Adequate characterization, mapping, and precise delineations of areas with soil pollution are essential for optimizing decontamination. Heterogeneity in soil sodification and associated growth of Silver oak (Grevillea robusta A. cunv.) trees were studied. Soil samples from 6 × 6-M grid patterns in four layers at a depth increment of 0.3 M were analyzed. Annual tree height observations recorded for 7 years were also statistically analyzed. Average (of four layers) CVs of the 1 N Ammoniumacetate-DTPA-extractable sodium (Na), sodium adsorption ratio (SAR), pH, and phosphorus (P) were 123, 117, 66 and 65%, respectively. CV in tree height was also high and fluctuated around 26% during the last 3 years. Two-way average trend variations were 41, 18, 20, 19, and 24% of variance of pH, SAR, Na, P, and tree height, respectively. About 29,43,55, 23, and 33% of variance of pH, SAR, Na, P, and tree height was locally spatially structured. Cross-variability among soil layers and SAR with tree height was predominantly structured. Mapping with optimal, unbiased, and minimum estimation variance was also validated. Delineation of soil sodicity with regionalized variable theory showed areas requiring different amounts of ameliorants. This technique was useful for optimizing soil reclamation.