PhD thesis

Tools’ O the Times: Understanding the common properties of species interaction networks across space

My thesis can be thought of in terms of two acts, which are driven by the central problem that we lack both information about species interactions and have a limited understanding of large scale ‘patterns’. Act 1 focuses on the development of tools that can help us fill in the gaps that we have when it comes to knowing which species interact and Act 2 represents a development of tools that we can use to interrogate networks

My thesis can be thought of in terms of two acts, which are driven by the central problem that we lack both information about species interactions and have a limited understanding of large scale ‘patterns’. Act 1 focuses on the development of tools that can help us fill in the gaps that we have when it comes to knowing which species interact and Act 2 represents a development of tools that we can use to interrogate networks

Publications

Tools for predicting species interactions

  • In Strydom & Catchen et al. (2021), we present a discussion (and roadmap) of both why and how we can use models to help us to predict species interactions as well as networks.

  • In Strydom et al. (2022) and Strydom et al. (2023), we present a an overview of graph embedding and transfer learning as a way to predict species interaction, using the prediction of the Canadian mammalian metaweb as a proof of concept.

Tools for understanding species interactions

  • In Strydom et al. (2021) we showcase the complexity in defining the complexity of networks. Specifically we use SVD entropy as a way to measure the physical complexity of networks.

  • In Strydom & Poisot. (2023), we present the SpatialBoundaries.jl package. Which uses the spatial wombling algorithm to detect areas of change (boundaries) in the landscape.