Determination of chemical soil properties using diffuse reflectance and ion-exchange resins

The management of soil nutrients is essential for sustainable agricultural production. The time requirements for soil nutrient determinations and the high cost per sample are problems that are attributed to traditional laboratory analyses that limit the adoption of precision agriculture techniques. Such problems arise because the sample density that is required to obtain soil fertility maps is greater than that required by conventional agricultural management. The use of radiometric sensors combined with a diffuse reflectance technique is quicker and less expensive than surveying soil fertility. However, the construction of robust models for the prediction of soil chemical properties based on spectral data requires samples with standardized physical characteristics. The objective of this work was to develop a model to predict the soil phosphorus (P), calcium (Ca), magnesium (Mg), and potassium (K) contents based on a multivariate analysis using spectroscopic data in the visible and near-infrared ranges. Ion-exchange resins were used to extract nutrients from the soil, and then diffuse reflectance spectra were collected. Models were constructed using partial least squares (PLS) regression, and the ordered predictors selection (OPS) algorithm was used for the selection of variables. The coefficients of determination (greater than 90%), ratios of the standard deviation to the root mean square error (higher than 2.20), and relative error percentages (lower than 25%) were obtained using the developed models. The mean values that were predicted by the models were significantly different from those measured in the laboratory only for K ions. For the other analyzed ions, including P, Ca and Mg, no significant differences were observed at the 5% level (p > 0.05). The results indicate that the PLS–OPS models based on the diffuse reflectance of ion-exchange resins are reliable for the fast and accurate prediction of these ions.