The significance of species groups for food web structure and functioning

• Main Author: Ohlsson, Mikael
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: Linköpings universitet, Ekologisk och miljövetenskaplig modellering 2023
• Subjects: Ecology; Ekologi; Barents Sea
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-198326; https://doi.org/10.3384/9789180753821

In ecosystems across the world, species co-exist, compete, and consume, all while adapting to environmental conditions and disruptions. An important key to the puzzle of understanding how species will respond to changes in the ecosystems, caused by for example climate change, pollution, habitat destruction, and overexploitation, is what current roles species have in a larger context. Species interactions are the basis for many ecological processes, for example, describing who eats whom in food webs. Finding groups of species that have similar interactions can provide insight into what roles species have in a food web, as well as identify core structures and functions of said food webs. Food webs are often based on data aggregates of large areas. Consequently, there is a possibility of blurring local aspects of the food web structure, thus blurring locally realized species roles. In Paper I, I used the group model to analyze local and regional group structures of a food web in the Barents Sea. The group model identifies groups linked to their niche, in which species eat, and are eaten by similar species. I found the large, regional food web diverged from the local group structures, indicating that locally scaled food webs may be required to find more accurately realized species roles. On a local scale, similar group structures were generally spatially clustered and environmentally similar. This was to some extent explained by similarities in species compositions, but more fine-grained patterns related to species identities further impacted the group structures. In essence, the group model is a type of community detection based on stochastic block models. Generated groups contain groups of species with similar sets of prey and predators. Groups are related to both trophic similarity and modularity, but the process itself is, as the name implies, stochastic. Various methodologies to determine a "best fit" group structure out of multiple iterations exist. Arguments can however be made, that discarded, alternative group ...