Assessing the utility of visible-to-shortwave infrared reflectance spectroscopy for analysis of soil weathering intensity and paleoclimate reconstruction

Abstract

Visible-to-shortwave infrared reflectance (VSWIR) spectroscopy is a fast and efficient approach for estimation of soil properties. In order to test whether soil weathering intensity can be predicted using reflectance spectra in the VSWIR range (350–2500nm), and to examine the efficacy of this methodology to supplement or substitute for traditional mineralogical and geochemical techniques of paleoclimate reconstruction, we investigated a Quaternary soil sequence at Shengli in the Sichuan Basin of southwestern China. VSWIR absorption bands can be characterized by geometrical parameters including position (P), depth (D), width (W), asymmetry (AS), and full width at half maximum (F). Our results indicate that spectral features in the visible-to-near-infrared (VNIR; 350–1000nm) range can be interpreted in terms of the presence and/or concentration of individual Fe-oxide phases. Spectrally detected water is mostly bound in clay minerals, and this single dominant water source facilitates interpretation of shortwave infrared (SWIR; 1000–2500nm) parameters. Soil spectral parameters were compared with the Fe-oxide mineralogy, major element composition, and clay mineralogy of soil samples obtained from conventional laboratory analyses. Strong correlations are found between VNIR parameters and Fe-oxide mineralogy, with D900 serving as the best proxy for total ferric iron concentration. P900 and D700/D500 exhibit similar variation trends and relationships to the hematite/(hematite+goethite) [H/(H+G)] ratio of soil samples, confirming their value as monsoonal weathering intensity proxies. Lower values of P900 and D700/D500 indicate warmer and seasonally drier pedoenvironments, reflecting a stronger East Asian summer monsoon. The utility of some SWIR parameters for assessment of weathering intensity declines in inhomogeneous soils (e.g., those having vermiform structure). Thus, spectral evaluation of weathering intensity is most effective in soils that are relatively homogeneous, for which weathering degree is highly correlated to AS2200, D2200/D1900, and AS1400. The utility of these spectral proxies depends mainly on the mineral composition of soils and the application type (i.e., laboratory versus field studies). The application of VSWIR in the field (e.g., imaging spectroscopy, or IS) shows promise for rapid, in-situ mapping of weathering intensity, thus providing a convenient approach to paleoclimate reconstruction in remote and inaccessible regions.