Document Type : Research Paper
Authors
1 Assistant Professor, Agricultural and Natural Resources Research Center, Western Azerbyjan, Iran
2 Associate Professor, Agricultural Research Organization, Iran
3 Professor, Tehran University, Iran
4 Assistant Professor, Soil and Water Research Institute, Iran
Abstract
Planning and suitable management is necessary for optimal use of soil and for this; spatial variability of soil characteristics is important which may be edcarried out through geostatistical methods of parametric and non-parametric predictors such as TPSS, WMA, Kriging and Co-kriging. This research work was done in Southern part of Uromieh plain with 36690 ha surface area in order to study the spatial variability of soil lime, sand and saturation moisture percentage. Distance between soil profiles ranged 1300 to 4700 meters. For estimation and prediction of them in non-sampled points, the Kriging, Co- kriging and Weighted Moving Average were used in Geographic Information System environment. For selecting suitable interpolation method, Cross validation and MAE and MBE parameters were used. Selected method was also used for estimating and mapping of the selected soil characteristics. The Sturges rule was used for defining map classification. Results showed that the Kriging method has the highest accuracy with correlation coefficient of 0.83 and error of 3.98 percent for prediction of soil characteristics in non-sampled points.
Keywords
)