Spatial Variability of Soil Fertility and Identification of Management Zones on Hummocky Terrain

A field study was conducted on hummocky terrain at the Manitoba Zero Tillage Association Research Farm near Brandon, Manitoba, Canada to determine the variability of crop yield as related to landscape position, soil properties, weed populations and plant disease. This information was used to develop a method for delineation of management units related to precision farming. Variable-rate fertilizer management systems can improve efficiency of fertilizer use and environmental sustainability. Adoption of this technology has been hampered due to the difficulty of classifying fields into management units, the high cost of sampling soils on a grid basis, and the variability of soil and plant properties in the landscape. Soil fertility, surface curvature, crop yield, weed populations and plant diseases varied considerably across the landscape in this study. Yield zones were identified based on cluster analysis of 5 years of normalized crop yield data. Although distinct yield zones were identified, soil nitrate-nitrogen and soil phosphorus were highly variable within these zones. Furthermore soil fertility and crop yield varied considerably between years within yield zones. This variability was attributed to excess soil moisture and flooding in the landscape. These sources of spatial and temporal variability of crop yield and soil fertility will be considered in the development of management strategies for variable application of nitrogen fertilizer. Introduction Variable-rate fertilizer management systems can improve efficiency of fertilizer use and environmental sustainability. Adoption of this technology has been hampered due to the difficulty of classifying fields into management units, the high cost of sampling soils on a grid basis, and the variability of soil and plant properties in the landscape. Technology for variable rate fertilizer systems is available, but there is little information available related to yield response in clay soils on hummocky terrain, and the relationship of plant tissue test levels in relation to soil fertility as measured by soil test nitrate nitrogen. Current soil test recommendations for nitrogen are based on soil test nitrate nitrogen from samples bulked from samples in several locations in the field preferably grouped according to topography. Variability of nitrate nitrogen can vary considerably in landscapes, Cahn et al. (1994) determined the spatial pattern of nitrate nitrogen changed with time, and the frequency distribution was skewed due to a few outlying values, which greatly exceeded the mean. Solohub et al. (1999) concluded mineral nitrogen is highly variable both across the landscape and over distances less than one meter. Current research on variablerate fertilizer management has focused on potential crop yield in relation to factors such as erosion and soil fertility as it varies from upper to lower slope positions. Variability of crop production has been related to the thickness of topsoil (Gregorich and Anderson 1985, Pennock and De Jong 1990) and soil moisture (Verity and Anderson 1990). Previous research (Moulin et al 1994) showed crop yield and soil erosion were related to a multivariate factor composed of relative elevation and surface curvature. Variability of crop yield in a glacial-till landscape was attributed to the effect of relative elevation and surface curvature on overland flow of water and related soil properties. Subsequent research (Beckie et al 1997) evaluated variable rate fertilizer management in a field classified according to slope position, organic matter and residual soil test nitrogen. The highest efficiency of fertilizer use was found when nitrogen fertilizer was variably applied to land classified according to topography.