The potential for site-specific management of soil and corn yield variability induced by tillage

Site-specific management (SSM) for agriculture is a recent concept in soil and crop management that involves the management of the localized conditions within a field. A required first step in the development of SSM strategies is the accurate assessment of the spatial variability of factors, both natural and use-dependent, which regulate the soil-plant-atmosphere system. Therefore, accurate knowledge of variability in soils and landscapes and understanding its relationship to variability of crop yield and/or quality is essential to SSM. This study examines three specific objectives relative to understanding the variability in soils and crops within a field. First, this study assesses the value of soil survey in delineating site-specific management zones by examining the quality of soil map units within an existing soil survey in terms of map unit purity and assessing the magnitude and nature of the spatial variability in soil physical properties as expressed by different sampling designs. Second, this study examines how corn yield varies spatially, how tillage alters soil properties and corn yield, and the extent to which variation in soil properties relate to variability in corn yield. Third, this study evaluates the potential for proven crop simulation models to account for spatial variability in corn yield within a field. A field study was initiated in 1995 to quantify how tillage affected the spatial variability in soils and corn within 3.72 ha field in central Michigan relative to its potential for site-specific management. Variability was assessed by applying different sampling schemes and measuring soil profile characteristics, soil physical properties and carbon contents of the surface horizon, soil water content and water table depth over time, and productivity of corn over two growing seasons. The soils within the experimental area were quite variable and did not correspond well to the soil map units given in the county soil survey maps. The current soil survey for this field is not accurate and of little value to SSM. The 30.5 m grid described the average condition well but did not accurately describe the spatial variability in soil physical properties. An intensive geostatistical sampling design produced very good estimates of the semivariogram and indicated a strong spatial dependence of all soil physical and soil texture properties measured. Tillage altered soil physical properties and soil water availability spatial structure and altered the spatial variability of corn grain yield differently each year. Soil properties and soil water measurements were of little value in explaining corn grain yields using stepwise regression techniques. However, simulation of the variability in corn yield using the CERES-maize crop simulation model correlated well to measured yields for 1995 but failed to account for water flux from the water table that appeared to regulate corn yields in 1996. There is a potential for crop models to simulate the spatial and temporal variability in yields. Overall, the potential for SSM in this field may be limited by a lack of correspondence between the spatial variability in soils and the temporal nature of variability in corn yield observed during the two years of this study.