Visible and near-infrared diffuse reflectance spectroscopy (VIS-NIR) has shown levels of accuracy comparable to conventional laboratory methods for estimating soil properties. Soil chemical and physical properties have been predicted by reflectance spectroscopy successfully on subtropical and temperate soils, whereas soils from tropical agro-forest regions have received less attention, especially those from tropical rainforests. A spectral characterization provides a proficient pathway for soil characterization. The first step in this process is to develop a comprehensive VIS-NIR soil library of multiple key soil properties to be used in future soil surveys. This paper presents the first VIS-NIR soil library for a remote region in the Central Amazon. We evaluated the performance of VIS-NIR for the prediction of soil properties in the Central Amazon, Brazil. Soil properties measured and predicted were: pH, Ca, Mg, Al, H, H+Al, P, organic C (SOC), sum of bases, cation exchange capacity (CEC), percentage of base saturation (V), Al saturation (m), clay, sand, silt, silt/clay (S/C), and degree of flocculation. Soil samples were scanned in the laboratory in the VIS-NIR range (350–2500 nm), and forty-one pre-processing methods were tested to improve predictions. Clay content was predicted with the highest accuracy, followed by SOC. Sand, S/C, H, Al, H+Al, CEC, m and V predictions were reasonably good. The other soil properties were poorly predicted. Among the soil properties predicted well, SOC is one of the critical soil indicators in the global carbon cycle. Besides the soil property of interest, the landscape position, soil order and depth influenced in the model performance. For silt content, pH and S/C, the model performed better in well-drained soils, whereas for SOC best predictions were obtained in poorly drained soils. The association of VIS-NIR spectral data to landforms, vegetation classes, and soil types demonstrate potential for soil characterization.

The United States NRCS has a soil database that has data collected from across the country over the last several decades. This also includes soil spectral scans. This data is available, but it may not be used to its full potential. For this study, pedon, horizon and spectral data was extracted from the database for samples collected from 2011 to 2015. Only sites that had been fully described and horizons that had been analyzed for the full suite of desired properties were used. This resulted in over 14,000 samples that were used for modeling and eight soil properties: soil organic carbon (SOC); total nitrogen (TN); total sulfur (TS); clay; sand; exchangeable calcium (Ca_ex); cation exchange capacity (CEC); and pH. Four chemometric methods were employed for soil property prediction: partial least squares (PLSR); Random Forest (RF); Cubist; and multivariable adaptive regression splines (MARS). The dataset was sufficiently large that only random subsetting was used to create calibration (70%) and validation (30%) sets. SOC, TN, and TS had the strongest prediction results, with an R² value of over 0.9. Ca_ex, CEC and pH were predicted moderately well. Clay and sand models had slightly lower performance. Of the four methods, Cubist produced the strongest models, while PLSR produced the weakest. This may be due to the complex relationships between soil properties and spectra that PLSR could not capture. The only drawback of Cubist is the difficult interpretability of variable importance. Future research should include the use of environmental variables to improve prediction results. Future work may also avoid the use of PLSR when dealing with large datasets that cover large areas and have high degrees of variability.

Summary: Few studies have systematically investigated the effects of subsetting strategies on soil modelling or explored the potential of emergent methods from other fields not previously applied to pedometrics. This study considers smallholder agricultural villages in southern India that have been understudied in terms of chemometric modelling intended to support soil health, fertility and management. Therefore, the objective was to investigate the application of visible near‐infrared spectroscopy and chemometrics to predict soil properties in this setting. In addition, this study evaluated the effects of methods of calibration subsetting and new parametric models on the prediction of soil properties. These novel methods were transferred from the genomics field to soil science. Three strategic subsetting methods were used to produce calibration subsets that consider the variation in the soil properties, the spectra and both together; this is in addition to standard random calibration subsetting. Partial least squares regression (PLSR) and two methods from genomics that impose variable reduction were used for modelling; the latter were sparse PLSR (SPLSR) and the heteroscedastic effects model (HEM). Soil samples were collected from two villages and analysed for texture, soil carbon and available macro‐ and micro‐nutrients. The results showed that soil texture and carbon could be predicted moderately to strongly, whereas plant nutrient properties were predicted poorly to moderately. Random subsetting and subsetting by property distribution were more appropriate when spectra varied less overall, whereas subsetting that incorporates variation in spectra and properties improved results when spectral variation increased. The SPLSR and HEM models improved results over PLSR in some cases, or at least maintained prediction strength while using fewer predictors. Subsetting methods improved prediction results in 75% of cases. This study filled an important research gap by systematically studying local subsetting behaviour under different degrees of spectral and attribute variation. Highlights: Explored new calibration subsetting methods and chemometric models in soil spectral modelling. Compared the methods and models for 17 soil properties in an understudied area of India. Random subsetting was not always optimal; subsetting matters and depends on data characteristics. Sparse models from genomics performed better in 75% of cases than a standard method.