ABSTRACT
Forward scattered sound from the ocean surface is inverted for wave shape during three periods: low wind, mix of wind and swell, and stormy. Derived wave profiles are spatially limited to a Fresnel region at or near the nominal surface specular reflection point. In some cases, the surface wave profiles exhibit unrealistic temporal and spatial properties. To remedy this, the spatial gradient of inverted waves is constrained to a maximum slope of 0.88. Under this global constraint, only surface waves during low wind conditions result in a modeled surface multipath that accurately matches data. The power spectral density of the inverted surface wave field saturates around a frequency of 8 Hz while upward looking SONAR saturates at 1 Hz. Each shows a high frequency spectral slope of -4 that is in agreement with various empirical ocean wave spectra. The improved high frequency resolution provided by the scattering inversion indicates that it is possible to remotely gain information about high frequency components of ocean waves. The inability of the inversion algorithm to determine physically realistic surface waves in periods of high wind indicates that bubbles and out of plane scattering become important in those operating scenarios.
ABSTRACT
The interaction of vhf 100-1000 kHz underwater sound with the ocean surface is explored. The bistatic forward scatter of 300 kHz sound is measured in a wind driven wave channel. Fluctuations in arrival amplitude are described by the scintillation index (SI) which is a measure of arrival intensity variance. SI initially increases with wind speed but eventually saturates to a value of 0.5 when the root-mean-square (rms) roughness is 0.5 mm. An adjusted scintillation index (SI*) is suggested that accounts for the multiple arrivals and properly saturates to a value of 1. Fluctuations in arrival time do not saturate and increase proportionately to the dominant surface wave component. Forward scattering is modeled at frequencies ranging from 50 to 2000 kHz using the Helmholtz-Kirchhoff integral with surface wave realizations derived from wave gauge data. The amplitude and temporal statistics of the simulated scattering agree well with measured data. Intensity saturation occurs at lower wind speeds for higher frequency sound. Both measured and modeled vhf sound is characterized by many surface arrivals at saturation. Doppler shifts associated with wave motion are expected to vary rapidly for vhf sound however further analysis is required.
ABSTRACT
Surface wave shapes are determined by analyzing underwater reflected acoustic signals collected at multiple receivers. The transmitted signals are of nominal frequency 300 kHz and are reflected off surface gravity waves that are paddle-generated in a wave tank. An inverse processing algorithm reconstructs 50 surface wave shapes over a length span of 2.10 m. The inverse scheme uses a broadband forward scattering model based on Kirchhoff's diffraction formula to determine wave shapes. The surface reconstruction algorithm is self-starting in that source and receiver geometry and initial estimates of wave shape are determined from the same acoustic signals used in the inverse processing. A high speed camera provides ground-truth measurements of the surface wave field for comparison with the acoustically derived surface waves. Within Fresnel zone regions the statistical confidence of the inversely optimized surface profile exceeds that of the camera profile. Reconstructed surfaces are accurate to a resolution of about a quarter-wavelength of the acoustic pulse only within Fresnel zones associated with each source and receiver pair. Multiple isolated Fresnel zones from multiple receivers extend the spatial extent of accurate surface reconstruction while overlapping Fresnel zones increase confidence in the optimized profiles there.
ABSTRACT
Surface wave shapes are determined by analyzing underwater reflected acoustic signals. The acoustic signals (of nominal frequency 200 kHz) are forward scattered from the underside of surface waves that are generated in a wave tank and scaled to model smooth ocean swell. An inverse processing algorithm is designed and implemented to reconstruct the surface displacement profiles of the waves over one complete period. The inverse processing uses the surface scattered pulses collected at the receiver, an initial wave profile (two are considered), and a broadband forward scattering model based on Kirchhoff's diffraction formula to iteratively adjust the surface until it is considered optimized or reconstructed. Two physical length scales over which information can be known about the surface are confirmed. An outer length scale, the Fresnel zone surrounding each specular reflection point, is the only region where optimized surfaces resulting from each initial profile converge within a resolution set by the inner length scale, a quarter-wavelength of the acoustic pulse. The statistical confidence of each optimized surface is also highest within a Fresnel zone. Future design considerations are suggested such as an array of receivers that increases the region of surface reconstruction by a factor of 2 to 3.
