EXPERIMENTAL-DESIGN AND DATA MANAGEMENT OF ROTATION EXPERIMENTS

Well summarized data from long-term cropping systems experiments give agriculturists and others direction for improving and sustaining crop productivity in a cost-effective manner while recognizing environmental concerns. Long-term agronomic studies range from a few demonstration plots to multisite, replicated and randomized trials with complex experimental and treatment designs. Basic design principles need to be used with additional requirements, e.g., each phase of each cropping sequence should occur each year. Data management becomes more complex as multiple crop-soil variables are measured over time. The multiyear data must be stored efficiently and must be easily retrievable to users with varied uses. Analysis-of-variance procedures are relevant for evaluation of interactions and are followed by specialized modeling applications. The commonly held statistical assumptions of experimental error for annual experiments may not be valid in long-term studies where yields are measured on the same plots over years. Selection of an analysis approach depends on the assumed error structure. The renewed interest in existing long-term studies encourages innovative approaches to data management, graphics and experimental modeling. The challenge is to make summarized data, including means, other model estimates and standard errors, accessible to users in a format that answers their questions.