A trial network answers a harder question than a single trial: which hybrid and treatment combination performs, and where. Scattered protocols, different raters and uneven scoring dates make the interaction terms unreadable, and the recommendation ends up being made site by site.
PhenoScale flights under one protocol at every site measure Green Cover, Plant Biovolume and NDVI on every plot and date. Cloverfield models the hybrid × treatment × site interactions across the network, with analysis of variance and post-hoc comparisons on the season-long response.
Stable, high-performing combinations are identified across the network so each product is positioned where it wins, turning a scattered trial network into a coherent, interaction-aware recommendation for maize and soybean.
