Page 26 - WINTER2019

Utilizing CTBT Hydroacoustic Stations
Figure 7. Time-frequency spectrograms of the 2017 San Juan impulse measured at the mid-Atlantic station HA10, having propagated within the SOFAR channel (left), and at the southern Indian Ocean station HA04, having propagated through a polar environment (right).
a slower propagation speed in the polar profile) as the fre- quency increases.
A comparison of the two signals makes clear the task at hand, which is accurately “picking” signal features and associating them to the correct propagation speed. Although multiple isochrons can be formed from the modal arrivals, triangula- tion really benefits from forming isochrons along different propagation paths (triangulation does not work when three listeners are at the same location). In triangulation of the San Juan with the two hydroacoustic stations, additional iso- chrons are formed from three-dimensional (3-D) arrivals that had propagated off the geodesic or shortest path.
Triangulating the ARA San Juan
At low frequencies, relevant to the CTBT network, variability
in modal propagation speed depends primarily on the ocean bottom topography (bathymetry). As sound propagates near sea mounts, islands, or the continental shelf, interaction with the bottom steepens the angle of a mode. This steepening causes sound waves to turn or refract away from (but depend- ing on oceanography sometime toward) bathymetric features (Munk and Zachariasen, 1991). As sound energy travels up the slope and approaches the apex (the depth where a mode no longer fits), it is translating along the slope analogous to a
ball rolling up an inclined plane at an angle. The sound turns and then travels back down, out toward deeper water.
Bottom topography in the vicinity of the San Juan was conducive to bathymetric refraction, and a plethora of 3-D arrivals were received at the mid-Atlantic IMS station (HA10). One of these arrivals had considerable amplitude and corresponded to energy refracted by the continental slope. The propagating wave front of mode-1 gets folded by the refraction process (its apex occurs at a depth of ~200 m), and from this, an additional isochron is formed based on that path (Dall’Osto, 2019).
Considering these two mode-1 arrivals and the arrival along the southern path, we have the three necessary for triangulation. After computing propagation paths and tabu- lating the propagation speed along each path, isochrons are formed perpendicular to the propagation paths. The iso- chron intersection occurs within a few kilometers of the actual location (see Figure 8). What remains that causes the error or mismatch between the actual and triangulated locations are inaccuracies in picking out the timing of the mode arrival and an ill-constrained estimate of the climate state of the ocean. Through inclusion of additional arrivals, these errors can be resolved.
26 | Acoustics Today | Winter 2019