Abstract
Distributed Acoustic Sensing (DAS) is revolutionizing passive acoustic monitoring, offering the potential to create vast arrays of uniformly spaced virtual hydrophones every few meters. Applications are expanding rapidly and now include marine mammal monitoring, which often requires higher (>1 kHz) bandwidths. DAS is unlike traditional hydrophones since the sensitivity is determined by a complex stress-to-strain conversion from acoustic pressure into extensional strain in the fiber with a directional response. Furthermore, the nature of the extended physical aperture over which strain rates are measured by the laser pulses injected into the fiber creates a frequency-dependent directivity, both from the spatial weighting of the laser pulse autocorrelation function and from the “gauge length” over which strain rate estimates are averaged to suppress noise. We develop a simple theoretical model and apply finite element numerical modeling to account for these phenomena, generating predicted sensitivity curves as a function of the acoustic frequency and incident angle to the fiber, accounting for interrogator self-noise. These models are compared to experimental observations of several thousand received acoustic signal pulses over a large range of angles at frequencies up to 3 kHz by an OptoDAS interrogator connected to a fiber-optic cable in the Trondheim fjord, Norway.