Submitted to: Agriculture, Ecosystems and Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/28/2008
Publication Date: 1/1/2009
Citation: Shillito, R., Timlin, D.J., Fleisher, D.H., Reddy, V., Quebedeaux, B. 2009. Yield response of potato to spatially patterned nitrogen application. Agriculture, Ecosystems and Environment. 129:107-116. Interpretive Summary: The variability of field soil properties presents problems in agricultural research where scientists rely on a limited number of small plots to quantify the effects of treatments such as fertilizer on yield. We know how the plots respond but don’t know how the whole field responds. It is important to understand how soil properties interact with fertilizer applications and yield over the entire field because fertilizer shortages decrease yield, and fertilizer excess can have widespread environmental effects. We have to better understand how these soil properties vary over a field. Our goal in this two-year study was to quantify the response of potato to nitrogen fertilizer using a new experimental design and statistical techniques. We found that yield response to nitrogen fertilizer was sensitive to even slight changes in soil texture. We also found that low yielding areas of a field stayed low-yielding even when more fertilizer was added. These results are useful not only to understand the interaction between soil properties and yield, but to assess the effectiveness of fertilizer application over an entire field. Providing management recommendations based on whole-field experimentation and analyses can provide more accurate information for farmers and policy-makers than results based on small, discrete plots which cannot capture the spatial variability present in all agricultural fields.
Technical Abstract: The difference between experimental results and their application at the farm level is highlighted by the ever-present research issue of nitrogen fertilizer. Although crop response to nitrogen fertilizer has long been studied, inappropriate experimental designs led to the inadequate accounting of spatial variability in field and yield properties, and may be a major component of this gap. Analytical methods to explicitly account for spatial variability now exist, but the complementary modification of experimental design is still developing. There is a need to combine these analytical methods with non-traditional experimental design. A two-year study was implemented to assess the response of potato (Solanum tuberosum cv. Kennebec) yield to nitrogen fertilizer. We used a whole-field design where four nitrogen treatments (0, 56, 112, and 280 kg N ha-1) were applied systematically throughout the entire field in a sinusoidal pattern. Measured field properties included topography, soil texture, pre-application soil nitrate levels, and plant available soil water content. Yield was measured along two continuous two-row transects the length of the field. A random field linear model was used to simultaneously account for treatment effects, spatial covariance and soil properties along the transects. Specific areas of the field had consistently low yields for the four nitrogen levels over the two years of the study suggesting high rates of nitrogen leaching out of the root zone. Of the covariates, only clay content proved important in explaining yield variability over the field, but the impact of clay content was different each year. The results showed the effectiveness of whole-field experimentation, the importance of spatially distributed data, and the ease with which spatial variability can be addressed.