Software and idea sharing for the IRA-EDA Active Acoustics Strategic Initiative
Revolutionize and expand opportunities for active acoustic data in ecosystem-based fisheries science and management through a NOAA-wide program that advances analytical methods for multi-disciplinary data collected from innovative platforms.
Add IRA & EDA stuff here
Active acoustics here refers to multi-frequency and wide-bandwidth echosounders that span frequency spectra from kiloHertz to megaHertz and data-processing and analytical software used to assess fish and zooplankton populations for fisheries research and management. Echosounders transmit pulses of sound into the ocean and record frequency-dependent echoes from marine organisms that are used to classify the echoes to taxa or species, map their distributions, and estimate abundances. These classifications are performed by manual or automated echo-data segmentation and attribution and may include ancillary information to enhance their accuracy. Development of these approaches are often stifled by large data volumes, required improvements in processing emerging sources of wide-bandwidth and low-frequency acoustic data, and uncertainty in covariation with habitat factors.
NOAA collects active acoustic data from an increasingly diverse number of crewed and uncrewed platforms such as Fisheries Survey Vessels, autonomous surface and underwater vehicles, moorings, and landers. Compared to data from multi-frequency echosounders, wide-bandwidth echosounder data are much more voluminous, providing higher resolution frequency response information but also presenting challenges for data storage and analysis. Artificial intelligence (AI), machine learning (ML), and statistical techniques comprise advanced analytical methods that are needed to significantly accelerate automated echo classification through the combination of high-resolution echosounder, biological, and habitat data.
The principal goal of this NOAA-wide program is to advance active acoustic data acquisition, storage, and analytical capacity at all Fisheries Science Centers (FSCs) to enable accurate automated classification of echoes from most or all assessed species and their predators and prey, as well as to develop novel survey designs for stock and ecosystem assessment purposes. The analytical tools will be applicable to multi-disciplinary data sent from collection platforms to a cloud server, or directly aboard the data collection platforms. These tools will greatly expedite data processing, improve classification accuracy, expand applicability of acoustic data, promote adaptive sampling, and facilitate further and faster collaborative development. The proposed activities will be conducted on a national level to leverage resources and expertise across FSCs, NOAA, academic institutions and commercial entities and to ensure parity in expertise and support among all FSCs.
Revolutionizing echo classification (Activity A) is our primary goal, and the collection of activities contribute substantially to achieving this. Compiling validated data sets (Activity B) and developing cloud infrastructure (Activity C) are critical to the success of this SI. Ensuring high-bandwidth capacity at sea (Activity D) and testing and implementing the echo classification advancements using novel data collection opportunities (Activity E) are required for routine inclusion in stock assessment and ecosystem-based management. At the $8M and $6M levels of funding, critical activities are under or unfunded, severely diminishing the impact that this SI will have on NOAA’s data acquisition and analysis capabilities, capacity, and breadth toward climate-ready fisheries. A substantial portion of our budget is allocated to cloud infrastructure. Reducing or eliminating costs associated with cloud infrastructure and data up/downloads through partnerships with NOAA OCIO and OMAO will contribute significantly to funding our other activities.