Assessing soil mineralogy and weathering degree by a multi-range sensor synergistic approach: from parent rock to topsoil

Abstract

The weathering degree (WD) of soils is usually based on the presence of weatherable primary minerals and the content of chemical elements along the soil profile, which traditionally demand expensive, time-consuming and environmentally unfriendly laboratory analyses (e.g. wet chemistry). To address this issue, proximal sensors often operating in only a single spectral range have been used in soil WD assessments. In this sense, we analyzed in detail the performance of proximal sensors operating at several spectral ranges to assess the soil mineralogy and WD. We selected three soil profiles developed from sedimentary rocks and analyzed samples from parent rock to topsoil. Samples from each horizon of the soil profiles and from the parent rock were collected and submitted to chemical, physical and mineralogical laboratory analyses. These samples were also analyzed using several sensors operating within different spectral ranges, i.e., gamma-ray, X-ray fluorescence (XRF), visible (VIS), infrared (IR) (near infrared, NIR, shortwave infrared, SWIR, and middle infrared, Mid-IR), and the AISA Fenix hyperspectral imaging systems (VIS-NIR-SWIR). Several WD indices were calculated using laboratory (reference) and multi-range sensor data for comparison. The spectral differences in the VIS-IR range were evaluated according to the soil mineralogy and WD. Gamma-ray and XRF sensors presented high correlation with the main soil elements in oxide minerals used in the WD quantification (R ranging from 0.70 to 0.99). We detected differences associated with the WD of each horizon of all profiles using data from VIS to Mid-IR ranges. The application of linear models to AISA images enabled the mapping and quantification of main elements along one soil profile (R2 ranging from 0.48 to 0.8) using a pixel-to-pixel approach, which facilitated the interpretation of the WD. Due to the specificity of each spectral region, the use of several sensors at different wavelengths provided a holistic overview of mineralogy and WD of the analyzed soil profiles, using a quick and non-destructive approach. This technique produces overlapping and/or complementary information of WD and mineralogy from all spectral ranges and can significantly assist traditional laboratory methods.