The influence of thresholding passive bubble-acoustic signals on the quantification of physical effects

Passive emissions of sound by bubbles provide potentially useful data in many natural and engineered systems. Bubble-acoustic data are used to predict the severity of volcanic eruptions, identify undersea gas seeps, and measure ocean wave breaking. In these natural cases, and in many chemical engineering and metallurgical processes, statistics on the bubble size are sought; the bubble size controls the rate of gas to liquid mass transfer as well as fluid dynamical aspects such as mixing. However, owing to uncertainty in the physics generating the acoustic amplitude, bubble-size determination remain essentially empirical, requiring an threshold to select data. The impact of various thresholding and data-windowing techniques is discussed. Examples are given on the data from a chemical engineering aerator, a plunging jet, and a case in which diagnostic statistics can be used to identify the threshold giving the optimum sensitivity and specificity for detecting the breaking of wind-waves. Once found, the optimum threshold can be used to infer relevant physics from the acoustic signal alone.