"International Journal of Geography, Geology and Environment"

Prediction models have been widely used to create a statistical model and understand the relationship between environmental variables and soil attributes. The present study was conducted to access the spatial variability in soil parameters by using geostatistical techniques. The study area selected was the National Capital Territory (NCT) of Delhi within which 22 sampling points were chosen for sample collection from the surface layer. The parameters (pH, Electrical Conductivity, Soil Moisture (SM)%, and Soil Organic Carbon (SOC) % were analyzed against the prediction estimates, and a regression analysis supports the inter-relation of the observed and predicted set of values. The method of interpolation used in the study was RBF (Radial Basis Function) which was carried out by using the ArcGIS software. The cross-validation of the data set was also analyzed by calculating the Mean and Root Mean Square Error (RMSE). According to the results, sample distance is adequate for interpolation and the RBF can clearly show the geographical distribution of soil attributes.