Kenneth M. Arnoult

Explosion localization via infrasound

Two acoustic source localization techniques were applied to infrasonic data and their relative performance was assessed. The standard approach for low-frequency localization uses an ensemble of small arrays to separately estimate far-field source bearings, resulting in a solution from the various back azimuths. This method was compared to one developed by the authors that treats the smaller subarrays as a single, meta-array. In numerical simulation and a field experiment, the latter technique was found to provide improved localization precision everywhere in the vicinity of a 3-km-aperture meta-array, often by an order of magnitude.

Discrimination of near-field infrasound sources based on time-difference of arrival information

A computationally efficient method for discriminating between near- and far-field infrasound sources using array time-difference of arrival (TDOA) information is described. Rather than assess wave-front curvature, the discriminant quantifies the statistical departure of TDOA information from that of a plane wave passing the array. Since the method constrains neither the functional form nor the amplitude characteristics of a signal it is suited for discrimination of signals across large-aperture infrasound arrays. Experimental results confirm theoretical predictions to a range of order ten array apertures. The discriminant is applied to data from an Antarctic infrasound array.

Locating explosions, volcanoes, and more with infrasound

Sound with frequencies below the human threshold of hearing has long been used for global monitoring of large acoustic outbursts such as those in nuclear tests. Now infrasound is helping scientists detect much smaller events at close range.

Performance of an infrasound source localization algorithm

We present a performance analysis of a method of acoustic source localization based on time difference of arrival (TDOA) information for an arbitrary array of sensors. The method begins with the construction of a vector containing estimates of time delays for each unique sensor pair in the array via cross correlation. An optimization in the space of geographic location and sound speed is then conducted to minimize the difference between the observed vector and one calculated from a search in that space. The technique uses an analytic method of seeding based on the acoustic analog of a light cone. The source localization procedure is tested as a function of a number of parameters, including array aperture, number of sensors, sample rate, signal‐to‐noise ratio, and signal type. The technique is shown in application to real and synthetic infrasound signals and its statistical behavior is given.

A method for locating nearby sources

A fast method for locating nearby sources based solely on the time‐of‐flight information is described. The time delays for a signal passing between all sensor pairs is determined using cross‐correlations and a vector of time delays is constructed. In the noise free case, where the speed of sound is constant, each point in the plane is associated with a unique value of the time delay vector. The method we present uses a fast, simplex‐based search of the plane for a source location by minimizing the difference between the vector associated with the candidate location on the plane and the value estimated from cross‐correlations. The search takes place over the three dimensional space that includes two coordinates in the plane and the propagation speed. The starting point for the search is constructed from an analytic fit of circular wave fronts to groups of sensors within the array. This method has been useful in identifying near‐field sources in the 153US and 155US CTBT/IMS infrasound arrays. Examples of th...