Page 45 - Summer 2021

2020), but for the purposes of imaging the deep Earth, they are unwanted. Last, the ocean is full of sound (Dahl et al., 2007) that also must be filtered out. Examples are ship propellers, volcanic eruptions, glacial calving, rain, mysterious plane crashes, and, of course, sounds of bio- logical origin. Adaptations targeting rather than avoiding such sources are straightforward. Acoustic packages tai- lor-made for whale census research (Matsumoto et al., 2013) and meteorology (Riser et al., 2008) are part of MERMAID’s extended family.
The Second Coming
MERMAID’s breakthrough came in the form of the second- generation model (Hello et al., 2011) that reported data via
Figure 3. The first-generation MERMAID prototype: an oceanographic profiling float equipped with an externally mounted hydrophone, and a recording and processing unit. Reproduced from Simons et al., 2009, with permission.
Figure 4. Seismic events detected at a 700 m depth offshore of San Diego, CA, during the recovered deployments of MERMAID-001. a: Only the large (magnitude [mag] 6) event is from a distant source 5,170 km from the epicenter as measured along the surface of the Earth and is thereby suited to map seismic wave speed variations in the Earth’s mantle. b: a nearby (144 km epicentral distance) smaller (magnitude 2.7) crustal earthquake. Left column: filtered pressure records aligned on the first-arriving seismic “P” wave. Right column: Fourier spectrograms. These reveal how earthquake propagation over long distances filters out the high frequencies. Pressure conversions due to the close-by earthquake remain prominent in the range of 0-10 Hz, whereas the distant quake registers as a relatively small blip of power around 0.5 Hz. MERMAID’s onboard processing unit recognizes these differences and is now tuned to preferentially report large earthquakes. The horizontal bands of energy are caused by nonseismic oceanic acoustic noise. Reproduced from Simons et al., 2009, with permission.
global seismology (Simons et al., 2009). Several smaller magnitude events were, strictly speaking, bycatch and not useful for deep-Earth imaging. They are, however, suitable for assessments of local and regional seismicity and to study crustal structure. The difference in char- acter between global and local seismicity (Figure 4) is apparent from the records themselves. Figure 4 shows the time-domain records, centered on the arrival of the earthquake’s compressional wave marked “P” for “pri- mary,” and their Fourier spectrograms.
Every type of earthquake (large or small, deep or shallow, distant or close-by) has a distinct acoustic fingerprint. Efficient signal-processing techniques onboard MER- MAID reduce the streaming time-domain samples to an evolving “bar code” whose character reveals which records are due to the types of waves most useful for whole-Earth wave speed imaging. Acoustic “T” waves from very shallow sources, for example, carry over long distances, bouncing around in the Sound Fixing and Ranging (SOFAR) low-velocity channel of the ocean itself. Recently, these have been used for water thermom- etry over long spatial and temporal baselines (Wu et al.,
Summer 2021 • Acoustics Today 45