|MILLIKEN, GEORGE - Milliken And Associates|
|MCCARTER, KEVIN - Louisiana State University|
Submitted to: Operational Research: An International Journal (ORIJ)
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/2/2009
Publication Date: 11/1/2010
Citation: Milliken, G., Willers, J.L., McCarter, K., Jenkins, J.N. 2010. Designing experiments to evaluate the effectiveness of precision agricultural practices on research fields. Part 1. Concepts for their formulation. Operational Research: An International Journal (ORIJ). 10:329-348.
Interpretive Summary: Variable-rate controllers on farming equipment represent novel sources of geographically referenced information useful for the analysis of precision agricultural management decisions. Field topography and crop attributes, when captured by remote sensing systems, are other sources of information useful for the analysis of site-specific decisions. This paper describes how these diverse sources of georeferenced agricultural information can be statistically modeled to build an analysis approach useful for the analysis of precision agricultural experiments conducted upon research fields.
Technical Abstract: The objective of this paper is to present a unique formulation methodology for designing experiments to evaluate the effectiveness of a precision agricultural practice on a research farm field. We demonstrate an efficient method of combining the georeferenced treatment structure and the georeferenced design structure that describe and are necessary for the analysis of the site-specific experiment. Covariates included in this methodology are one or more layers of information (obtained by various remote-sensing systems) that geographically describe the topography of the research field and/or its crop attributes. This information also defines the geographical zones of a field where one or several rates (levels) of site-specific treatments can be assigned. The effectiveness of site-specific practices on yield in each zone are evaluated by a geographically-based statistical analysis. Concepts are illustrated using a simplistic hypothetical field. Current limitations are also discussed.