• Citer

Spatial heterogeneity in Antarctic shallow benthos: Insights from ROV imagery and Bayesian network inference

par Katz, Lea
Président du jury Dellicour, Simon
Promoteur Danis, Bruno ;Mitchell, Emily
Publication Non publié, 2025-12-05

Thèse de doctorat

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

The Western Antarctic Peninsula (WAP) is undergoing some of the most dramatic environmental changes on Earth, with sea-ice regression, extensive glacier retreats, and ice-shelf collapses intensifying the already steep environmental gradients. These changes are affecting entire marine ecosystems, including the local benthos, which hosts the most diverse communities in Antarctica. Therefore, it is crucial to understand what shapes their composition and distribution to anticipate their responses to climate change. Yet, the complex interplay of physical, chemical and biological processes acting at different spatial and temporal scales makes disentangling what drives their distribution particularly challenging, especially with the high heterogeneity of the WAP benthos.This thesis aims to quantify spatial heterogeneity in benthic communities of the WAP and identify environmental and biological factors driving this variability across multiple spatial scales. I combined underwater imagery from a portable Remotely Operated Vehicle with Bayesian Network Inference (BNI) analysis to model statistical dependencies between taxa and their environment.In Chapter 2, I found through a fine-scale survey in a bay of the southern WAP, high heterogeneity within and between sites. Substrate granulometry, depth and distance to the glacier emerged as key environmental structuring factors, while only few taxa, such as starfish, showed comparable network connectivity. BNI demonstrated how local changes in these variables could cascade through the network, highlighting the sensitivity of community structure to subtle abiotic or biotic variation. In Chapter 3, I used multi-scale surveys in the northern WAP to show that heterogeneity appears even at the meter scale. Benthic assemblages were organized into distinct microhabitats not strictly defined by geography or substrate type, but rather by dominant macroalgae species and habitat complexity. These microhabitats represent ecologically meaningful units where early shifts may be detectable before community-level changes occur. In Chapter 4, by integrating BNI across nested spatial scales, I demonstrated that drivers shaping benthic communities are scale dependent. Large scale networks captured broad ecological patterns, while fine-scale networks revealed species-specific associations, predator-prey interactions, abundance-dependent processes. This multiscale approach provided empirical evidence that benthic community organization in the WAP emerges from the interplay of abiotic gradients and biotic dependencies operating at different levels.Altogether, this work refined methodological tools for Antarctic benthic ecology, underlined the critical role of scale in biodiversity assessments, and contributed to understanding how benthic ecosystems may reorganize under ongoing climate-driven environmental change.