Deriving and Assessing Spectrotransfer Function and Pedotransfer Function in Predicting Soil Cation Exchange Capacity

Cation exchange capacity (CEC) is an important soil physicochemical property that has great effect on fertility and soil quality management. Measurement of CEC is difficult, time-consuming and expensive. The objective of this study was to assess whether inclusion of soil spectral data as a unique set of the predictors and alternative to soil basic properties would improve CEC predictions. Consequently, a total of 120 soil samples were collected from surface soil layer. The CEC and easily-determined soil properties were measured by standard laboratory methods. The spectral reflectance of soils over 350 to 2500 nm range were also determined using a handheld spectroradiometer apparatus. Different pre-processing techniques were evaluated after recording the spectra. Stepwise multiple linear regression (SMLR) was used to estimate some soil properties and CEC. Three scenarios including spectrotransfer functions (STF), pedotransfer functions (PTF) and spectropedotransfer functions (SPTF) were investigated. Results showed that STF had higher accuracy (RPD=1.50; RMSE=2.57 cmolc/kg) than the others in predicting soil CEC. PTF (RPD=1.09; RMSE=3.55 cmolc/kg) and SPTF (RPD=0.95; RMSR=4.06 cmolc/kg) provided poor predictions accuracy. These results suggest the efficacy of the spectral data, which can be used as an indirect, simple, and fast method to predict soil cation exchange capacity.