Effect of moisture content on the prediction of cation exchange capacity using visible and near infrared spectroscopy

This study was undertaken to investigate the effect of moisture content (MC) on the prediction accuracy of the cation exchange capacity (CEC) of soils collected from Turkey using a fibre-type Vis-NIR spectrophotometer. The diffuse reflectance spectra of 150 soil samples were measured under 6 gravimetric MC levels of 0, 5, 10, 15, 20 and 25%. The results showed that the effect of MC on the prediction accuracy of CEC was nonlinear. Due to the direct spectral response in the NIR range, the prediction of CEC in the independent validation set was successful with a root mean square error of prediction (RMSEP) ranging from 5.82 to 8.14% and a ratio of prediction deviation (RPD) ranging from 1.59 to 2.14. The highest accuracy achieved for CEC was with the 10 and 15% MC soil samples (RMSEP 5.82%; RPD 2.14 and 2.12, respectively), and the lowest accuracy was found for wet soil samples with a 20% MC (RMSEP 8.14%; RPD 1.59). These results suggested that there was a clear effect of MC on the prediction of CEC. Because the differences between the models at different levels of MC were significant for CEC.