Statistical protocols for analyzing wind noise levels for site selection

Noise generated at an infrasound sensor by local turbulent pressure fluctuations in the atmosphere, generally referred to as wind noise, is arguably the greatest impediment to infrasound signal detection. In recent years there has been extensive research on the underlying mechanisms for wind noise generation. While considerable progress has been, and continues to be, made there is still no comprehensive method for estimating, from some first principles, wind noise levels at a given site. At present, the best method for determining wind noise levels at potential deployment sites is to place test sensors at the sites and collect data for an extended period of time. Here we present a protocol we have developed for the analysis of such data. The method entails identifying the frequency band relevant to the proposed application and then estimating the appropriate band limited short time rms pressure levels at the test sensors. This generates a time history of rms pressure levels. We then develop a statistical model for the expected band limited noise levels which can be compared to an appropriate threshold model for the determination of site acceptability.