Vis-NIR, MIR, and XRF spectral data fusion for predicting soil weathering indices – A case study using 150 Alfisols
Catena, cilt.274, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 274
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.catena.2026.110523
- Dergi Adı: Catena
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Artic & Antarctic Regions, BIOSIS, Environment Index, Geobase, Academic Search Ultimate (EBSCO)
- Anahtar Kelimeler: Alfisols, Glossic horizon, Loess, Machine learning, Soil spectra, Weathering
- Eskişehir Osmangazi Üniversitesi Adresli: Evet
Özet
Elemental ratios and weathering indices (WIs) are widely used to assess soil weathering under different climatic and environmental conditions. We investigated weathering in 150 Alfisols developed in loess over glacial till in central Wisconsin using 517 samples collected from Ap, glossic, Bt, and 2B(t) horizons. The loess-over-till sequence provided an opportunity to evaluate weathering across contrasting parent materials. Elemental concentrations (Al, Ca, Fe, K, Mg, P, Si, Ti, Zn, and Zr) determined by X-ray fluorescence (XRF) were used to calculate six weathering indices: the Chemical Index of Alteration (CIA), Chemical Index of Weathering (CIW), Chemical Proxy of Alteration (CPA), Desilication Index (DI), Ruxton Index (RI), and Weathering Index of Parker (WIP). Visible–near infrared (Vis–NIR; 350–2500 nm), mid-infrared (MIR; 4000–600 cm−1), and XRF spectra were used to predict these indices using partial least squares regression (PLSR) and Cubist models. Weathering intensity generally increased from the Ap to the 2B(t) horizons, reflecting pedogenic differentiation through leaching of base cations from the surface horizons, illuviation of clay and Fe-Al oxides into the Bt horizons, and the greater weathering of the older underlying glacial till. Weathering indices varied with land use and soil drainage class. MIR spectroscopy produced the most accurate predictions, with R2 values of 0.95 for DI and 0.93 for RI in glossic horizons using PLSR. Combining Vis–NIR, MIR, and XRF data further improved prediction accuracy, yielding R2 values of 0.95 for RI, 0.83 for CIW, and 0.78 for CPA using Cubist models. Spectral data fusion enhanced the prediction of soil weathering indices compared with individual sensors. Overall, spectrometer type, prediction model, and soil horizon were the principal factors controlling prediction accuracy, demonstrating that integrated proximal sensing provides a rapid and reliable approach for quantifying weathering intensity in Alfisols developed from different parent materials.