Document Type : Research Paper
Authors
1
Assistant professor, Soil and Water Research Institute of Iran
2
Assistant professor, Jiroft University
3
Lecturer Department, Jiroft University
Abstract
Determining the variability of soil parameters is a necessity for precision agriculture. To achieve higher yield with better management, knowledge of physico-chemical properties of soil in the fields is essential. Geostatistics is one of the methods developed for investigating the spatial variability of soil properties. For this purpose, 188 soil samples were taken from the Eastern farms of Mazandaran province in 2009. Soil samples were analyzed and the amounts of organic carbon (OC), P, K, Fe, Mn, and Cu were determined. The spatial correlation of variables and the best fitted model were determined by variogram. Analysis indicated that OC, K, P, and Mn were best fitted to Gaussian model. Also, Fe and Cu were best fitted to exponential and spherical models, respectively. The effective ranges of these parameters were 58, 26, 58, 5, 58, and 3 km, respectively. In order to determine the distribution map, Kriging, Inverse Distance Weighted (IDW) and Splines (RBF) methods were used. The precision of interpolations were calculated using mean base error (MBE), mean absolute error (MAE), and root mean square error (RMSE).The results showed that kriging method had a higher accuracy compared to IDW and RBF. Kriging was the best method to estimate OC, P, K, and Cu because it had the highest precision and lowest error. IDW and RBF had the highest precision for estimation of Mn and Fe, respectively, in this area.
Keywords
)
