Abstract
In order to better understand the trophic interactions in marine ecosystems during the Middle Triassic, this study applies 2D dental microwear analysis to two European microfossil assemblages. The two localities analysed for this project are the Anisian of Winterswijk (The Netherlands) and the Ladinian of Henarejos (Spain). The teeth from the assemblages consist mostly of Chondrichthyes, Osteichthyes, and marine reptiles. Dating from roughly 7 (Winterswijk) to 20 (Henarejos) million years after the Permian–Triassic mass extinction, these assemblages offer valuable insights into the functioning of marine ecosystems after the extinction. The teeth were obtained by acid extraction, sieving and picking under a light microscope.
Dental microwear features were imaged using scanning electron microscopy (SEM). Manually identifying and counting microwear features is labour-intensive and introduces observer bias. To combat these issues, an Artificial Intelligence (AI) model is being developed. The AI model is being trained to identify and quantify 2D microwear features (e.g., pits, scratches, feature size) from the SEM images. The model aims to reduce observer bias and increase the reproducibility of 2D microwear analysis. Where necessary, manual assessment will supplement the AI-derived data. The model will be released as an open-source tool accessible to everyone. By reducing the time investment required for 2D dental microwear analysis, this tool aims to make the method more efficient and widely applicable for the reconstruction of
trophic interactions.
Microwear patterns will be compared across taxa and between Winterswijk and Henarejos. Based on these patterns and other dietary proxies (for example, see Lily Hardeveld Ook Genaamd Kleuver et al. abstract, this volume), diets will be inferred, providing input for the reconstruction of the trophic web. These reconstructed trophic relationships can reveal whether taxa present both in Winterswijk and Henarejos (e.g., Gyrolepis, Saurichthys) occupy similar dietary niches across space and time.
Dental microwear features were imaged using scanning electron microscopy (SEM). Manually identifying and counting microwear features is labour-intensive and introduces observer bias. To combat these issues, an Artificial Intelligence (AI) model is being developed. The AI model is being trained to identify and quantify 2D microwear features (e.g., pits, scratches, feature size) from the SEM images. The model aims to reduce observer bias and increase the reproducibility of 2D microwear analysis. Where necessary, manual assessment will supplement the AI-derived data. The model will be released as an open-source tool accessible to everyone. By reducing the time investment required for 2D dental microwear analysis, this tool aims to make the method more efficient and widely applicable for the reconstruction of
trophic interactions.
Microwear patterns will be compared across taxa and between Winterswijk and Henarejos. Based on these patterns and other dietary proxies (for example, see Lily Hardeveld Ook Genaamd Kleuver et al. abstract, this volume), diets will be inferred, providing input for the reconstruction of the trophic web. These reconstructed trophic relationships can reveal whether taxa present both in Winterswijk and Henarejos (e.g., Gyrolepis, Saurichthys) occupy similar dietary niches across space and time.
| Original language | English |
|---|---|
| Pages | 340-341 |
| Number of pages | 2 |
| Publication status | Published - 2025 |
| Event | XXII Annual Meeting of the European Association of Vertebrate Palaeontologists - Kraków, Poland Duration: 30 Jun 2025 → 5 Jul 2025 Conference number: 22 https://eavp2025.wixsite.com/eavp2025 |
Conference
| Conference | XXII Annual Meeting of the European Association of Vertebrate Palaeontologists |
|---|---|
| Abbreviated title | EAVP 2025 |
| Country/Territory | Poland |
| City | Kraków |
| Period | 30/06/25 → 5/07/25 |
| Internet address |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 14 Life Below Water
Fingerprint
Dive into the research topics of 'Using artificial intelligence for 2D dental microwear analysis of European Middle Triassic marine vertebrates'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver