Geostatistical Modeling of Some Soil Fertility Parameters at an Agricultural Field Formed on Alluvial Material in Çukurova

This study was focused on determining spatial variability structure and spatial dependence level of some selected soil fertility parameters in a farmer land, site-specific management field, where maize has been planted every year, in Çukurova (the Mediterranean Region), Southern Turkey. Some soil properties [total nitrogen (tN) (mg kg-1), available phosphorous (aP) (mg kg-1), available potassium (aK) (mg kg-1), OM (%), EC (dS m-1), pH, CEC (cmol kg-1), sand (%), silt (%), clay (%), CaCO3 (%)] at the depth of 0-30 cm were evaluated geostatistically by georeferenced 260 samples, which were collected in different sampling intervals, within 38 ha area. An isotropic variogram structure was determined for aP, where EC had pure nugget effect and other soil properties showed anisotropic variogram structures within different azimuth angles. The highest spatial dependence ranks were observed in the geostatistical model results of clay and CEC, with the values of 2.6 % and 3.6 %, while the least dependencies were obtained from those of aP and pH, with the values of 56.1 % and 53.2 %, respectively. Major and minor structural ranges were modeled together to remove the anisotropy and while OMs’ structural range was found as the longest spatio-structural range with 500 m, silts’ was found as the shortest spatiostructural range with 220 m. The majority of soil properties showed strong spatio-dependency level in this study area. Spatial distribution of soil properties was interpolated by ordinary kriging method, depending on variogram structures (isotropic or anisotropic), which were verified by the cross validation. There was strong similarity among the kriging spatial distributions of clay content, CEC, OM, and aK as well as among their semivariograms. These similarities can be a promising indicator for employing co-kriging technique among these soil properties.