Page 36 - Summer2022

WEIRD DATA IN UNDERWATER ACOUSTICS Dolphins and whales are not the only creatures that can create external interference for acoustic systems, as Emma Cotter of the Pacific Northwest National Labora- tory, Richland, Washington, experienced with a passive acoustic data recorder in the ocean: “When we deployed an autonomous system, the hydrophone data showed periodic sound at low frequencies (<1 kHz) that we couldn’t explain. We initially thought it might be electrical noise or flow noise, but eventually realized it was the result of crabs scraping their carapaces on the metal surfaces of the lander” (personal email, 2022, used with permission). Another common biological noise source is snapping shrimp. When these extremely loud, impulsive signals are present, they can significantly affect the acoustic system signal-to-noise ratio (SNR) in frequency ranges from low kilohertz to 10s of kilohertz. If you ever go scuba diving, you might hear the signal from snapping shrimp as a crackling high-frequency sound near reefs. In addition to biology, the ocean environment itself is a source of natural external noise. Examples include sea ice (Worcester et al., 2020), glacial action (Deane et al., 2019), sediment noise, surface weather and waves, magma dis- placement, and earthquakes. Anthropogenic Noise Sources The actual noise level due to shipping is highly linked to latitude, longitude, and depth relative to strong propa- gation paths from shipping lanes. Worldwide automatic identification system (AIS) maps are available that show historical counts and densities of ship traffic, revealing striking patterns in ship movements. Figure 2 shows an example of one of these maps from 2016 for the waters around Massachusetts using publicly available National Oceanographic and Atmospheric Association (NOAA) data. Unsurprisingly, choosing a sensor location near a shipping channel or ferry route significantly increases the ambient-noise level due to ships (e.g., B and C in Figure 2), whereas a location with only a few ships per year (A and D in Figure 2) will have a much lower noise level due to shipping. The impact of nearby boats on an instantaneous SNR is striking when seen in a spectrogram or beamformed passive acoustic data from an underwater hydrophone array. Figure 3, top and bottom right, shows the results of a boat crossing near an array of recorded acoustic data. Ship noise is complicated because it overwhelms most other signals, is aspect (angle) dependent, and is sub- ject to Doppler shift and multipath effects; this causes an interference pattern that shifts in frequency with range Figure 2. Automatic identification system (AIS)-based ship counts from National Oceanographic and Atmospheric Association (NOAA) for 2016 around Cape Cod, Martha’s Vineyard, and Nantucket Island in Massachusetts. Geographical location in the ocean has a large impact on the ambient-noise level. For example, a sensor located at point A or D would experience far lower noise from ships and boats than a sensor at a location experiencing multiple times daily ferry traffic (point B or C). 36 Acoustics Today • Summer 2022