Multiple sampling methods to develop indices of mid‐trophic levels abundance in open ocean ecosystems. (19th October 2022)

• Record Type: Journal Article
• Authors: Escobar‐Flores, Pablo C.; Décima, Moira; O'Driscoll, Richard L.; Ladroit, Yoann; Roberts, Jim
• Is Part Of: Limnology and oceanography, methods. Volume 20:Number 12(2022)
• Page Start: 789
• Page End: 810
• Publication Date: 2022-10-19
• Subjects: Limnology -- Methodology -- Periodicals; Oceanography -- Methodology -- Periodicals; 551.48
• DOI: 10.1002/lom3.10522
• Languages: English
• ISSNs: 1541-5856

Abstract

Abstract: Mid‐trophic level (MTL) organisms play a key role in the sub‐Antarctic ecosystem food web, linking primary producers and tertiary consumers transferring energy across trophic levels. In this region, the relative abundance of MTLs has been monitored using single‐frequency acoustic data collected opportunistically in a time series of trawl surveys carried out in the austral late spring/early summer. To advance this approach, we developed a methodology that enabled us to disaggregate multifrequency acoustic data into three different MTL groups to assess their temporal and spatial patterns: mesopelagic fish, gelatinous zooplankton (excluding pyrosomes), and euphausiids. We used acoustic data and biological sampling information collected with two midwater trawl nets from a dedicated voyage, to train a classification tree algorithm. We complemented these data with footage from an in‐trawl camera system that gave us new insights on vertical distribution and presence of pelagic taxa. The testing dataset used to implement our classification algorithm was a time series of acoustic data from the sub‐Antarctic trawl survey. This approach was successful at disaggregating acoustic data and provided useful information to study the relative abundance and temporal and spatial patterns of pelagic taxa. Our analysis revealed the presence of dense shallow layers of gelatinous zooplankton which showed a statistically significant increasing trend over the time series. Although we were unable to detect pyrosomes acoustically, we encountered large biovolumes of these organisms during our biological sampling, which was supported by the in‐trawl camera system.