Research.ganse.org

From 2008-2015 I participated in the North Pacific Acoustic Laboratory (NPAL), which was a research consortium of multiple universities working together with the Naval Research Laboratory and the Office of Naval Research on problems in long-range ocean acoustics. The focus of our APL group was large-scale statistical data analysis of our measured field data to test the limits of ocean acoustic scattering theories.

Overview
Acoustic propagation over long ranges (hundreds or thousands of kilometers) in the deep ocean becomes progressively randomized due to ocean variability such as internal waves and "spice" (neutrally buoyant soundspeed fluctuations) as well as ambient noise, with a direct bearing on Naval capabilities. The consortium conducts major sea experiments to study these topics, the latest being in the Philippine Sea in 2009 and 2010. The Philippine Sea experiment took place about 600km east of Taiwan and Luzon, in five kilometer deep water. The Scripps Institute of Oceanography (SIO) team placed transceivers in a pentagon layout for ocean acoustic tomography investigations, as well as a five-kilometer-long "distributed vertical line array" (DVLA) of hydrophones covering almost the whole water column. Our APL-UW team operated a low-frequency acoustic source system that we lowered from a ship down to a kilometer's depth, transmitting coded signals across 100km (2009) and 500km (2010) to the DVLA so that we could analyze the effects of the internal waves and spice on those acoustic signals along that path.

Sound in the ocean does not travel in straight lines, for the speed of sound in the ocean is not constant. Soundspeed in the ocean is governed mainly by the water's temperature and pressure. In the upper kilometer or so the temperature gradually drops with depth so the soundspeed does as well. But at depths below that top kilometer, the temperature remains fairly constant (it's just plain cold!) so the increasing pressure with depth becomes the dominant effect and causes the soundspeed to start increasing with depth again. Thus there is a soundspeed minimum at about one kilometer's depth in the world's oceans, and this has a "waveguide" effect on the sound paths -- the sound waves are guided up and down and up and down again along that 1km deep axis very efficiently for thousands of miles (for low frequencies). It is worth noting that it is not an unbearably loud source that allows the sound to be heard at such great distances, it is the amazingly high efficiency of the waveguide propagation and lack of acoustic attenuation in seawater that accomplishes this. Still, marine permits had to be obtained each time for the source levels that we did use, for they still had to be loud enough to keep the signal up above the background ambient noise, which is caused predominantly by commercial shipping.

The sound splits up into multiple paths within the waveguide as it travels to the receiver array, at which the pings in the multiple paths come in at different times. It's almost analogous to multiple echoes coming in from walls at different distances, except that here there are no reflections (as with the walls). It is the splitting nature of the waveguide -- differing refraction for differing take-off directions from the source -- that causes the multiple arrivals. The timings of these multiple arrivals are different depending on receiver depth, and when you plot them all together you get a cool accordion-shaped pattern (seen here in real data received on the 5km long vertical array of receivers).

The Navy has taken advantage of these waveguide features of ocean sound propagation since WWII, but the picture is complicated by the fluctuations of internal waves, spice, and ambient noise, which all continue to make it difficult to accurately predict the received fluctuations in ocean acoustic signals. Understanding and better modeling these fluctuations (so they can always be distinguished from targets like submarines) is a longtime interest of the Navy, who funds this work. Yet there are other benefits too -- the oceans are affected by global warming just as are the continents, and sound measurements can be used to study long-term temperature variations in the ocean. Also, "T-phase" energy from oceanic earthquakes gets into that sound axis and can be detected from far away, making these ocean acoustic methods a key tool in studying these earthquakes. In any event, the axis causes sound to split into multiple paths in the ocean, causing the distinctive "accordion" shaped pattern in hydrophone depth vs arrival time plots. This accordion and parts within it shudder and shimmer over time due to the ocean fluctuations, and we use these shudders and shimmers to learn about the ocean fluctuations.