Ocean Ambient Noise from Autonomous MERMAID Floats in the South Pacific

Abstract:

Noise in the ocean is generated by the interaction of the
atmosphere, the ocean, and the solid Earth. Noise decreases data
quality from ocean stations by lowering signal-to-noise
ratios. Noise can be used to study Earth's interior as well as
ocean-atmosphere interaction, and to correct the timing of ocean
bottom seismometer arrays where GNSS synchronization is not
available. Understanding ocean noise is furthermore important to
develop better ocean seismometers, and to improve earthquake
signal detections. We present a unique data set of 12 months of
continuous ocean noise recorded in 2018--2019, at 1500~m depth in
the South Pacific by MERMAID, an autonomously drifting
hydrophone. MERMAID is generally not recovered after deployment,
surfacing only to report triggered data segments containing likely
teleseismic events suitable for global tomography. Exceptionally,
one MERMAID instrument was recovered and redeployed, and its
one-year buffer read out for analysis. We apply time-domain,
frequency-domain, and time-frequency analysis to isolate the
background noise from earthquake signals and other kinds of
hydroacoustic transients. We present the temporal evolution of the
noise energy and its spectral density, in the period range
0.05--100~s (frequencies of 0.01~Hz to 20~Hz) over the twelve
months of the recording, focusing on time scales from weekly to
biweekly, monthly, and seasonal. We identify various regimes of
variability in distinct frequency bands, and link them to the
frequency-dependent energy input at the ocean surface as derived
from an ocean wind-wave model. With better understanding of the
ocean noise we hope to improve data quality from ocean stations,
to understand earthquake detection rates, and ultimately to
provide useful information for global Earth imaging.