Hydroacoustic Propagation and Reflection Loss Using Explosions Found in the Indian Ocean

We investigate propagation and reflection loss using hydroacoustic waves from explosion sources in the Indian Ocean recorded at the International Monitoring System’s (IMS) hydrophone stations. We developed a discovery procedure that searches the automatic processing results from the International Data Centre (IDC) for signals that match explosive characteristics. Potential candidate explosions are confirmed by examination of the raw waveform data. Using this procedure, we have identified hundreds of explosions throughout the Indian Ocean Basin. The instigators of the explosions are not known, but the locations and characteristics of the explosions lend themselves to likely candidates. These include military exercises, construction, geophysical exploration, and local fishing practices. We examined hydroacoustic data covering a 6-year period and identified 305 separate explosions. Of these, 25 were recorded at three hydrophone triads – Cape Leeuwin, Australia (H01W), Crozet Island (H04N) and Diego Garcia (H08S) – 79 were recorded at two triads – H04N and H08S – and 201 were recorded at a single triad. In addition, we have observed many airgun surveys, but they are not of interest here due to their local nature. We are able to locate explosions observed at three or two triads using direct arrivals. In some cases, we can locate explosions recorded at a single triad using observed reflections. Absolute arrival times can be estimated to less than 2 seconds, and relative times for closely spaced explosions can often be estimated to within a few samples (4 to 12 milliseconds). Backazimuth uncertainty estimates range from a few tenths of a degree to several degrees depending on the frequency content of the signal and noise conditions at the triad. The explosions are distributed in three general locations: the Arabian Sea off northwestern India, the Bay of Bengal, and along the islands of western Indonesia. The explosions in the Arabian Sea and the Bay of Bengal are likely of military origin, while the western Indonesian explosions are more likely from blast fishing. The Arabian Sea explosions occur near a test firing range off shore from Dwarka, India. The Indonesian explosions occur in clusters that correspond to the islands ranging from Simeulue Island in Aceh Province to the northwest tip of Java. Received signals have energy distributed in frequency starting at 2 to 10 Hz and continuing up to the Nyquist frequency (125 Hz). Peak received levels from the explosions range from 105 to 150 dB rel. 1 μPa. Basic attenuation corrections indicate the yield sizes range from less than one to tens of kilograms. Bubble pulse estimates place a trade-off constraint on source size and depth. The bubble pulse delay time estimates range from 0.1 to 0.5 seconds. The explosions along Indonesia have bubble pulses clustered near 0.2 seconds suggesting that the explosions have similar sizes and depths. Bubble pulse estimates for the same explosion measured at different triads (H04N and H08S) agree to within a few milliseconds. The explosions are located so that paths to different stations intersect different bathymetry. Most have direct deep water paths to at least one of the hydrophone triads. Attenuation estimates can be made by comparing amplitudes between stations. Relative amplitudes between triads for closely spaced explosions are stable (to within 3 dB) and demonstrate the robustness of attenuation estimates. Attenuation is observed to vary for paths that intersect specific bathymetric features, many of which do not reach the Sound Fixing and Ranging (SOFAR) channel axis. This is particularly evident for paths crossing near the Afanasiy-Nikitin Seamount south of Sri Lanka. Small changes in path (within 100 km) are observed to cause attenuation differences of as much as 20 dB. 29th Monitoring Research Review: Ground-Based Nuclear Explosion Monitoring Technologies