Journal of

GEOsciences

  (Formerly Journal of the Czech Geological Society)

Original paper

Namık Aysal, Fatma Şişman Tükel

Thermobarometry of Clinopyroxene, Amphibole and Biotite for the Oligo-Miocene Plutons (NW Türkiye): Comparison of Conventional and AI-Based Methods

Journal of Geosciences, volume 70 (2025), issue 4, 199 - 209

DOI: http://doi.org/10.3190/jgeosci.405

AI-based algorithms have increasingly found applications across various fields of geology, emerging as robust alternatives to conventional thermobarometric calibration strategies. In this study, the compatibility, distinctions, and reliability of these novel AI-driven approaches were assessed through a comparative analysis with conventional methods. Focusing on three representative case-study samples from the Oligo-Miocene plutons in northwestern Turkey, the pressure-temperature (P-T) conditions governing clinopyroxene, amphibole, and biotite crystallization were evaluated.
The comparative analysis revealed significant methodological divergences. Conventional clinopyroxene barometers yielded widely scattered and often petrologically unrealistic pressure estimates (frequently >10 kbar), whereas AI-based algorithms provided tightly clustered results consistent with shallow-to-mid crustal emplacement (2-4 kbar). For hydrous phases, while conventional methods indicated only shallow crystallization depths (<2 kbar), AI-based models for amphibole and biotite consistently revealed deeper, mid-crustal crystallization conditions (~5-6 kbar). These findings suggest that the studied plutons underwent polybaric crystallization, a history that is effectively captured by AI algorithms but partially obscured by conventional calibrations. Consequently, while AI-based approaches offer superior resolution for complex magmatic systems, the observed discrepancies highlight the necessity of integrating whole-rock geochemistry with mineral chemistry for robust geological interpretations.