Development of an X-band Radar Depth Inversion Model at the Sand Motor

Development of an X-band Radar Depth Inversion Model at the Sand Motor

Friedman, J.

Stive, M.J.F. (mentor)
Reniers, A.J.H.M. (mentor)
Luijendijk, A.P. (mentor)
Hoekstra, R. (mentor)
Swinkels, C.M. (mentor)
Radermacher, M. (mentor)

Civil Engineering and Geosciences

Hydraulic Engineering

2014-06-11

A large-scale nourishment known as the Sand Motor has been implemented along the Dutch coast as a Building with Nature solution designed for the upcoming 20 years. Given the long-term period of the project, a combination of in situ measurements and remote sensing techniques are currently in use. An X-band radar system is deployed at the Sand Motor, but requires further research into its applicability in such a dynamic coastal climate. 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 applications since it presents a relatively effortless method to capture high resolution spatial and temporal hydrodynamic parameters. The objective of this research is to develop an X-band radar depth inversion model at the Sand Motor for further investigation into remote sensing as an accurate tool for estimating nearshore hydrodynamics. The developed model should be able to accurately estimate hydrodynamic parameters from raw X-band radar images with high temporal and spatial resolution. This thesis explains the development, calibration and validation of the X-band MATLAB Fitting (XMFit) model at the Sand Motor for a single storm in October 2013. XMFit proved to be a valuable remote sensing tool for extracting nearshore hydrodynamics based on in situ comparisons. The SeaDarQ software developed by Nortek B.V. is also used as a quality benchmark. The storm results showed that the XMFit is more robust and accurate relative to the currently available SeaDarQ software. A sensitivity analysis was completed to further analyze the spatial and temporal patterns associated with XMFit accuracy. Spatial statistics indicated high error around the edges of the radar domain, which led to a reduced radar footprint by implementing a spatial cutoff of 2.5 km. The smaller domain results in much less scatter with a near-constant linear bias of 2 m. High accuracy in XMFit is directly linked with waves exceeding 1 m, wind speeds exceeding 12 m/s and the alignment of wind and waves. The metocean limits help conclude that XMFit requires spectra spreading in wavelength-frequency space to help constrain the dispersion shell. This finding directly links with locally generated wind waves, more commonly referred to as wind sea. Ideal results based on the spatial and temporal limits further reduced the linear bias to approximately 1.6 m. The ideal conditions show a much better agreement between radar-derived and in situ bathymetry and hydrodynamics. A relationship between the inaccuracy of XMFit during flood tide was linked to complex nearshore hydrodynamics around the Sand Motor. Note that the flood tide at the Sand Motor exhibits complex flow structures (i.e. stratification and large scale eddy formation on the lee side). XMFit averages these complex 3D flow structures evident during flood tide into a single large computational cube, which drastically simplifies the hydrodynamics. This work concludes by emphasizing the need for additional research into XMFit since it proved applicable at the Ameland inlet along the Dutch coast. Instantaneous results increase confidence in XMFit given its ability to extract the complex ebb-tidal delta, the orientation of the flood channel and coherent wave-induced currents. This work proves that XMFit is a valuable model for extracting high resolution spatial and temporal hydrodynamic parameters.

depth inversion
Sand Motor
remote sensing
X-band radar

http://resolver.tudelft.nl/uuid:2d4773bd-50e5-4507-ae64-1b7c333fb7c1

2016-06-11

52.052285, 4.184252

Student theses

master thesis

(c) 2014 Friedman, J.