The race for lipids : ontogeny of the fine-scale vertical co-distribution of arctic calanoid copepods and their phytoplankton food as elucidated by artificial intelligence coupled with an imaging profiler
|Advisor:||Fortier, Louis; Babin, Marcel|
|Abstract:||The grazing of phytoplankton by Arctic copepods channels energy from primary producers to higher trophic levels. However, the predator-prey interactions between phytoplankton and zooplankton in the water column are difficult to study since zooplankton sampling still relies heavily on nets that roughly stratify the water column. The quantification of physiological parameters of copepods, such as lipid content, is also made at coarse vertical resolution. To overcome this limitation, this research used the Lightframe On-sight Keyspecies Investigation (LOKI) system, an underwater camera that provides 1 m vertical resolution. An automatic zooplankton identification model, based on artificial intelligence, was developed for the analysis of profiles sampled in fall 2013 in the North Water Polynya (NOW) and Nares Strait (NS), in the Canadian Arctic. The model turns LOKI images into taxonomic information and can differentiate 114 taxa (organisms and particles), including the developmental stages of copepods. Two studies were carried out based on automatically identified LOKI images. First, during a Lagrangian drift, fine-scale vertical distributions (1-m resolution) of the copepods Calanus hyperboreus, C. glacialis and Metridia longa were studied in relation to their total lipids (TL, mg) and lipid fullness (LF, %). C. hyperboreus and C. glacialis with low LF performed diel vertical migration to surface waters at night to feed, while same-stage individuals with high LF ceased migrating and remained at depth to diapause. Migration to diapause in C. hyperboreus occurred at a LF of approximately 50%, while C. glacialis needed a higher LF (60%). A bioenergetics model showed that Calanus females had enough lipids stored to diapause for over 365 days, highlighting their capacity for capital breeding. In a second study, the fine-scale vertical coupling of C. hyperboreus and C. glacialis developmental stages with their phytoplankton food was studied in the NOW and NS. Three types of copepod vertical distributions in response to the subsurface chlorophyll maximum (SCM) and incident photosynthetic active radiation levels were identified, all of them being in accordance with the predator avoidance hypothesis. At stations where copepod abundances peaked in the SCM, C4 and C5 C. hyperboreus and C. glacialis were vertically partitioned on a fine scale (1-2 m). While C. hyperboreusC4 and C5 abundance peaks were found in the core of the SCM, C. glacialis C4 and C5 peaked just above and below their congeners. The partitioning could be explained by optimal foraging theory or the copepods’ feeding preferences for phytoplankton taxa occupying the SCM. Insight into the fine scale vertical coupling of phyto- and zooplankton is important for a better understanding of climate change effects on the Arctic marine ecosystem.|
|Document Type:||Thèse de doctorat|
|Open Access Date:||24 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.