Vertical Migration: Structure and function of pelagic ecosystems

Abstract

Diel Vertical Migration (DVM) is the daily movement of marine organisms as diverse as zooplankton, fish and squids between the depths and the surface of the oceans. DVM is primarily an antipredator response, organisms typically hiding from visual predation in deep waters during daytime before ascending to feed in food-rich surface waters at night. DVM is such a widespread phenomenon that it is believed to be the largest natural daily movement of biomass of the planet (only potentially exceeded by human commuters). Behind this apparent simplicity, a vast diversity of patterns can be observed in the oceans. The day residency depth, the night residency depth, the speed and timing of ascent and descent all vary according to the environmental conditions considered. In addition, DVM being an anti-predator response, trophic interactions and the food-web structure have important consequences on the realised migration patterns. Consequently, prey and predator can react to each other and adapt their behaviours simultaneously. Most studies modelling DVM usually consider varying environmental conditions (e.g. light levels) but keep predator and prey fields constant, omitting the cascading effects that changes in the DVM patterns of one population can have on other populations. The first aim of this PhD thesis was to consider explicitly these feedbacks to understand mechanistically the drivers of DVM patterns (especially day and night residency depths). We merged notions of game theory, trait-based approach and classic ecology to reproduce DVM patterns observed in nature. Each organism is seen as a player that tries to optimize its fitness. Its fitness depends on its traits, but also on its own behaviour and on the behaviours of its prey and predators. Organisms modify their vertical position in the water column, exposing themselves to different light levels, i.e. different growth opportunities and mortality risks. We compute simultaneously the optimal behaviour of all organisms by finding the Nash equilibrium of the system – the point at which no organism can get a better fitness by changing unilaterally its behaviour. We first apply this method to a system consisting of a zooplankton prey and its visual predator, and investigate how changes in predator traits impact prey (and predators) DVM patterns. Then, we apply this method to a size-based zooplankton food-web. We divide the zooplankton community in 100 sub-populations based on their size and feeding modes and recreate size-based non-linear DVM patterns observed in the oceans. Further, as a feature modifying food intake and mortality risks, DVM can potentially impact population dynamics. We investigate the consequences of behaviour at the individual time scale on population dynamics (taking place at a larger time scale). We show that considering the adaptive behaviour of prey and predator affects population dynamics, but also ecosystem functions such as trophic transfer efficiency and active carbon export. Indeed, if DVM is a fascinating feature by itself, it is also important globally as it has cascading consequences for marine ecosystem and the global carbon cycle. Global marine phytoplankton biomass represents “only” about 3 PgC (compared with 60 PgC for terrestrial vegetation), but they both fix carbon at a comparable rate (60PgC/yr for terrestrial vegetation vs. approx. 50 PgC/yr for marine phytoplankton). Of these 50 PgC/yr, between 5 and 11 PgC/yr are transported below the euphotic zone, either through sinking (dead organisms, fecal material) or active transport of vertically migrating organisms. This carbon export rate is crucial for the global carbon cycle as carbon reaching the ocean depths is sequestrated for extended periods, consequently helping to reduce atmospheric CO2 concentrations. It is estimated that active transport of carbon by organisms performing DVM represents approx. 15% of this total. However, this estimation is mainly for migrating zooplankton, and the exact global contribution of other functional groups such as fish is blurry. In the final part of this thesis, we develop a model of DVM for different functional groups of zooplankton and fish that provides us with global DVM patterns. This enables us to disentangle the relative contribution of each functional group to the carbon export flux below the euphotic zone, as well as the relative importance of DVM in the carbon export mediated by each functional group. We couple these outputs with a global ocean circulation model and generate global estimates of carbon sequestrated through the different pathways of the active biological pump. We estimate that higher trophic levels (mesozooplankton and up) sequestrate around 751 PgC.