Submitted to: Agricultural Systems
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/2/2007
Publication Date: 1/1/2008
Citation: Willers, J.L., Milliken, G.A., Jenkins, J. N., O'Hara, C.G., Gerard, P.D., Reynolds, D.B., Boykin, D.L., Good, P.V., Hood, K.B. 2008. Defining the experimental unit for the design and analysis of site-specific experiments in commercial cotton fields. Agricultural Systems. 96:237-249.
Interpretive Summary: A method is developed to define experimental units in commercial agricultural fields in which site-specific management practices are applied. The method defines these units according to the largest tool bar (or boom) on farm equipment that applies the site-specific practice along paths of travel across the field. The intersection of the travel paths, the width of the tool bar, and prescription map indicating where different rates of inputs are to be applied creates the experimental unit. These units, even for similar application rates, differ in size. The value of defining these units is that it is then possible to use modern statistical methods to determine the effectiveness of the site-specific practices. The results of the analysis provide important feedback information to producers and consultants to improve making management decisions for future production seasons.
Technical Abstract: Traditional experimental designs based upon small plots have proven useful for investigating many aspects of cotton production. However, on larger scales such as an entire field (or sets of adjacent fields), the small plot experimental design approach is inadequate. The availability of remote sensing data, variable-rate equipped machinery, yield monitors, and other spatial sensor systems provide novel opportunities to apply management practices in a site-specific manner. These same spatial technologies, when coupled to general linear mixed model software, also offer a method of analysis. The experimental units for site-specific experiments in commercial fields are formed by relationships between the width of the largest field implement’s boom (or tool bar), its travel paths, and the decision makers’ intent to treat various areas differently. Replications of experimental units differ in number, size, shape, and interspersion within every travel path. A general linear mixed model can describe test statistics based on topological relationships among features (or themes) of agricultural topography, including traditional and site-specific management practices. Concepts and non-farm cotton variety trial that included a one-time, site-specific application of a plant growth regulator is analyzed.