Modelling acoustic propagation under-ice in support of AUV missions in Antarctica

This work identifes and investigates acoustic conditions relevant to ship based operation of Autonomous Underwater Vehicles (AUVs) under Antarctic sea ice. The principal objective of the project was to predict the detectability of acoustic status and emergency beacons used by AUVs in an Antarctic sea-ice deployment environment. The key acoustic issues identified were deployment noise, the inuence of water column and sea ice canopy properties on acoustic propagation, and the source and receiver locations. This work contributes to this endeavour in three ways: the first is the measurement of a noise profile of the Australian Antarctic Division's research and resupply vessel the Aurora Australis while parked in Antarctic sea ice; the second is an investigation of the influence of water column properties and sea ice roughness on signal propagation using acoustic modelling and simulation; and the third is the inclusion of field measured Antarctic water and ice data in acoustic models to assess variability and create detectability estimates. Experiments were carried out in the Antarctic Ocean in November 2010 and September- November 2012. The noise profile of the Aurora Australis showed the reduction in deployment noise through shutting down of the main engine which resulted in a 10 dB reduction at 10 kHz when measured below 200 m depth. The influence of a generalised sea ice canopy was assessed with multiple statistically generated simulated ice draft profiles in acoustic modelling tool BELLHOP. Initial simulation using a sound speed profile from Antarctic field trials highlighted the impact of the surface reflected path in regions where defocusing of the direct path was created by the sound speed profile. Further exploration of the Antarctic water and ice environment examined the variability of these media and found that the shape of the sound speed profile was the most significant influence for consistency of direct path propagation. Inhomogeneities in direct path propagation resulting from the characteristics of the sound speed profile led to a greater influence of the surface reflected path than in the case of a linear upward refracting sound speed profile. When modelling acoustic reflections from the underside of the ice canopy the inclusion of statistical surface roughness, obtained from AUV mounted multibeam and drill line measurements, was found to have a signifcant influence on transmission loss and range estimates. Inclusion of the field results in detectability calculations provided a method for combining noise and signal variability into a useful form for estimating safe working ranges. Graphs are presented showing detectability fields for a 10 kHz signal with a probability of detection of 0:5 and probability of false alarm of 10. A range of field measured noise and environmental variables were used to show the variability in detection field. Safe operating ranges were found to vary from 5 km to over 20 km for different ship noise and ice surface conditions. The main contributions of this work are the combination and analysis of Antarctic field measured environmental and AUV deployment conditions with an acoustic propagation modelling tool. A method is provided for estimating safe operating ranges using ice roughness statistics and highlights the importance of realistic sound speed profiles. This thesis presents recommendations for deployment parameters, frequency, receiver and transmitter depths and provides a reference for including ice in a BELLHOP ray/beam acoustic propagation model.