Submitted to: International Symposium on Precision Agriculture
Publication Type: Proceedings
Publication Acceptance Date: 7/14/2004
Publication Date: 2/14/2005
Citation: Shillito, R.M., Timlin, D.J., Fleisher, D.H., Reddy, V.R., Quebedeaux, B. 2005. Landscape-Scale Agricultural Experimentation: Spatial Autocorrelation and Potato Yield. In: Proceedings of the 7th International Conference on Precision Agriculture and Other Precision Resources Management, July 25-28, 2004, Minneapolis, Minnesota. 2005 CDROM. Interpretive Summary: Variability of soil and terrain presents problems in agricultural research where scientists rely on limited numbers of field plots to quantify the effects of treatments such as fertilizer or irrigation on plant growth and yield. But agriculture takes place on a landscape with continuous and varied properties such as elevation, soil type, and nutrient and soil water status. These properties are related to each other in time and space and to crop yield and can confound statistical analyses. Our goal in this study was to quantify the response of potato to nitrogen through the use of continuous plots laid out over a gradient in soil properties. A further goal was to develop statistical methods to analyze data that are spatially related. We found that yield responds differently to nitrogen fertilizer mainly as a function of slight changes in soil texture. The effects of soil texture on yield response could be mapped using information from neighboring plots. These results are useful not only to understand the spatial interaction of yield and soil properties, but to assess the effectiveness of fertilizer application on an entire field. Providing management recommendations based on spatial experimentation can provide better information than results based on small, discrete plots which ignore or negate the spatial dimension of the agricultural landscape.
Technical Abstract: Spatial variability of field conditions has been recognized as problematic in agricultural experimentation. Traditionally, spatial variability has been either ignored or removed from analyses, but spatial variability may be indicative of underlying patterns and processes. By sampling on a continuous field scale in contrast to a discrete plot scale, these patterns can be revealed and used to understand and predict spatially varied responses. The objective of this study is to explore the response of potato yield to spatially varied nitrogen application, soil nitrate, and soil texture. We planted a 0.18 ha field with potatoes. A one-time application of four levels of nitrogen fertilizer was applied in a continuous sinusoidal pattern along the length of the field. Ultimately, this allowed for the estimation of whole field response to nitrogen fertilizer subject to initial soil conditions, essentially increasing the number of experiments by four.