The effects of sound propagation and avoidance behaviour on naval sonar levels received by cetaceans
In: http://hdl.handle.net/10023/3194
The use of active sonar is deemed to be essential for naval operations, but its potential impact on marine life has raised concerns worldwide. In a risk-assessment framework, characterisation of risk of harm is accomplished by combining exposure assessment and dose−response relationships. The overall topic of this thesis is an evaluation of factors that influence exposure assessment, including analysis of how sound levels received by cetaceans are affected by in-situ sound propagation and the influence of diving, movement and possible avoidance behaviour of the whales themselves. Data from an international research programme based on controlled exposure experiments (CEEs) were available for this study. During these experiments, low-frequency active sonar (LFAS: 1-2 kHz band) and mid-frequency active sonar (MFAS: 6-7 kHz band) signals were recorded by suction-cup tags attached to killer whales, long-finned pilot whales and sperm whales, and by a hydrophone array towed near the whales. Chapter two describes how the sonar signals recorded by these systems were quantified, and investigates the influences of range, depth and propagation conditions on the received sound levels. Chapter three focuses upon the effect of simulated vertical and horizontal exposure-avoidance strategies of whales in response to an approaching source on the received sound levels. A total of 7,091 sonar signals were analysed from the towed-array (2,794) and tag (4,297) recordings. Transmission loss (TL) and excess attenuation (EA) from a simple 20log(range) model were compared among species, signal types and acoustic receivers. TLs followed expected geometric spreading versus range and TL coefficients were 15.5−20.1 for LFAS and 18.8−23.6 for MFAS. One experiment where levels on the animal-attached tag were attenuated due to 'body shielding' (when the animal's body is interposed between the sound recording tag and the sound source) was documented, and other sources of variation in received level dataset were discussed. Variations in EA with depth were consistent with TL patterns predicted using the acoustic propagation model Bellhop with the highest EAs occurring near the sea surface. The effect of depth on EA was clearest in killer and pilot whale experiments which occurred at locations with stronger gradients in the sound-speed profile, while sperm whale experiments in deeper homogenous offshore waters showed little influence of depth on EA. The results indicate that a simple TL model like 20log(range)+absorption does not accurately predict attenuation levels over the distances (0.1−11.1 km) from a sonar source to a freely-diving animal, but that the overall patterns of TL can be fairly well explained using sound propagation models that take into account local environmental conditions. A consistent different in TL between LFAS and MFAS signals was not explained by the Bellhop model, however, indicating that unidentified sources of variation do influence the sonar signals recorded on freely-diving whales. To evaluate the potential effect of avoidance strategies of whales on received sound levels, whale positions were simulated with a Monte Carlo method in which a simulated source vessel directly approached the whale. The cumulative sound exposure level (SELcum) received by the whales was estimated using the Bellhop model. Horizontally-stationary animals received the highest levels. The optimal course in terms of reducing SELcum for animals moving in a straight line was 100° from the heading of the source vessel, while 120−130° was optimal for animals dynamically moving relative to the position of the source. Moving horizontally in the optimal direction away from the vessel path yielded 9−17 dB reduction of SELcum and vertical avoidance led to reductions of up to 10 dB in certain circumstances. Actual observations of the whales during the sonar experiments indicated that animals often move sideways out of the path of the approaching vessel, close to the optimal angle predicted. The simulation approach is therefore potentially useful to predict how whales react to an approaching sound source. This type of analysis may also be useful to understand the patterns of cetacean strandings relative to the movement of sonar-transmitting military vessels.