Utility of visible and near-infrared spectroscopy to predict base neutralizing capacity and lime requirement of quaternary soils

Abstract Detailed knowledge of a soil’s lime requirement (LR) is prerequisite for demand-based fertilization to achieve the optimum soil pH and thus sustainably increasing fertility crop yields. LR can be directly determined by base neutralizing capacity (BNC) obtained soil-base titration. For site-specific acidity management, detailed information on within-field variation BNC required. However, titrations determination are too laborious extensively applied in routine testing. In contrast, visible near-infrared spectroscopy (visNIRS) time cost-effective alternative that analyze several characteristics within single spectrum. VisNIRS was tested laboratory 170 air-dried sieved samples nine agricultural fields quaternary landscape North-east Germany predicting corresponding (LR ) at target 6.5. Seven spectral pre-processing methods were including new technique based normalized differences (ND). Furthermore, six multivariate regression conducted method combining forward stagewise subset selection algorithm with PLSR (FS-PLSR). The models validated using an independent sample set. best model most variables FS-PLSR combined second Savitzky-Golay derivation as achieving R 2 s from 0.68 0.82. Finally, performance direct prediction (R = 0.68) compared indirect calculated predicted parameters. This resulted slightly higher correlation coefficients 0.75.