Switching off the sun to observe the twilight zone spatial dynamics across Saint-Paul and New-Amsterdam Islands, Southern Indian Ocean [résumé]
Résumé
Information on micronekton (> 1 cm organisms) is globally scarce in the open ocean, and its ver-tical and horizontal distribution in relation to oceanographic structures is poorly known. The complex biodiversity composing micronektonic functional groups lead to even more challenging interpretations of their spatial dynamics. Advanced generations of echosounders emit simulta-neously several acoustic signals (multi-frequency device), allowing a finer view of the micron-ektonic community. While data becomes more abundant and complex, it is crucial to develop statistical tools aiming to objectively extract key components of its variability. In this study, we analyse data recorded onboard the R/V Marion Dufresne from an EK80 echosounder emitting at 18, 38, 70, 120 and 200 kHz. We developed a Multivariate Functional Data Analysis method to identify patterns in micronekton structures across Saint-Paul and New Amsterdam economic exclusive zone, at the boundary between vast oceanic domains. This approach proposes an ob-jective method to analyse the vertical backscatter distribution and quantify temporal and spatial modes of variability in multivariate acoustic data. By filtering the temporal mode, we uncovered a latitudinal acoustic pattern in concordance with hydrological features and biological samples distribution. Such methods could be implemented at a global or local scale and allow 3-D mod-elling of micronekton structuring.