Measuring underwater ambient noise: problems and pitfalls

Ambient noise is a key factor in determining the performance of underwater acoustic systems. Ambient noise is that sound received by an omni-directional sensor which is not from the sensor itself or the manner in which it is mounted. Noise from the sensor or its mounting is termed self-noise. Ambient noise is made up of contributions from many sources, both natural and anthropogenic. These sounds combine to give the continuum of noise against which all acoustic receivers have to detect the signals they are looking for. Ambient noise is deceptively easy to measure, just deploy a hydrophone, make a recording, transform into the frequency domain, and you then have the level and spectral characteristics of ambient noise. Unfortunately this is rarely adequate to fully characterise ambient noise and further difficulties may be introduced by the way the hydrophone is deployed. Ambient noise is made up of a number of contributions, each of which can vary in a random and/or cyclic manner. The cyclic variation may have a period corresponding to the tidal, diurnal, weekly, lunar, monthly or annual cycles. For any one application it is important to match the measurement time to the relevant cyclic variations. It is also important to consider how the raw acoustic data can be processed to reduce the information to manageable levels while still retaining the important factors. Noise radiated from a discrete source is called radiated noise and is measured in a similar manner to ambient noise. Measurements are usually made close to the noise source so that the radiated noise is sufficiently stronger than the ambient noise to ensure that it is the radiated noise being measured. Radiated noise can be highly directional so it is important to make a series of measurements around the source. If the characteristics of the individual contributions to ambient noise, including radiated noise from individual sources, are understood then it is possible to model the ambient noise field. For many applications this can considerably reduce the need for detailed, and expensive, field measurements. Making the ambient noise measurements can be hampered by poor measurement procedure. Care must be exercised to minimise the self-noise of the measurement equipment and the platform from which it is deployed. It is also important to consider the depth of the measurement sensor and any directional characteristics of both the sensor and the noise field. This paper will review how ambient noise can be measured, the length of time for which the measurements need to be made, and the equipment that should be used to collect the data. It will also consider how the data should be processed to obtain meaningful, but manageable, information.