Modelling the structure and seasonal evolution of planktonic food webs in Arctic polynyas.

摘要

Polynyas are areas of open water in ice-covered seas characterized by high biological productivity. Because they have the capacity to take up large quantities of inorganic carbon through photosynthesis they play an important role locally in the sequestration of carbon to depth. To understand the fate of carbon in these high latitude ecosystems, and how the high productivity is sustained, I explore the structure and seasonal evolution of polynya food webs, using multivariate analyses and inverse methods. I first examine the use of inverse models in an ecological context, focusing on the optimization criterion traditionally used to achieve mass balance, i.e. a least squares criterion, and discuss possible variations to the methodology. I then use principal component analyses with missing data to explore temporal and structural patterns among physical, chemical and ecological characteristics of two Arctic polynyas: the NEW (Northeast Water) and the NOW (North Water) polynyas. I further use the temporal patterns identified by the multivariate analyses to establish flow networks for the food web of the NOW polynya, via 1) an inverse model using the least squares optimization, which provides a unique solution for each period and 2) a Markov Chain Monte Carlo (MCMC) inverse model, which provides a representative range of possible solutions for each period and allows the estimation of the probability of these solutions. My results suggest that the NEW polynya food web is strongly structured by physical processes and characterized by a weak seasonal variation in food web structure. On the contrary, the NOW polynya food web is structured more strongly by biological interactions, such as predation by higher trophic levels, and characterized by a strong seasonal variation with differences in transfer efficiency of organic matter to top predators. The food web development in the NOW polynya, however, is not linear, as generally expected from the observation of temperate and tropical blooms. Instead, it is characterized by the interruption of a balanced microbial-metazoan state by a massive bloom, which shifts the food web towards a herbivory state. The combined use of multivariate analyses and inverse methods appear to be a powerful approach to resolving both seasonal-scales differences in food web structure and transfer efficiencies. As such, this approach helps understand the mechanisms structuring planktonic food webs in polynyas, providing important insights into how climate change in the Arctic might impact the pathways of carbon.