Machine Learning Model for Assessing Melting Pressure Conditions of the TTG Genesis - Implications for the Timing of the Onset of Global Plate Tectonics

Aug 3, 2026

Upload by Aurelia

Machine learning enables accurate prediction of melting pressures for Archean tonalite-trondhjemite-granodiorite (TTG) gneisses using only whole-rock major element compositions. The developed approach supports batch processing and explicitly incorporates fractional crystallization, mineral segregation, and related petrogenetic effects that commonly hinder conventional pressure estimates. Results indicate two transitions from low to medium pressure recorded by Earth’s oldest rock unit, Acasta gneiss, with high-pressure rocks and melting emerging near 3.8 Ga and becoming globally extensive after 3.5 Ga. Fractional crystallization and source effects suggest true global high-pressure TTGs constitute only about 8%, implying increasing crustal rigidity and enabling widespread high-pressure melting and plate tectonic activity after 3.5 Ga.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Key Points

Abstract

Plain Language Summary

1. Introduction

TTG suite and pressure classifications

Motivation: uncertainties in pressure history

RESEARCH ARTICLE 10.1029/2025GC012503 Key Points: • A new model to accurately predict the formation pressures of Archean TTGs was built based on machine learning • The world's oldest rock unit, the Acasta gneiss, underwent two pressure transitions from low to medium pressures • High‐pressure rocks and melting first appeared at 3.8 Ga, with a global occurrence by 3.5 Ga Supporting Information: Supporting Information may be found in the online version of this article. Correspondence to: J. Liu, [liujia85@zju.edu.cn](liujia85@zju.edu.cn) Citation: Lu, D., Liu, J., Xia, Q., Luan, Z., Zhou, J., Yang, K., & Hanski, E. (2025) . Machine learning model for assessing melting pressure conditions of the TTG genesis: Implications for the timing of the onset of global plate tectonics. Geochemistry, Geophysics, Geosystems, 26, e2025GC012503. [https://doi.org/10.1029/](https://doi.org/10.1029/)[ ](https://doi.org/10.1029/)2025GC012503 Received 16 JUN 2025 Accepted 12 NOV 2025 Author Contributions: Conceptualization: Jia Liu Data curation: Denggang Lu Funding acquisition: Jia Liu Investigation: Denggang Lu, Jia Liu, Qunke Xia, Zhikang Luan, Jingjun Zhou, Kailong Yang Software: Denggang Lu Writing – original draft: Denggang Lu Writing – review & editing: Jia Liu, Qunke Xia, Eero Hanski © 2025 The Author(s) . Geochemistry, Geophysics, Geosystems published by Wiley Periodicals LLC on behalf of American Geophysical Union. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Machine Learning Model for Assessing Melting Pressure Conditions of the TTG Genesis: Implications for the Timing of the Onset of Global Plate Tectonics Denggang Lu1, Jia Liu1 , Qunke Xia1 , Zhikang Luan1 , Jingjun Zhou1 , Kailong Yang1, and Eero Hanski2 1Research Center for Earth and Planetary Material Sciences, School of Earth Sciences, Zhejiang University, Hangzhou, China, 2Oulu Mining School, University of Oulu, Oulu, Finland Abstract Understanding the pressure conditions of the formation of tonalite‐trondhjemite‐granodiorite (TTG) series rocks is important, as these rocks are key constituents of the Archean continental crust. However, the detailed pressure history of Archean TTG gneisses remains ambiguous due to uncertainties surrounding their sources and the effects of fractional crystallization. To address this problem, we developed a machine learning‐ based tool to predict the melting pressures of Archean TTG gneisses from their major element compositions. This method provides high accuracy, supports batch processing, and incorporates the effects of fractional crystallization, mineral segregation, and other factors, making it suitable for application to global data sets. Our results indicate that the Earth's oldest rocks record two key transitions from low to medium pressure conditions, suggesting that the formation of the earliest continental crust may not have required deep subduction‐related geodynamic processes. High‐pressure rocks and melting conditions first emerged around 3.8 Ga (up to 1.5 GPa), becoming more extensive globally after 3.5 Ga (up to 1.9–2.0 GPa) . Fractional crystallization, source, and other petrogenetic factors significantly influence the identification of high‐pressure TTGs, suggesting that true global high‐pressure TTGs may make up only ∼8% of the total rock record, much lower than earlier estimates. Our findings suggest that after 3.5 Ga, increased crustal rigidity likely facilitated widespread high‐ pressure melting and plate tectonic activity across numerous terranes. This research provides new insights into Archean geodynamic processes and demonstrates the power of machine learning in advancing our understanding of ancient Earth systems. Plain Language Summary The melting pressures of Archean rocks provide key insights into the propert

FAQs

What does the machine learning model predict in the TTG genesis study?

It predicts the melting pressure conditions of Archean TTG gneisses from their major element compositions, aiming to recover the pressure history behind TTG formation.

Why is fractional crystallization important for identifying high-pressure TTGs?

Fractional crystallization, mineral segregation, and source-related factors significantly affect how melting pressures are inferred from composition, which can bias the inferred proportion of high-pressure rocks.

What do the results suggest about the timing of high-pressure melting and plate tectonics?

High-pressure rocks and melting first appeared around 3.8 Ga, then expanded globally after about 3.5 Ga, suggesting increased crustal rigidity and enabling widespread high-pressure melting and plate tectonic activity after 3.5 Ga.

Aug 3, 2026

Upload by Aurelia