In this study,we present a machine learning-based method to predict trace element concentrations from major and minor element concentration data using a legacy lithogeochemical database of magmatic rocks from the Karo...In this study,we present a machine learning-based method to predict trace element concentrations from major and minor element concentration data using a legacy lithogeochemical database of magmatic rocks from the Karoo large igneous province(Gondwana Supercontinent).Wedemonstrate that a variety of trace elements,including most of the lanthanides,chalcophile,lithophile,and siderophile elements,can be predicted with excellent accuracy.This finding reveals that there are reliable,high-dimensional elemental associations that can be used to predict trace elements in a range of plutonic and volcanic rocks.Since the major and minor elements are used as predictors,prediction performance can be used as a direct proxy for geochemical anomalies.As such,our proposed method is suitable for prospective exploration by identifying anomalous trace element concentrations.Compared to multivariate compositional data analysis methods,the new method does not rely on assumptions of stoichiometric combinations of elements in the data to discover geochemical anomalies.Because we do not use multivariate compositional data analysis techniques(e.g.principal component analysis and combined use of major,minor and trace elements data),we also show that log-ratio transforms do not increase the performance of the proposed approach and are unnecessary for algorithms that are not spatially aware in the feature space.Therefore,we demonstrate that high-dimensional elemental associations can be modelled in an automated manner through a data-driven approach and without assumptions of stoichiometry within the data.The approach proposed in this study can be used as a replacement method to the multivariate compositional data analysis technique that is used for prospectivity mapping,or be used as a pre-processor to reduce the detection of false geochemical anomalies,particularly where the data is of variable quality.展开更多
The Mesoproterozoic(1.11 Ga)Umkondo large igneous province(LIP)in southern Africa and Antarctica was emplaced in<5 Myr and is dominated by low-Ti tholeiitic doleritic-gabbroic sills.It is of particular interest bec...The Mesoproterozoic(1.11 Ga)Umkondo large igneous province(LIP)in southern Africa and Antarctica was emplaced in<5 Myr and is dominated by low-Ti tholeiitic doleritic-gabbroic sills.It is of particular interest because it is the least studied LIP in southern Africa with both sublithospheric and lithospheric mantle sources proposed and it coincides with the early assembly of Rodinia,so it has importance in understanding the nature of magmatism and tectonics in and around the Kalahari craton during the Mesoproterozoic.In this study,we compiled a large database of existing(~750)and new(~100)major and trace element data for the Umkondo province,as well as 42 new Sr-Nd isotopic measurements,to provide constraints on its magma sources and geochemical evolution.Major element compositional variations in the low-Ti tholeiites are explained by low-pressure(1 kbar)three-phase fractional crystallisation(olivine,clinopyroxene and plagioclase)of a parent magma with~10 wt.%MgO in oxidising conditions(QFM+1).Inverse models show that the low-Ti tholeiitic magmas were derived as residual melts after the crystallization of 12%-33%olivine from primary komatiitic-basaltic magmas(up to~20 wt.%MgO)in equilibrium with mantle olivine(Fo90).Low Sm/Yb and TiO2/Yb-Nb/Yb indicate that the primary magmas were derived by 2%-20%shallow(40-50 km)partial melting of spinel lherzolite.High Sm/Yb is restricted to dyke swarms and may imply limited magma production from deeper(up to~70 km)garnet lherzolite-like sources.The low-Ti tholeiites of the Umkondo province are enriched in large ion lithophile elements(Rb-Sr-Cs-K)and depleted in high-field strength elements(Zr-Hf-Nb-Ta),indicating the involvement of crustal material and/or the subcontinental lithospheric mantle.This is supported by covariations in Th/Nb,Nb/Yb,Nb/La and Ce/Sm with generally negativeΔNb.Sr-Nd isotopes lend support to the notion that the Umkondo magmas were derived from depleted and/or enriched sublithospheric mantle sources and subsequently contaminated by enriched lithospheric material during emplacement(initial(at 1.11 Ga)^(87)Sr/^(86)Sr between 0.704820 and 0.737464 andεNd between-8.9 and+5.3).The Vredefort sills are significant as they display the most depleted Sr-Nd isotopic signatures(average initial ^(87)Sr/^(86)Sr of 0.705342 and averageεNd of 0.4)and are the least contaminated magma suite in the Umkondo province.Because of(i)the large volume of low-Ti magmas,(ii)evidence of a primary hot and MgO-rich(komatiitic)magma,and(iii)the short duration of magmatism,we suggest that the Umkondo province was formed by plume-induced melting of the sublithospheric mantle beneath the Kalahari craton in an extensional setting.This contrasts with previous suggestions that the heat source developed in response to the“thermal insulation”of the mantle beneath a thickened Kalahari craton in the absence of a mantle plume.There is further evidence from the elevated Zn/Fe that the sublithospheric mantle was lithologically heterogeneous and consisted of mixed peridotite and pyroxenite domains.There is a general lack of ultramafic cumulates in the low-Ti magma suite that may imply there was deeper ponding and storage of the primary magmas that fractionated large quantities of ultramafic rocks.There is also a paucity of high-Ti rocks in the Umkondo province that may reflect limited direct melting of the lithospheric mantle or that they are simply not as well-preserved in this province compared to the Karoo province.The similar trace element and Sr-Nd isotopic compositions of the Umkondo sills in southern Africa with the Borgmassivet sills in Antarctica support the concept that the Kalahari craton and Grunehogna terrane were adjoined at 1.11 Ga.The timing of the Umkondo province indicates there was localised lithospheric extension and upwelling asthenospheric mantle during a time of dominantly compressional tectonics on Earth at the end of the‘boring billion’.展开更多
文摘In this study,we present a machine learning-based method to predict trace element concentrations from major and minor element concentration data using a legacy lithogeochemical database of magmatic rocks from the Karoo large igneous province(Gondwana Supercontinent).Wedemonstrate that a variety of trace elements,including most of the lanthanides,chalcophile,lithophile,and siderophile elements,can be predicted with excellent accuracy.This finding reveals that there are reliable,high-dimensional elemental associations that can be used to predict trace elements in a range of plutonic and volcanic rocks.Since the major and minor elements are used as predictors,prediction performance can be used as a direct proxy for geochemical anomalies.As such,our proposed method is suitable for prospective exploration by identifying anomalous trace element concentrations.Compared to multivariate compositional data analysis methods,the new method does not rely on assumptions of stoichiometric combinations of elements in the data to discover geochemical anomalies.Because we do not use multivariate compositional data analysis techniques(e.g.principal component analysis and combined use of major,minor and trace elements data),we also show that log-ratio transforms do not increase the performance of the proposed approach and are unnecessary for algorithms that are not spatially aware in the feature space.Therefore,we demonstrate that high-dimensional elemental associations can be modelled in an automated manner through a data-driven approach and without assumptions of stoichiometry within the data.The approach proposed in this study can be used as a replacement method to the multivariate compositional data analysis technique that is used for prospectivity mapping,or be used as a pre-processor to reduce the detection of false geochemical anomalies,particularly where the data is of variable quality.
文摘The Mesoproterozoic(1.11 Ga)Umkondo large igneous province(LIP)in southern Africa and Antarctica was emplaced in<5 Myr and is dominated by low-Ti tholeiitic doleritic-gabbroic sills.It is of particular interest because it is the least studied LIP in southern Africa with both sublithospheric and lithospheric mantle sources proposed and it coincides with the early assembly of Rodinia,so it has importance in understanding the nature of magmatism and tectonics in and around the Kalahari craton during the Mesoproterozoic.In this study,we compiled a large database of existing(~750)and new(~100)major and trace element data for the Umkondo province,as well as 42 new Sr-Nd isotopic measurements,to provide constraints on its magma sources and geochemical evolution.Major element compositional variations in the low-Ti tholeiites are explained by low-pressure(1 kbar)three-phase fractional crystallisation(olivine,clinopyroxene and plagioclase)of a parent magma with~10 wt.%MgO in oxidising conditions(QFM+1).Inverse models show that the low-Ti tholeiitic magmas were derived as residual melts after the crystallization of 12%-33%olivine from primary komatiitic-basaltic magmas(up to~20 wt.%MgO)in equilibrium with mantle olivine(Fo90).Low Sm/Yb and TiO2/Yb-Nb/Yb indicate that the primary magmas were derived by 2%-20%shallow(40-50 km)partial melting of spinel lherzolite.High Sm/Yb is restricted to dyke swarms and may imply limited magma production from deeper(up to~70 km)garnet lherzolite-like sources.The low-Ti tholeiites of the Umkondo province are enriched in large ion lithophile elements(Rb-Sr-Cs-K)and depleted in high-field strength elements(Zr-Hf-Nb-Ta),indicating the involvement of crustal material and/or the subcontinental lithospheric mantle.This is supported by covariations in Th/Nb,Nb/Yb,Nb/La and Ce/Sm with generally negativeΔNb.Sr-Nd isotopes lend support to the notion that the Umkondo magmas were derived from depleted and/or enriched sublithospheric mantle sources and subsequently contaminated by enriched lithospheric material during emplacement(initial(at 1.11 Ga)^(87)Sr/^(86)Sr between 0.704820 and 0.737464 andεNd between-8.9 and+5.3).The Vredefort sills are significant as they display the most depleted Sr-Nd isotopic signatures(average initial ^(87)Sr/^(86)Sr of 0.705342 and averageεNd of 0.4)and are the least contaminated magma suite in the Umkondo province.Because of(i)the large volume of low-Ti magmas,(ii)evidence of a primary hot and MgO-rich(komatiitic)magma,and(iii)the short duration of magmatism,we suggest that the Umkondo province was formed by plume-induced melting of the sublithospheric mantle beneath the Kalahari craton in an extensional setting.This contrasts with previous suggestions that the heat source developed in response to the“thermal insulation”of the mantle beneath a thickened Kalahari craton in the absence of a mantle plume.There is further evidence from the elevated Zn/Fe that the sublithospheric mantle was lithologically heterogeneous and consisted of mixed peridotite and pyroxenite domains.There is a general lack of ultramafic cumulates in the low-Ti magma suite that may imply there was deeper ponding and storage of the primary magmas that fractionated large quantities of ultramafic rocks.There is also a paucity of high-Ti rocks in the Umkondo province that may reflect limited direct melting of the lithospheric mantle or that they are simply not as well-preserved in this province compared to the Karoo province.The similar trace element and Sr-Nd isotopic compositions of the Umkondo sills in southern Africa with the Borgmassivet sills in Antarctica support the concept that the Kalahari craton and Grunehogna terrane were adjoined at 1.11 Ga.The timing of the Umkondo province indicates there was localised lithospheric extension and upwelling asthenospheric mantle during a time of dominantly compressional tectonics on Earth at the end of the‘boring billion’.
