Using ambient noise to extract coherent environmental information

The ambient noise recorded in the ocean has traditionally been treated solely as the noise that a desired acoustic signal must overcome. Nevertheless, it contains environmental information relating to the physical characteristics of the ocean where the noise was recorded. Different types of ambient noise (i.e. diffuse fields vs. independent sources of opportunity) provide the signals for extracting various types of information from the environment. By cross-correlating time-aligned recordings of ambient noise, it is possible to extract time-of- arrival information of the acoustic ray path between the sensors which recorded the data - even with no a priori knowledge of the sound field which contributed to the noise recording. When correlation processing fails due to anisotropy in the ambient noise field, information about the physical environment may still be extracted given trackable sources of opportunity. Due to the nature of the ocean noise field and the fluctuations of the environment itself, difficulties arise in extracting the time-of- arrival information quickly and accurately before environmental changes corrupt the result. The use of arrays of sensors incorporating standard array processing techniques into the noise correlation function is shown to dramatically improve the correlation results both theoretically and when applied to experimental data. In addition, the effects of the directionality of the noise field and the relationship between the directionality of the incoming energy and the signal-to-noise ratio defined from the noise correlation function are examined both theoretically and using ambient noise recordings from experiments. When the ambient noise field is dominated by loud source events, these sources of opportunity may be used directly to invert for environmental characteristics. A model is presented for the use of independent, low frequency, loud sources of opportunity (ice calving events) to invert for the propagating sound speed over very long distances