Performance Analysis of Geostatistical Approach and PCA Techniques for Mapping Cation Exchange Capacity in the South West of Iran.

The mapping of cation exchange capacity (CEC) was assessed using the geostatistical technique based on principal component analysis (PCA) in the Khuzestan province, southwestern Iran, for the proper management of agricultural lands. To this end, 210 topsoil samples were obtained from the area at the depth of 0–30 cm by systematic sampling in regular networks with dimensions of 350 × 350 m. Soil attributes were fully analyzed using both classical and geostatistical methods, and the prediction map of soil CEC was created by ordinary kriging (OK) and cokriging (CK) methods. The findings revealed that soil properties dominating the first principal component (PC1) explained the high correlation observed between soil properties such as CEC (r2 = 0.89). Still, a weak or no correlation between other components and CEC was obtained in the correlation matrix. According to the PCA results, the PC1 was used as an auxiliary variable to estimate the soil CEC in the CK method. The precision of the prepared maps was determined by the cross-validation analysis using NRMSE, MAE, and R2. Based on the cross-validation result of the predicted dataset, the NRMSE and MAE for OK were 0.25, 2.9, and 0.34 cmol(+) kg–1, and 0.18 and 2.1 cmol(+) kg–1 for CK, respectively. The cross-validation R2 for the prediction dataset was 0.31 for OK and 0.84 for CK methods at the 0.01 level, respectively. The result also revealed that the CK interpolation method with PC1 could effectively predict soil CEC variation. Furthermore, the combination of the CK method and PC1 derived from soil attributes was reasonably able to predict soil CEC. Principal components, which have a good correlation with the dependent variables, the best potential for CK predictions, and the application of CK interpolation method with PC1 can reduce the time and cost for CEC mapping. [ABSTRACT FROM AUTHOR]