International Journal of Applied Research
Vol. 2, Issue 11, Part A (2016)
Employing of regression analysis for prediction of sodium adsorption ratio of a soil
For practical soil management, indirect estimation of soil sodium adsorption ration (SAR, dimensionless) may be a good tool due to its important role in estimating of both the amount requirements of amendments and gypsum for soil reclamation. SAR is calculated from sodium, calcium and magnesium concentrations which are often determined with high costs. Therefore, developing of a simple tool to estimate SAR indirectly is more economical. Input data of the present study were collected from literature. The validation data were collected from actual laboratory soil analysis. Different regression models were developed to predict soil SAR based on soil electric conductivity (EC, dSm-1), soil pH (pH) and soil texture index (STI, dimensionless). The best regression model for estimating SAR was selected based on higher coefficient of determination (R2), lower both root mean square error (RMSE) and mean absolute error (MAE). The best regression model had the following form:
The performance of the best developed regression model was evaluated using an independent test data set. In order to evaluate the model, R2 was used. The value of R2 derived by the model for testing data was 0.9927. The proposed model is simple to be used by soil scientists and agricultural engineers to have a rapid check on sodium adsorption ratio at wide range of soil conditions within the studied range without the necessity of any time consuming and laboratory tests.
How to cite this article:
Abdulwahed M Aboukarima, Adel M Ghoneim, Mohamed S El-Marazky. Employing of regression analysis for prediction of sodium adsorption ratio of a soil. Int J Appl Res 2016;2(11):44-48.