Integrating predictive modelling and molecular approaches to investigate deep-sea Antipatharia biodiversity and distribution along the Northeast Atlantic margin

Black corals (Hexacorallia: Antipatharia) are ecosystem engineers that play an important role in sustaining biodiversity, as they provide other invertebrates with shelter and food, and can create gardens used as nursery areas by other species. Black coral gardens are Vulnerable Marine Ecosystems (VMEs), and are subject to a variety of threats caused by the intensification of human activities in deep waters. Species distribution modelling is a valuable ecological tool that can be used to inform conservation management on the presence of VMEs and VME indicator species, but it mostly relies on imagery data, and image-based identification of species is uncertain if not supported by traditional or molecular taxonomy. In the northeast Atlantic Margin, a territory characterised by oceanographic and geomorphological heterogeneity, only a few studies have focused on the biodiversity and distribution of black corals, and no attempts have been made to inform image-based identification through reliable taxonomic approaches. This project combines presence-only species distribution modelling and molecular species delimitation to investigate the biodiversity and distribution of deep-sea Antipatharia along the northeast Atlantic margin. The predicted distribution of the most occurring morphospecies was modelled in Chapter II using two top-performing algorithms, MaxEnt and Gradient Boosting of Regression Trees, and with or without selecting variables prior to modelling. To help quantify the generalisation capability of the models, the novel Average TSS Index was designed. MaxEnt models outperformed those built using GBM, and variable preselection did not improve model performance. Only two MaxEnt models for Stichopathes gravieri were satisfactory, and thus were used to delineate the environmental niche of the coral. In Chapter III, the models were re-built with more occurrence points that added environmental heterogeneity, and were tested on independent data. These revised models identified suitable niches for one morphospecies of Stichopathes (m1) and for Leiopathes, but tended to underperform in underexplored regions. Species delimitation analyses in Chapter IV were based on three mitochondrial markers and two nuclear markers. These analyses identified that some of the morphospecies whose models had underperformed were not coherent with molecularly delimited species. This highlights the usefulness of supporting image-based identification with information on molecular variation and DNA barcodes. The species delimitation study consistently revealed the presence of fourteen species across the margin, likely underestimating the biodiversity of black corals, particularly in the family Antipathidae. Future species delimitation studies should consider the use of next-generation sequencing approaches to better identify species boundaries, and future image-based distribution studies should consider the integration of molecular delimitation analyses on target species to ensure environmental niches are correctly delineated. Future modelling approaches should consider modelling to smaller spatial extent to improve model performance.

Funder

University of Galway
Research Ireland

Publisher

University of Galway

Rights

CC BY-NC-ND

cbn

Collections

University of Galway Theses (PhD Theses)