A Comparison of Mechanical Wind Filters for Infrasound Sensor Noise Reduction

Abstract

AbstractInfrasound sensors record the ambient acoustic field that contains not only signals of interest but also noise and clutter. Noise is defined as atmospheric turbulence that is incoherent over the distances of meters, whereas, signals of interest and clutter are acoustic pressure waves that are coherent over 10s of meters to 100s of kilometers. There is a growing interest in monitoring sources that extend across the acoustic spectrum from infrasound (below 20 Hz) into the low-end audible acoustic (20–1000 Hz). Monitoring of these extended band signals with a single sensor is made possible with the development of contemporary infrasound sensors, such as Hyperion IFS-3000 with a flat response from 0.01 to 1000 Hz. Combining infrasound sensors with seismometers provides opportunity to better assess noise contributions for both sensor types and improve characterization of sources that occur close to the solid earth–atmosphere boundary. Because sensors are installed to target these broadband acoustic sources, considerations need to be made when selecting a mechanical wind filter to mitigate the noise, while minimizing the impacts to the signals of interest across these frequency ranges. Motivated by these opportunities, this article compares traditional infrasound wind filter designs, that is, porous hoses rosettes and domes, in an urban setting for frequencies 0.01–45 Hz. Data analysis compares the filters, in terms of their response to noise and signals with direct comparisons of wind filters, as a function of frequency. The quantification of performance of these filters in an urban setting provides insight into their effects on detection of sources of interest in this environment.