand Motor”的大型人工育滩工程已经在荷兰沿海实施，以作为一种为未来20年设计的“自然环境”解决方案。鉴于该项目的长期性，目前正在使用现场测量和遥感技术相结合的方法。一部X波段雷达系统部署在Sand Motor，但需要进一步研究其在这样一个动态沿海气候中的适用性。利用三维快速傅立叶变换（FFT）可以将雷达数据处理成波浪、海流、水深等水动力参数。这项技术是非常理想的海岸工程应用，因为它提出了一个相对轻松的方法来捕获高分辨率的时空水动力参数。

本研究的目的是研发一个Sand Motor X波段雷达的深度反演模型，以进一步研究遥感作为估计近岸水深和水动力学的精确工具。该模型应能从高时空分辨率的X波段雷达原始图像中准确估计水动力参数。本文介绍了2013年10月Sand Motor X波段MATLABr拟合（XMFit）模型的建立、标定和验证。XMFit是一种基于现场对比提取近岸水动力的有价值的遥感工具。Nortek B.V.开发的SeaDarQ软件也被用作质量基准。Sand Motor结果表明，与目前可用的SeaDarQ软件相比，XMFit更加健壮和准确。本文完成了灵敏性分析，以进一步分析与XMFit精度相关的时空模式。空间统计表明，雷达区域边缘的误差很大，通过实施2.5公里的空间截止，雷达覆盖的范围减少了。较小的区域使得散射小得多，近似恒定的线性偏差为2m。以下海洋气象条件被发现与XMFit中的高精度周期测量相关：

•有效波高大于1 m

•风速应大于12 m/s

•风和波浪应在45°观测范围内。

海洋气象条件有助于得出以下结论：XMFit需要在k-ω空间中传播光谱，以帮助约束弥散面。这一发现与当地产生的风浪（通常称为风海）直接相关。基于空间和时间限制的理想结果进一步将线性偏差降低到约1.6 m。理想条件表明，雷达反演和原位水深测量与流体力学之间有更好的一致性。涨潮期间XMFit的误差与Sand Motor周围复杂的近岸流体力学有关。注意，Sand Motor处的涨潮显示出复杂的流动结构（即背风面上的分层和大规模涡流形成）。XMFit将这些在涨潮期间明显的复杂三维水流结构平均成一个大的计算立方体，这大大简化了流体力学的分析。最后，这项工作强调需要对XMFit进行更多的研究，因为XMFit被证明适用于荷兰沿岸的Ameland海湾。瞬时结果增加了XMFit的可信度，因为它能够提取复杂的落潮三角洲、洪道方向和相干波诱导流。

A large-scale nourishment known as the SandMotor has been implemented along the Dutch coast as a “Building with Nature”solution designed for the upcoming 20 years. Given the longterm period of theproject, a combination of in situ measurements and remote sensing techniquesare currently in use. An X-band radar system is deployed at the Sand Motor, butrequires further research into its applicability in such a dynamic coastalclimate. Radar data can be processed into hydrodynamic parameters such as waves,currents and bathymetry information through use of a 3D Fast Fourier Transform(FFT). This technology is highly desirable for coastal engineering applicationssince it presents a relatively effortless method to capture high resolutionspatial and temporal hydrodynamic parameters. The objective of this research isto develop an X-band radar depth inversion model at the Sand Motor for furtherinvestigation into remote sensing as an accurate tool for estimating nearshorebathymetry and hydrodynamics. The developed model should be able to accuratelyestimate hydrodynamic parameters from raw X-band radar images with hightemporal and spatial resolution. This thesis explains the development,calibration and validation of the X-band MATLABr Fitting (XMFit) model at theSand Motor for a single storm in October 2013. XMFit proved to be a valuableremote sensing tool for extracting nearshore hydrodynamics based on in situcomparisons. The SeaDarQ software developed by Nortek B.V. is also used as aquality benchmark. The storm results showed that XMFit is more robust andaccurate relative to the currently available SeaDarQ software. A sensitivityanalysis was completed to further analyze the spatial and temporal patternsassociated with XMFit accuracy. Spatial statistics indicated high error aroundthe edges of the radar domain, which led to a reduced radar footprint byimplementing a spatial cutoff of 2.5 km. The smaller domain results in muchless scatter with a near-constant linear bias of 2 m. The following metocean limitswere found to be associated with periods of high accuracy in XMFit.

• Significant wave height greater than 1 m• Wind speed should be greater than 12 m/s • Wind and waves should align within45◦ The metocean limits help conclude that XMFit requires spectra spreading ink − ω space to help constrain the dispersion shell. This finding directly linkswith locally generated wind waves, more commonly referred to as wind sea. Idealresults based on the spatial and temporal limits further reduced the linear biasto approximately 1.6 m. The ideal conditions show a much better agreementbetween radar-derived and in situ bathymetry and hydrodynamics. A relationshipbetween the inaccuracy of XMFit during flood tide was linked to complexnearshore hydrodynamics around the Sand Motor. Note that the flood tide at theSand Motor exhibits complex flow structures (i.e. stratification and largescale eddy formation on the lee side). XMFit averages these complex 3D flowstructures evident during flood tide into a single large computational cube,which drastically simplifies the hydrodynamics. Lastly, this work concludes byemphasizing the need for additional research into XMFit since it provedapplicable at the Ameland inlet along the Dutch coast. Instantaneous results increaseconfidence in XMFit given its ability to extract the complex ebb-tidal delta,the orientation of the flood channel and coherent wave-induced currents.

1. 引言
2. 文献回顾
3. 研究目标
4. 数据库
5. XMFit的研发
6. 理论验证
7. 结果
8. 讨论
9. 结论
10. 研究建议
    附录A XMFit空间校正
    附录B 非线性波理论测试
    附录C 单时间步进比较

更多精彩文章请关注公众号：