5:15 PM - 6:45 PM
[BCG06-P03] Estimate of the Eoarchean seawater composition using an independent component analysis
Keywords:Seawater composition, Eoarchean, Independent component analysis, Evolution of life
Estimate of the chemical composition of ancient seawater is one of the key issues in the decoding history of the Earth because it is widely considered that it has contributed significantly to the evolution of organisms throughout Earth's history. However, it is difficult to estimate the chemical composition of ancient seawater; in particular, previous studies lacked quantification.
In general, ancient seawater composition is estimated from the chemical compositions of chemical sedimentary rocks. However, they contain not only authigenic minerals precipitated from seawater, but also a variety of clastic materials; thus, their chemical composition does not necessarily reflect the seawater composition. In addition, the influence of elemental movement due to later alteration and metamorphism must be considered. Konhauser et al. (2009) showed that the 3.8–2.7 Ga banded iron formations (BIFs) have higher Ni/Fe ratios than iron-rich rocks since 2.7 Ga, and suggested that the decrease in Ni content of seawater resulted in the decline of methanogens and the relative prosperity of oxygenic photosynthetic organisms because Ni, included in Cofactor F430, is a bioessential element for methanogens. This turnover led to the formation of an oxic environment on Earth. However, the estimated secular change is still so obscure that the hypothesis is not verifiable because the Ni/Fe ratios are highly scattered, possibly due to the influence of clastic material contamination. Aoki et al (2018) estimated the Ni/Fe ratios of the “pure” iron oxide components from the correlation with Zr because Zr contents in BIFs are usable as an indicator of clastic materials. However, this estimate is still insufficient because BIFs contain various sources of contamination such as terrestrial and volcanic materials.
We employed independent component analysis (ICA) to the Eoarchean sedimentary rocks (3.9–3.8 Ga) to estimate the compositions of sources with different origins in the sedimentary rocks. ICA enables the reduction of multidimensional data, but unlike principal component analysis, it extracts dissociations from a non-normal distribution. Each extracted component is also 'independent' of each other, which may be helpful in estimating the compositions of components from multiple origins. The datasets are composed of BIF (36 samples), carbonate rocks (24 samples), and clastic rocks (18 samples) from Isua and BIF (36 samples), carbonate rocks (8 samples) and clastic rocks (59 samples) from Labrador, with 10 major elements and 31 (or 33) trace elements.
The components particularly rich in FeOT, SiO2, TiO2+Al2O3, and MgO+CaO contents are extracted, corresponding to iron oxyhydroxides, silica minerals, clastic components, and carbonate minerals, respectively. Several types of clastic components were also separated, as well as a silicification component, causing SiO2 enrichment in carbonate rocks. The estimated iron oxide component in the Isua BIF is characterized by positive Eu and Y/Ho anomalies and enrichment in P. The silica mineral component also possesses positive Eu and Y/Ho anomalies, whereas the silicification component, which also causes an increase in SiO2, does not exhibit these characteristics. All these components can be found in the BIF, carbonate rocks, and clastic rocks. Furthermore, the hornblende-rich part of the BIF shows depletion in light rare-earth elements. Ni is mainly positively correlated with the clastic component, supporting the involvement of Ni in clastic components rather than iron oxyhydroxides in the BIF. We expect to remove the contamination of clastic materials and more quantitatively estimate the Ni content in seawater by calculating the each contribution of Ni from iron oxyhydroxides or clastic materials.
In addition, ICA using only major elements can remove the influence of metamorphism and alteration to estimate the original mineral compositions, whereas ICA using major elements, REEs, and transition elements allows us to estimate depositional processes, such as hydrothermal environments. The combination is expected to enable more detailed analysis of sedimentary rocks that have undergone alteration and metamorphism.
In general, ancient seawater composition is estimated from the chemical compositions of chemical sedimentary rocks. However, they contain not only authigenic minerals precipitated from seawater, but also a variety of clastic materials; thus, their chemical composition does not necessarily reflect the seawater composition. In addition, the influence of elemental movement due to later alteration and metamorphism must be considered. Konhauser et al. (2009) showed that the 3.8–2.7 Ga banded iron formations (BIFs) have higher Ni/Fe ratios than iron-rich rocks since 2.7 Ga, and suggested that the decrease in Ni content of seawater resulted in the decline of methanogens and the relative prosperity of oxygenic photosynthetic organisms because Ni, included in Cofactor F430, is a bioessential element for methanogens. This turnover led to the formation of an oxic environment on Earth. However, the estimated secular change is still so obscure that the hypothesis is not verifiable because the Ni/Fe ratios are highly scattered, possibly due to the influence of clastic material contamination. Aoki et al (2018) estimated the Ni/Fe ratios of the “pure” iron oxide components from the correlation with Zr because Zr contents in BIFs are usable as an indicator of clastic materials. However, this estimate is still insufficient because BIFs contain various sources of contamination such as terrestrial and volcanic materials.
We employed independent component analysis (ICA) to the Eoarchean sedimentary rocks (3.9–3.8 Ga) to estimate the compositions of sources with different origins in the sedimentary rocks. ICA enables the reduction of multidimensional data, but unlike principal component analysis, it extracts dissociations from a non-normal distribution. Each extracted component is also 'independent' of each other, which may be helpful in estimating the compositions of components from multiple origins. The datasets are composed of BIF (36 samples), carbonate rocks (24 samples), and clastic rocks (18 samples) from Isua and BIF (36 samples), carbonate rocks (8 samples) and clastic rocks (59 samples) from Labrador, with 10 major elements and 31 (or 33) trace elements.
The components particularly rich in FeOT, SiO2, TiO2+Al2O3, and MgO+CaO contents are extracted, corresponding to iron oxyhydroxides, silica minerals, clastic components, and carbonate minerals, respectively. Several types of clastic components were also separated, as well as a silicification component, causing SiO2 enrichment in carbonate rocks. The estimated iron oxide component in the Isua BIF is characterized by positive Eu and Y/Ho anomalies and enrichment in P. The silica mineral component also possesses positive Eu and Y/Ho anomalies, whereas the silicification component, which also causes an increase in SiO2, does not exhibit these characteristics. All these components can be found in the BIF, carbonate rocks, and clastic rocks. Furthermore, the hornblende-rich part of the BIF shows depletion in light rare-earth elements. Ni is mainly positively correlated with the clastic component, supporting the involvement of Ni in clastic components rather than iron oxyhydroxides in the BIF. We expect to remove the contamination of clastic materials and more quantitatively estimate the Ni content in seawater by calculating the each contribution of Ni from iron oxyhydroxides or clastic materials.
In addition, ICA using only major elements can remove the influence of metamorphism and alteration to estimate the original mineral compositions, whereas ICA using major elements, REEs, and transition elements allows us to estimate depositional processes, such as hydrothermal environments. The combination is expected to enable more detailed analysis of sedimentary rocks that have undergone alteration and metamorphism.