Author: Seamus Doherty
Doherty, Seamus, 2025 Constructing network models to predict spatial and temporal shifts in the ecological community of the Australian Wet Tropics, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Predicting how trophic cascades and coextinction influence extinction risk in ecological communities has steadily gained traction over the last few decades, because they are likely to play increasingly important roles under future changing environmental conditions predominantly in freshwater and marine ecosystems. Terrestrial ecosystems have comparatively received less attention given their higher complexity and the dearth of data needed to assess these hidden extinction pathways. Thus, predicting how cascades and coextinctions affect the extinction risk of species in terrestrial ecosystems at fine spatio-temporal scales using empirical data is required, particularly if we are to mitigate such threats. The overarching aim of my thesis was to quantify how anthropogenic threats might influence the extinction risk of vertebrate species in a terrestrial ecosystem at a fine spatio-temporal scale through the processes of trophic cascades and coextinctions. My research specifically examines how climate change may affect the extinction risk of vertebrate species in the Australian Wet Tropics through coextinction and trophic cascades. I developed a novel framework to represent terrestrial food webs more accurately than previous models generating more realistic simulations of trophic cascades under future environmental perturbations. By applying this approach to the Australian Wet Tropics, I provide a comprehensive, ecosystem-wide assessment of extinction risk in terrestrial communities. My research showed that analysing both alpha and beta diversity alongside multidimensional trait space provided a useful way to define terrestrial vertebrate communities across latitudinal and elevational gradients in the Australian Wet Tropics. Although compositional differences appeared, particularly across latitude, the trait-based analyses showed that these communities occupied broadly similar functional spaces. In contrast, distinct differences in both composition and trait space emerged between low- and high-elevational areas. Consequently, these findings suggested that the studied vertebrates largely function as a single community across this bioregion, with marking distinctions emerging only between low- and high-elevation zones based on their functional characteristics. Building on this foundation, I showed how combining such community boundaries with machine-learning algorithms facilitated the construction and analysis of multi-taxa trophic network models. This approach revealed structural differences associated with elevation while highlighting overarching similarities in predator-prey dynamics, and their potential links to extinction risks. Through simulations incorporating basal resources (plants and invertebrates), I demonstrated that bottom-up trophic cascades, particularly in constructed high-elevation networks, could amplify the extinction risk of vertebrate species by pushing ecosystems past tipping points. Coextinction thresholds and resource loss severity emerged as determinant factors in species persistence. Finally, I assessed potential climate-induced coextinctions under multiple future scenarios towards the end of this century, showing that direct climate impacts could rival or exceed coextinction rates, especially in highland vertebrate communities where limited space constrain species’ adaptive potential. These results highlight the need for conservation strategies that extend beyond ad hoc responses and single-species management to account for broader ecological consequences of species loss, ensuring more effective biodiversity protection in an increasingly fragile biosphere.
Keywords: Ecological network models, Trophic cascades, Coextinctions, Climate change, Terrestrial ecosystems, Conservation, Wet Tropics
Subject: Biodiversity and Conservation thesis
Thesis type: Doctor of Philosophy
Completed: 2025
School: College of Science and Engineering
Supervisor: Corey Bradshaw