Benthic ecosystem cascade effects in Antarctica using Bayesian network inference

Abstract

Antarctic sea-floor communities are unique, and more closely resemble those of the Palaeozoic than equivalent contemporary habitats. However, comparatively little is known about the processes that structure these communities or how they might respond to anthropogenic change. In order to investigate likely consequences of a decline or removal of key taxa on community dynamics we use Bayesian network inference to reconstruct ecological networks and infer changes of taxon removal. Here we show that sponges have the greatest influence on the dynamics of the Antarctic benthos. When we removed sponges from the network, the abundances of all major taxa reduced by a mean of 42%, significantly more than changes of substrate. To our knowledge, this study is the first to demonstrate the cascade effects of removing key ecosystem structuring organisms from statistical analyses of Antarctica data and demonstrates the importance of considering the community dynamics when planning ecosystem management.

Bayesian Network Analysis reveals resilience of the jellyfish Aurelia aurita to an Irish Sea regime shift

“…Inference of historical jellyfish abundance. One of the most powerful aspects of BNs is the ability to make inferences of how the probability of one node (taxa or physical variable) being in each state (zero, low or high for a taxon) is likely to change given that another node is in a given state (zero, low or high for a taxon) 38 . This inference is made by calculating the probability of node A being in a given state given node B is in a given state.…”

“…Correlations between species can be purely trophic, in which case the network represents a food web 36 , or can include other sorts of ecological interactions such as facilitation or competition for resources 37 . Including physical variables such as temperature or—in the aquatic realm—depth, also enables mutual habitat associations to be found 38 . Such multi-process networks can be reconstructed statistically using methods such as Discrete Bayesian Network Inference Algorithms (BNIAs) which can find network structures, including non-linear dependencies between nodes.…”

“…Such multi-process networks can be reconstructed statistically using methods such as Discrete Bayesian Network Inference Algorithms (BNIAs) which can find network structures, including non-linear dependencies between nodes. BNIAs have primarily been used to calculate gene regulatory and neural information-flow networks 39 , 40 , but more recently have been applied successfully to reconstruct ecological 35 , 41 , palaeontological 42 and—in the marine realm—to abyssal plain 43 , Antarctic 38 and deep-sea benthic networks 44 . To the best of our knowledge this study is the first application of BNIAs to a pelagic ecosystem.…”

Mortality, Population and Community Dynamics of the Glass Sponge Dominated Community “The Forest of the Weird” From the Ridge Seamount, Johnston Atoll, Pacific Ocean

“…Deep-sea sponges create these biodiversity hotspots by providing biogenic structures which increase vertical habitat complexity, providing substrate and refugia for macroinvertebrates and demersal fish (Dayton, 1972;Dayton et al, 2013;Kazanidis et al, 2016;Maldonado et al, 2017;Dunham et al, 2018;Meyer et al, 2019;Vieira et al, 2020). Sponges also provide a key link between benthic and pelagic systems, by pumping and filtering large quantities of water (Reiswig, 1974;Bell, 2008), and so increase diversity beyond providing a hard substrate (Mitchell et al, 2020b). Environmental settings have a strong influence on benthic community composition and density across multiple different scales, from global latitudinal and depth gradients to the kilometer and meter scale.…”

“…Bayesian networks are probabilistic models that show causal relationships between variables, whereby different variables are the network nodes; and where a dependency exists between two nodes, this is depicted as an edge (Heckerman et al, 1995). BNIs can infer network structures and non-linear interactions, and have been used extensively to reveal gene regulatory networks (Yu et al, 2002), neural information flow networks and ecological networks (Smith et al, 2006;Milns et al, 2010;Mitchell et al, 2020a), paleontological communities (Mitchell and Butterfield, 2018) and more recently using these networks to infer likely changes for benthic systems (Mitchell et al, 2020b). Note that the Bayesian network found reflects the associations caused by co-localizations rather than a specific association or interaction, which is why SPPA is needed to then infer the most likely underlying process.…”