Knowledge of the variability of soluble salt content in saline soils can assist in designing experiments or developing management practices to manage and reclaim salt-affected soils. Geostatistical theory enables the use of spatial dependence of soil properties to obtain information about locations in the field that are not actually measured, but classical statistical methods do not consider spatial correlation and the relative location of samples. A study was carried out using both classical statistics and geostatistical methods to delineate salinity and sand content and their variability in a small area of irrigated saline soil. Soil samples were taken for electrical conductivity (EC) and particle size distribution determinations at 64 locations from a 20 × 25-m area, on an 8 × 8-grid pattern at depth intervals of 0–15, 15–30, 30–60, 60–90 and 90–120 cm. The high coefficient of variation (CV) values of both EC and sand content indicated that the soil was highly variable with respect to these soil properties. The semivariograms of sand content of the first two depth intervals and EC of all the depth intervals showed strong spatial relationships. Contour maps, generated by block kriging, based on spatial relationships provide estimated variances which are smaller than general variances calculated by the classical statistical method. The interpolated EC results by both ordinary and universal kriging methods were compared and were almost identical. The kriged maps can provide information useful for designing experiments and for determining soil sampling strategy. Key words: Salinity, texture, variability, geostatistics, semivariogram, kriging
Linking management practices and soil properties to Ecosystem Services in Mediterranean mixed orchards