Dr Florian Baletaud1
1CSIRO, Perth/Crawley, Australia
Biography:
Florian Baletaud is a CSIRO CERC Postdoctoral Fellow whose research focuses on understanding ecological processes through innovative monitoring technologies and ecological modelling. He completed a PhD in New Caledonia on tropical fish biodiversity assessment using environmental DNA (eDNA), baited remote underwater video systems (BRUVS), and artificial intelligence for image analysis. He has worked across both the private and public sectors through marine consultancy and postdoctoral positions, including extensive field sampling in remote tropical environments and seamounts. Drawing on these experiences, he aims to translate ecological knowledge into practical tools and evidence that support environmental management and stakeholder decision-making.
Abstract:
Marine biodiversity is undergoing rapid redistribution and reorganisation under the combined influence of climate change, marine heatwaves and increasing human pressures. However, distinguishing climate-driven redistribution from local anthropogenic impacts remains challenging because observations are spatially uneven, taxonomically incomplete, and rarely represent an ecological baseline. This work aims to address this challenge by integrating environmental DNA (eDNA), traditional records, and environmental drivers within a common modelling framework to reconstruct contemporary biodiversity patterns and estimate “counter-factual” biodiversity states.
We use an ensemble modelling method compiling traditional machine learning algorithms (Random Forest, Boosted Regression Trees and Generalized Additive Models) to map species distributions under influence of environmental and human drivers. Biodiversity indices and community assemblages’ metrics are then derived while explicitly accounting for uncertainty and imperfect detection. By isolating response curves and importance of drivers to the assemblage derived biodiversity metrics, we identify scenario-based modification of environmental and human drivers to generate “counter-factual” biodiversity maps. These maps provide estimates of biodiversity under alternative conditions, allowing comparison between observed and expected ecological states across the region. Such comparisons may generalize biodiversity departures that are consistent with historically observed warming trends, marine heatwaves, local human pressures, or interacting effects.
The resulting framework is intended to provide a scalable foundation for biodiversity attribution that will feed the identification of early-warning signals of ecosystem change and eventually ecological forecasting. Ultimately, these outputs will support development of operational biodiversity dashboards capable of translating complex ecological data into actionable information for conservation and management.