Send the important stuff back: An innovative approach to on-board stream processing for remotely deployed passive acoustic monitoring systems

Real-time monitoring of underwater soundscapes is necessary for rapid assessment of oceanic ecosystem health, controlling anthropogenic acoustic pollution, and establishing multi-scale spatiotemporal variability of the ocean acoustic environment. Such an assessment requires progression from existing regional and specialized passive acoustic monitoring (PAM) observational systems to a large-scale systematic passive acoustic monitoring of the World Ocean, complementary to the ARGO float Integrated Marine Observing System. This research develops an innovative approach to on-board stream processing of acoustic data on underwater autonomous mobile platforms using PyPAM, a Python-based package for calculating hybrid millidecade spectra and statistics. The design optimizes limited computational resources while preserving sufficient information to characterize the sampled soundscape’s noise levels and assemblage of sound sources. The association between frequency-temporal acoustic signatures of biophonic, geophonic, and anthropophonic sounds and corresponding features in percentiles of Spectral Probability Density have been established from PAM data collected at stationary receivers and test deployments of PAM-enabled underwater profilers. Financial support to Joseph Ross was provided by ASA SURIEA. Hardware development was supported by NPS CRUSER program. The PyPAM software is developed by Clea Parcerisas (LifeWatch). The research is conducted at NPS in collaboration with John Ryan (MBARI) and Josh Laney (Seatrec Inc.).