Soil spatial variability: Structures, models, and their effects on crop yield variability

The current crop production practice in central Illinois involves uniform application of inputs to the soil. Variable rate technology (VRT) may boost crop production by reducing the problems associated with uniform input application. However, VRT necessarily requires that spatial variability in soil properties in any one field be characterized and mapped in detail for meaningful variable rate application. The objectives of this study were to determine and map the spatial variability structures, the spatial relationship between crop yield and soil properties, and develop a model to evaluate the crop-soil production system.; Geostatistical methods were used to study and map spatial variability structures in soil properties, and the spatial relationship between crop yield and these properties. Effects of variability in soil properties on crop yield were determined using principal component analysis and component regression techniques.; Spatial variability in most soil properties was best described by a spherical semivariogram model. However, soil properties associated with elevation changes or erosion-deposition dynamics such as Ahorizon thickness, percent sand and clay, and available water capacity were best described by the exponential model, while properties affected by tillage operations such as hydraulic conductivity were pure nugget. The range of spatial correlation, was 80–60 m, 60–100 m, and 70–160 m for morphological, physical and hydraulic properties respectively.; Correlograms indicated significant spatial correlation between crop yield and elevation (r = −0.5 and r = 0.5), organic matter (r = −0.3, r = 0.3), hydraulic conductivity (−0.1, r = 0.1), percent sand (−0.15, r = 0.15), and available water capacity (−0.15). The range of spatial correlation varied from zero to about 300 m.; Principal component analysis showed that only 15 of 51 variables accounted for more than 80% of total soil variability. Component regression accounted for about 52%, and 39% of soybean and corn yields variability respectively. Particle size distribution and organic matter by influencing moisture and nutrient dynamics, were the most important factors determining yield variability in these soils.