Bathymetry Estimates on Open Beaches and in Tidal Inlets via Remote Sensing

Abstract:

Despite its vital importance to coastal hydrodynamics and maritime safety, bathymetry is often unknown or poorly constrained; remote sensing techniques show promise to help fill this gap. It has been shown that spatial gradients of cross-spectral phase can be successfully exploited to estimate highly-resolved, frequency-dependent wavenumber vectors in wave resolving image time series. These frequency-wavenumber vector pairs have proven useful for bathymetry estimation not only through basic inversion of the linear dispersion relation, but also through more complex data assimilation schemes. However, previous efforts are generally limited to open beaches.

As part of the DARLA and RIVET I/II experiments (Duck, NC; New River Inlet, NC; Columbia River Mouth, OR/WA), we present application of this methodology to two complementary remote sensors (X-band marine radar and optical video) and to a range of environments, from simple (open beach) to complex (a micro-tidal inlet mouth and an energetic meso-tidal estuary mouth). Bathymetry estimates via linear wave theory (the “cBathy” algorithm), spatially resolved on the order of one wavelength, are shown to reproduce bar gaps along open beaches and the structure of ebb-tidal shoals. These quasi-continuous estimates are provided at significantly greater resolution than those calculated via 3D-FFT methods over similar domain sizes (up to 40+ km²).

Mean currents can affect depth inversion estimates through the Doppler shift of the wave field. As the effect scales with the current vector component in the direction of the waves, this term is often negligible on open beaches, but it becomes significant near tidal inlets. Tidal averaging of the Doppler effect induces a depth estimate bias, but it does not appreciably deteriorate depth estimate accuracy in the presence of low-amplitude, oscillatory tidal currents. Additionally, if slack-water events can be identified, the problem can be divided into separate depth- and current-inversions.