Scalable, cloud-native processing of water column sonar data
Scientists commonly use active sonar systems to collect data about mid-trophic level animals like zooplankton and small fish, which play an important role in the marine ecosystems. Echosounders, or fish-finders, are high-frequency sonar systems that emit pulses of sound and record the reflections from animals, the seabed, and other objects. These instruments have been proven to be more efficient and effective for collecting data over a large survey area or a long time period than many other sampling methods, such as underwater imaging and net trawls. This technology has been widely adopted by the ocean science and commercial fishing communities and more recently has been integrated with autonomous vehicles, resulting in a massive amount of data. However, these datasets can be difficult to analyze and are often underutilized. We will address this issue by adopting and advancing data standards, developing a streamlined data processing workflow, and integrating open-source software tools that capitalize on recent advancements in cloud computing technologies to efficiently transform large quantities of ocean sonar data into information that is useful for exploring, monitoring, and managing living marine resources.
Funding agency: NOAA Office of Ocean Exploration and Research FY2021 grants
- Machine learning in fisheries acoustics
- Updates from Echopype developers: changes and roadmap
- EchoPro workflow modernization
- Scalable, interoperable processing of water column sonar data for biological applications using the echopype Python package
- Building a toolbox for studying marine ecology using large ocean sonar datasets