Mr Jeremy Boughenout1, Dr Michael Bunce2, Dr David Harasti3, Dr Megan Huggett1, Dr Darren Southwell1, Dr Michael Stat1
1University Of Newcastle, Newcastle, Australia, 2Minderoo Foundation, Perth, Australia, 3Port Stephens Fisheries Institute, Taylors Beach, Australia
Biography:
Jeremy Boughenout is a PhD candidate in marine and molecular ecology at the University of Newcastle, supervised by Dr Darren Southwell, Dr David Harasti, A/Prof Megan Huggett and A/Prof Michael Stat, in collaboration with the Minderoo Foundation. He asks whether co-occurrence networks built from environmental DNA can reveal which species hold a marine community together, and how that structure shifts across environmental gradients/ under protection. He came to eDNA through his honours, developing metabarcoding assays for seahorses and pipefish. He is drawn to how ecosystems link into something larger, each one shaping the others.
Abstract:
Keystone species play a critical role in maintaining ecosystem structure and function and are therefore important for conservation and management. Traditionally, identifying keystones has required long term ecological studies and/or manipulative experiments, limiting the number of ecosystems in which they can be identified. Environmental DNA (eDNA) may provide an alternative approach, whereby taxa detected in samples are linked into co-occurrence networks based on frequency of recovery with the most connected species identified as potential keystones. However, identification of keystones via co-occurrence networks has rarely been evaluated.
We test the capacity of co-occurrence networks to resolve keystone species using existing eDNA datasets from three tropical island reef ecosystems: Mo'orea, Cocos (Keeling) Islands and Christmas Island. Early results from Mo'orea are promising: the two most connected species in the networks were the keystone grazing fishes Scarus globiceps and Ctenochaetus striatus, both of which have previously been shown to play an important role in preventing shifts from coral-dominated to seaweed-dominated states. The approach also presents limitations. Species that have a low frequency of detection such as sharks, are excluded from the network, despite their known ecological importance. If networks can reliably recover keystones, the approach may provide a powerful and scalable tool for identifying species for conservation and management.