EUVIMEDEUVIMED entdecken
Crossref

Using Bayesian network inference and underwater imagery to understand the influence of environmental heterogeneities on benthic community structure in the Antarctic Peninsula

Lea Katz, Tasnuva Ming Khan, Camille Moreau, Emily Mitchell, Bruno Danis

Polar Biology · 2025-09 · 48 (3)

DOI: 10.1007/s00300-025-03407-4 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Polar Biology, online erschienen am 28.07.2025. Verfasst von Lea Katz und 4 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 60 Literaturangaben. Gefördert von: Fonds De La Recherche Scientifique - FNRS, Cambridge Trust, Newnham College, University of Cambridge.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The Western Antarctic Peninsula (WAP) is one of Earth’s fastest-warming regions. Increasing temperatures and winds are disrupting sea-ice dynamics, which can impact benthic communities. Local heterogeneity in these benthic communities enhances biodiversity and provides redundancy, critical for ecosystem resilience. Therefore, understanding local heterogeneity and its primary drivers is essential to understand how benthic communities may shift under climate change. To assess local heterogeneity, we combined underwater imagery with a Bayesian network approach. Video sampling was done using a portable Remotely Operated Vehicle across four sites in Dodman Island, a small bay (1 km2) within the WAP. The dataset comprised 0.8 m2 photographs taken within 10 m2 areas, located within 50 m from each other to form a site. All data was collected within one week during Antarctic summer. We annotated seafloor images for morphotaxa and substrate type, then used Bayesian Network Inference (BNI) to infer an ecological network of dependencies between taxa and key environmental variables. Our results indicate high small-scale heterogeneity in benthic communities between sites, both in taxonomical and functional groups. In our Bayesian network, substrate granulometry, maximum depth, and distance to the glacier were the most connected environmental variables; only starfish showed similar connectivity among taxa. Using BNI, we inferred how changes propagate through the network via direct or cascading interactions. Substrate granulometry had a strong impact on all taxa, while macroalgal abundance changes had relatively weak impact. We demonstrate how BNI, coupled with non-destructive sampling, can support predictions of benthic change and inform conservation priorities.

Förderung

  • Fonds De La Recherche Scientifique - FNRS
  • Cambridge Trust
  • Newnham College, University of Cambridge
  • Natural Environment Research Council

Zitieren

APA

Katz, L., Khan, T. M., Moreau, C., Mitchell, E., & Danis, B. (2025). Using Bayesian network inference and underwater imagery to understand the influence of environmental heterogeneities on benthic community structure in the Antarctic Peninsula. Polar Biology, 48(3). https://doi.org/10.1007/s00300-025-03407-4

Vancouver

Katz L, Khan TM, Moreau C, Mitchell E, Danis B. Using Bayesian network inference and underwater imagery to understand the influence of environmental heterogeneities on benthic community structure in the Antarctic Peninsula. Polar Biology. 2025;48(3). doi:10.1007/s00300-025-03407-4

BibTeX
@article{Katz2025,
  author = {Katz, Lea and Khan, Tasnuva Ming and Moreau, Camille and Mitchell, Emily and Danis, Bruno},
  title = {Using Bayesian network inference and underwater imagery to understand the influence of environmental heterogeneities on benthic community structure in the Antarctic Peninsula},
  journal = {Polar Biology},
  year = {2025},
  volume = {48},
  number = {3},
  doi = {10.1007/s00300-025-03407-4}
}
Kostenlos starten →