Game theory: a method to mechanistically model optimal diel vertical migration behaviors of mesopelagic organisms

Diel vertical migration (DVM) is a key feature of pelagic and mesopelagic ecosystems, mainly driven by predator-prey interactions along a time-varying vertical gradient of light. As a consequence, the migration pattern of each organism is intrinsically linked to the patterns of its conspecifics, its prey and its predators, through feedbacks that can be difficult to understand. The presence of low oxygen zones can alter these patterns, by providing a refuge for species adapted to such conditions against predation from species that cannot survive in these waters. In addition to structuring marine food webs, DVM patterns also have important consequences for the biogeochemistry of the world’s oceans. Predators preying at the surface and migrating vertically actively bring carbon to depth, accelerating the rate of the biological pump, and directly connecting surface production with the mesopelagic ecosystem.

We present a trait-based mechanistic model to infer the DVM patterns of different groups of organisms simultaneously, shedding light on the direct effects that different trophic levels can have on each other. We also investigate the influence of oxygen levels on the optimal migration patterns of each group. Based on the traits of each population, the model is parametrized very simply, allowing us to test how different community assemblages and environmental conditions affects DVM patterns. Our model recreates observed patterns in nature, and explains the variability observed across time and space. We compare our results of DVM of a planktonic food web with size-based patterns from the California Current, and we show how oxygen minimum zones can alter DVM patterns of different functional groups of fish. Our results can be used to calculate carbon exports mediated by daily migrants.