SPATIAL VARIABILITY OF SOIL PROPERTIES AND POTENTIAL MANAGEMENT CLASSIFICATION OF A CHERNOZEMIC FIELD IN LOWER AUSTRIA

Abstract

Knowledge of the spatial variability of soil physical and chemical properties is important for precision agriculture. Fuzzy classification and geostatistical interpolation are useful tools for the simultaneous handling of continuous variation in both attributes and location. The objectives of this study were: (i) to classify data to capture optimally in-field variability and (ii) to delineate contiguous areas for site-specific management. Several soil physical and chemical parameters for a loam soil were measured in the 0- to 15-cm depth of a 50- × 25-m grid at the experimental farm of the University of Agricultural Sciences, Vienna, Austria. Using Variowin and ArcGIS, models were fitted to variograms, and continuous maps of individual attributes were generated by spline smoothing. The coefficient of variation (CV) for soil bulk density (ρb), pH, and silt and clay contents was <15%; for antecedent water content (θ), sand content, electrical conductivity (EC), organic carbon concentration (TC), soil organic carbon (SOC) and soil nitrogen (TN) it was 15 to 35%; and for saturated hydraulic conductivity (Ks) and matric flux potential (Φm) it was >35%. The nugget:sill ratio showed moderate spatial dependence for silt content, ρb, SOC, Ks, Φm, EC, clay content, and TC (from 38% for silt content to 59% for TC) and weak spatial dependence for TN and pH (<75%). The nugget:sill ratio from covariograms showed moderate spatial dependence for Ks and TC, TC and clay content, ρb and pH, TN and clay content, Ks and Φm, and ρb and θ (38 to 62%). Four classes captured total within-field variability optimally. Despite a low level of actual statistical variation in data, contiguous areas of each class were present in the field. The study demonstrated that by careful collection of data, site-specific management zones can be created for soil management within the field.