FULL FACTORIAL DESIGN VERSUS CENTRAL COMPOSITE DESIGN: STATISTICAL COMPARISON AND IMPLICATIONS FOR SPRAY DROPLET DEPOSITION EXPERIMENTS

摘要

The optimization of spraying parameters to improve pesticide efficacy and reduce environmental impacts requires a large number of experimental treatments. Some experimental designs can be used to reduce the required number of treatment combinations. A comparison between a 33 factorial design and a central composite design (CCD) regarding the model fitting ability and their respective capability in identifying significant coefficients and optimum sprayer operating conditions is presented. Experiments were conducted in a spray chamber simulating an air-assisted spray boom on broccoli and potatoes. The independent variables studied were air speed, airflow, and air jet angle, and the dependent variable was leaf area coverage. Small disks of Kromekote cards were used to collect the spray dyed with the fluorescent tracer Blankophor BA™. Disks were placed on the upper and under side of leaves in the top, middle, and bottom portion of the crop canopy. Results showed that both experimental designs performed equally well. Both designs lead to the same conclusions regarding the optimum settings of the operational sprayer parameters. Overall, the CCD tended to be more conservative than the factorial design. Theoretical considerations and the fact that the CCD requires fewer experimental units, justifies the recommendation to use of CCD in spray droplet deposition experimentation.