{"title":"使用机器学习约束重新考察太古宙地面地壳的成分演化","authors":"Junteng Lyu, Ziyi Guo, Ming Tang","doi":"10.1016/j.chemgeo.2025.122986","DOIUrl":null,"url":null,"abstract":"<div><div>The composition of the Archean subaerial crust, whether mafic or felsic, remains debated due to tectonic uncertainties and proxy limitations. Here, we employ Random Forest Regression algorithm-based machine learning model to analyze multi-element geochemical data from terrigenous fine-grained sedimentary rocks to reconstruct the MgO content in the subaerial crust. The model is trained using major oxides (TiO<sub>2</sub> and Al<sub>2</sub>O<sub>3</sub>) and trace elements (Sc, Cr, Co, Ni, Cu, Zn, Y, Zr, Nb and REEs) of simulated mixture samples from Archean craton igneous rocks and post-Archean igneous rocks. The validity of this model is demonstrated by using various modern samples to predict MgO content in the modern subaerial crust, whose MgO content has been well constrained. We also demonstrate that the accuracy of the predictions is not sensitive to tectonic settings. We then use this model to reconstruct the chemical evolution of the subaerial crust over Earth's history and compare the differences between our model and other proxies reflecting the composition of the subaerial crust. Our model results suggest that a systematic decline of MgO content in the subaerial crust (7–10 wt% to 3–4 wt%) and a shift from mafic (70–100 % komatiite-basalt) to felsic dominance (∼70–90 % TTG) of the subaerial crust in the Neoarchean period, after which the subaerial crust attained a similar composition to that of the modern subaerial crust. The results from our approach are similar to those from Ni/Co and Cr/Zn proxies, but differ from the results of Zr/TiO<sub>2</sub> and Al<sub>2</sub>O<sub>3</sub>/TiO<sub>2</sub> proxies. We suggest that many previously used composition proxies are sensitive to tectonic settings and applying these proxies to terrigenous sediments can lead to erroneous results on the subaerial crust composition in the Archean. In addition, we caution against the use of proxies whose correlations with MgO or SiO<sub>2</sub> content are highly nonlinear, which may result in large uncertainties in the projected MgO or SiO<sub>2</sub> content for the source rocks.</div></div>","PeriodicalId":9847,"journal":{"name":"Chemical Geology","volume":"693 ","pages":"Article 122986"},"PeriodicalIF":3.6000,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Compositional evolution of the Archean subaerial crust revisited using machine learning constraints\",\"authors\":\"Junteng Lyu, Ziyi Guo, Ming Tang\",\"doi\":\"10.1016/j.chemgeo.2025.122986\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>The composition of the Archean subaerial crust, whether mafic or felsic, remains debated due to tectonic uncertainties and proxy limitations. Here, we employ Random Forest Regression algorithm-based machine learning model to analyze multi-element geochemical data from terrigenous fine-grained sedimentary rocks to reconstruct the MgO content in the subaerial crust. The model is trained using major oxides (TiO<sub>2</sub> and Al<sub>2</sub>O<sub>3</sub>) and trace elements (Sc, Cr, Co, Ni, Cu, Zn, Y, Zr, Nb and REEs) of simulated mixture samples from Archean craton igneous rocks and post-Archean igneous rocks. The validity of this model is demonstrated by using various modern samples to predict MgO content in the modern subaerial crust, whose MgO content has been well constrained. We also demonstrate that the accuracy of the predictions is not sensitive to tectonic settings. We then use this model to reconstruct the chemical evolution of the subaerial crust over Earth's history and compare the differences between our model and other proxies reflecting the composition of the subaerial crust. Our model results suggest that a systematic decline of MgO content in the subaerial crust (7–10 wt% to 3–4 wt%) and a shift from mafic (70–100 % komatiite-basalt) to felsic dominance (∼70–90 % TTG) of the subaerial crust in the Neoarchean period, after which the subaerial crust attained a similar composition to that of the modern subaerial crust. The results from our approach are similar to those from Ni/Co and Cr/Zn proxies, but differ from the results of Zr/TiO<sub>2</sub> and Al<sub>2</sub>O<sub>3</sub>/TiO<sub>2</sub> proxies. We suggest that many previously used composition proxies are sensitive to tectonic settings and applying these proxies to terrigenous sediments can lead to erroneous results on the subaerial crust composition in the Archean. In addition, we caution against the use of proxies whose correlations with MgO or SiO<sub>2</sub> content are highly nonlinear, which may result in large uncertainties in the projected MgO or SiO<sub>2</sub> content for the source rocks.</div></div>\",\"PeriodicalId\":9847,\"journal\":{\"name\":\"Chemical Geology\",\"volume\":\"693 \",\"pages\":\"Article 122986\"},\"PeriodicalIF\":3.6000,\"publicationDate\":\"2025-07-26\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Chemical Geology\",\"FirstCategoryId\":\"89\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0009254125003766\",\"RegionNum\":2,\"RegionCategory\":\"地球科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"GEOCHEMISTRY & GEOPHYSICS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Chemical Geology","FirstCategoryId":"89","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0009254125003766","RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"GEOCHEMISTRY & GEOPHYSICS","Score":null,"Total":0}
Compositional evolution of the Archean subaerial crust revisited using machine learning constraints
The composition of the Archean subaerial crust, whether mafic or felsic, remains debated due to tectonic uncertainties and proxy limitations. Here, we employ Random Forest Regression algorithm-based machine learning model to analyze multi-element geochemical data from terrigenous fine-grained sedimentary rocks to reconstruct the MgO content in the subaerial crust. The model is trained using major oxides (TiO2 and Al2O3) and trace elements (Sc, Cr, Co, Ni, Cu, Zn, Y, Zr, Nb and REEs) of simulated mixture samples from Archean craton igneous rocks and post-Archean igneous rocks. The validity of this model is demonstrated by using various modern samples to predict MgO content in the modern subaerial crust, whose MgO content has been well constrained. We also demonstrate that the accuracy of the predictions is not sensitive to tectonic settings. We then use this model to reconstruct the chemical evolution of the subaerial crust over Earth's history and compare the differences between our model and other proxies reflecting the composition of the subaerial crust. Our model results suggest that a systematic decline of MgO content in the subaerial crust (7–10 wt% to 3–4 wt%) and a shift from mafic (70–100 % komatiite-basalt) to felsic dominance (∼70–90 % TTG) of the subaerial crust in the Neoarchean period, after which the subaerial crust attained a similar composition to that of the modern subaerial crust. The results from our approach are similar to those from Ni/Co and Cr/Zn proxies, but differ from the results of Zr/TiO2 and Al2O3/TiO2 proxies. We suggest that many previously used composition proxies are sensitive to tectonic settings and applying these proxies to terrigenous sediments can lead to erroneous results on the subaerial crust composition in the Archean. In addition, we caution against the use of proxies whose correlations with MgO or SiO2 content are highly nonlinear, which may result in large uncertainties in the projected MgO or SiO2 content for the source rocks.
期刊介绍:
Chemical Geology is an international journal that publishes original research papers on isotopic and elemental geochemistry, geochronology and cosmochemistry.
The Journal focuses on chemical processes in igneous, metamorphic, and sedimentary petrology, low- and high-temperature aqueous solutions, biogeochemistry, the environment and cosmochemistry.
Papers that are field, experimentally, or computationally based are appropriate if they are of broad international interest. The Journal generally does not publish papers that are primarily of regional or local interest, or which are primarily focused on remediation and applied geochemistry.
The Journal also welcomes innovative papers dealing with significant analytical advances that are of wide interest in the community and extend significantly beyond the scope of what would be included in the methods section of a standard research paper.