Investigation and prediction spatial variability in chemical properties of agricultural soil using geostatistics

Determination of the chemical characteristics of soil for balanced fertilization on large scales is an important factor in achieving a precision agriculture. Laboratory analyses of soil properties are usually expensive and time consuming. Surmounting these problems is possible using geostatistics. Therefore, this research aims at selecting a proper interpolation method using 213 soil samples for alfalfa farmland in Hamadan Province, Iran. Various factors such as pH, EC, , , K, P, Fe, Zn, B and Co were measured. Ordinary kriging and co-kriging were assessed to derive maps of soil physico-chemical properties, using mean absolute error (MAE), mean bias error (MBE), root mean squared error (RMSE) and average kriging standard error (AKSE) as statistical criteria. Variography analysis indicated that the ranges of influence for pH, EC, , , K, P, Zn, Fe, B and Co were 65, 55, 78, 79, 75, 60, 50, 65, 70 and 30 km, respectively, and the measuring error varied between 0.366 and 0.843. The results revealed that, based on precision criteria, co-kriging was the best method for interpolating the chemical properties of soil. Finally, using to the co-kriging for each determined variable, a related zoning map for fertility management of the study area was prepared.