Mineral-weathering Assessment And Paleoclimate Reconstruction Based On Visible And Near-infrared Reflectance Spectroscopy

Soils from Earth's surface record critical evolution information on climate,environment and human since the Quaternary.Soil is regarded as one of the most important carrier for studying the past geological history.The soil properties,such as ion transformation,grain-size distribution,mineral content and composition,are closely related to the coeval climatic and environmental change.Some properties are extracted as weathering proxies,which are widely used to reconstruct the past climate history.In the first half of this study we discussed the ability and stability of some soil properties?chemical properties,clay-mineral properties,magnetic properties,and non-magnetic iron properties?in reflecting soil weathering degree and reconstructing paleoclimate.Among all traditional weathering proxies,the chemical ratios are able to reconstruct climate history in various soil systems and are suggested as robust weathering proxies.However,it is not realistic for collecting chemical data with high precision across wide areas because it will cost lots of labor and financial resources.In the moderately weathered soil systems,the magnetic enhancement is mostly due to pedogenetic process and the increase of the weathering degree.In this case magnetic susceptibility is shown as a good weatehring proxy.In this study we focused the climatic significance,research methods and case studies of the authigenic mineral proxies-clay minerals and Fe-oxide minerals,and review their advantages and limitations in reflecting weathering degrees and climate conditions.Clay minerals and Fe-oxide minerals can be used as independent weathering and climate proxies.However,the applicability is conditioned by geological background,provenance,climate type,etc.Besides,a multi-proxy method is preferred in regional weathering and climate reconstruction.Moreover,looking for alternative weathering proxy is also feasible in future paleoclimate study.Visible and near-infrared reflectance?VNIR;350-2500 nm?spectroscopy is a rapid,non-destructive,and inexpensive method to effectively characterize soil samples with minimal sample preparation.The most diagnostic absorptions in the VNIR region are associated with climate-related mineral properties,therefore VNIR spectroscopy has great potential in analyzing soil-mineral data and being as new weatheirng proxies.The later part of this study is focused on the potential of VNIR spectroscopy for prediction of clay mineralogy and magnetic properties,as well as the establishment of new weathering proxies based on VNIR spectroscopy.First,we used principal component analysis?PCA?to analyze the original spectral data.Second,we used simple linear correlation and multiple linear correlation to correlate VNIR spectral data and the soil properties.PCA was performed on all spectra in order to better understand spectraldistributions and variations based on mineralogic composition,weathering,and geologic factors.It shows that the samples in different soil layers and different regions have distinct PCA distributions.In LC,the loess and paleosol samples have identific group features in the PCA plot;in SBC,the loess samples overlap with paleosol samples and show no group features in the PCA plot;in SL,the loam samples and clay samples also have distinct group features in the PCA plot.The soil spectral data of red earth samples and loess-paleosol samples exhibit large differences,resulting in no overlap in a PCA plot.However,the samples in LC and SBC loess-paleosol sequences have more common group features,causing more overlap in the PCA plot.The corralation matrix between VNIR spectra and soil properties show that soil-mineral properties have first-order relationships with spectral behavior?features and reflectance intensity?,soil-chemical properties have second-order relationships with spectral behavior?indirect relationships?.The variations of the soil-mineral and soil-chemical properties would significantly affect the VNIR spectral behaviors.The magnetic properties(?_lf and?_fd)have more influence on the spectral behaviors of the loess-paleosol samples,with the highest r of 0.80.The clay minerals have more influence on the spectral behaviors of the red earth soils,significantly affect the reflectance intensity of soil spectra in the SWIR region;however,they show little influence on the spectra of the loess-paleosol samples.The non-magnetic iron minerals affect the soil spectra in the 400-700nm region,which are directely related to the diagnostic absorption bands.As second-order impactor,soil-chemical properties?i.e.,?4Si and CIA?tend to have no rule in affecting the spectral behaviors.The chemical properties presented second-order relationship with soil spectra without their own absorption features where significant correlations of these properties were possible because their dependence on first-order ones since variations of such chemical attributes have been subjected to soil particle size,mineralogical and organic changes.The simple liner correlation between spectral data and soil properties show that the continuum-removal based VNIR morphological parameters?e.g.,depth and asymmetry?are significantly correlated with troditional weathering proxies?e.g.,clay-mineral proxies and iron properties?.Parameter D₉₀₀ is suggested as rubost proxy for total ferric iron concentration,it can remotely detect iron ore deposits and estimate their grades.Parameters P₉₀₀ and D₇₀₀/D₅₀₀ are highly correlated with hematite amount,showing excellent potential as paleomonsoon climate proxies,higher P₉₀₀ and D₇₀₀/D₅₀₀ values indicate weakening of the summer monsoon.Parameters AS₁₄₀₀,D₂₂₀₀/D₁₉₀₀,and AS₂₂₀₀are shown to be good weathering proxies for monsoonal climate reconstruction.Among them,parameter AS₁₄₀₀ and D₂₂₀₀/D₁₉₀₀ is not suitble for areas with vegetation and/or surface water,they are better for laboratory studies;However,parameter AS₂₂₀₀ is not influenced by vegetation and/or surface water,so it is particularly suited for field studies.The multiple liner correlation between the spectral data and soil properties indicate that VNIR spectra show potential for analysis of climatic related soil-mineral properties.Clay and iron mineralogy from the soils was compared with VNIR data using partial least squares regression?PLSR?and continuum-removal-based multiple linear regression(MLR_CR)approaches.Phyllosilicates and non-magnetic Fe-oxides are predicted more accurately for highly weathered red earth deposits,whereas soil magnetic properties are predicted more accurately for loess-paleosols.VNIR spectroscopy can be used as an alternative and broadly applicable tool in future paleoclimate studies based on soil mineralogy,substituting for XRD,DRS,and MS analyses.This will significantly reduce the workload and expense of experiments and improve the resolution of paleoclimate analyses.Soil type?weathering degree?and chemometric method are the main factors controlling the performance of predictive models,but model accuracy appears to be more dependent on the nature of soils than on the method selected.We suggest that model predictive ability is directly influenced by both concentration controls and variability controls,which are two factors related to soil type.In general,higher concentrations and wider variability in concentrations improve the capability of predictive models.We further infer that both concentration controls and variability controls are essential,and that a useable model should be built when both concentration and variability exceed certain thresholds.In general,VNIR spectra show potential for analysis of pedogenic weathering and paleoclimate reconstruction in weakly to highly weathered soils,in which spectral behavior is strongly related to both clay-mineral composition and magnetic properties.The parameters extracted by CR method and chemometric models(i.e.,D₉₀₀,AS₁₄₀₀,D₂₂₀₀/D₁₉₀₀,and AS₂₂₀₀)are proven to be robust weathering proxies in the moderately to highly weathered soils.We suggest that VNIR spectroscopy has the potential to be an alternative and broadly applicable tool for reconstructing paleoclimate using soil-mineral properties,substituting for a series of common mineralogic analyses.This would significantly reduce the workload and expense of experiments and improve the resolution of paleoclimate studies.