No abstract

No abstract

No abstract

One of the most critical aspects of agricultural experimentation is the proper choice of experimental design to control field heterogeneity, especially for large experiments. However, even with complex experimental designs, spatial variability may not be properly controlled if it occurs at scales smaller than blocks. Therefore, modeling spatial variability can be beneficial, and some studies even propose spatial modeling instead of experimental design. Our goal was to evaluate the effects of experimental design, spatial modeling, and a combination of both under real field conditions using GIS and simulating experiments. Yield data from cultivars was simulated using real spatial variability from a large uniformity trial of 100 independent locations and different sizes of experiments for four experimental designs: completely randomized design (CRD), randomized complete block design (RCBD), α‐lattice incomplete block design (ALPHA), and partially replicated design (PREP). Each realization was analyzed using different levels of spatial correction. Models were compared by precision, accuracy, and the recovery of superior genotypes. For moderate and large experiment sizes, ALPHA was the best experimental design in terms of precision and accuracy. In most situations, models that included spatial correlation were better than models with no spatial correlation, but they did not outperformed better experimental designs. Therefore, spatial modeling is not a substitute for good experimental design.

No abstract

No abstract

No abstract

No abstract