Assessing the impact of high-input management for reducing soybean yield gaps on high-productivity farms

To increase soybean actual yields (Yₐ), modern farmers are interested in exploring novel intensification options. We hypothesize that the inclusion of a high-input treatment in high-productivity systems may potentially reduce yield gaps (Yg), representing additional profits for farmers. Our objectives were to: (i) estimate Yₐ, water-limited yield potential (Ywₗᵢₘ), and Yg; (ii) evaluate a high-input treatment as a management option for reducing Yg; (iii) identify management and environmental factors associated to yield response to inputs and, (iv) determine the economic outcome of this high-input management on high-productivity soybean crops. We carried out 27 on-farm field experiments comparing a conventional management system versus a novel high-input treatment across the Central Pampas of Argentina. The conventional treatment was the most common management prescribed by farmers in the region. The high-input treatment included a set of inputs that were not usually applied and involved soil fertilizers with S and Zn, seeds treated with insecticides, fungicides, and inoculants, stress-reducing product, Ca and B foliar fertilizers and preventive insecticide. The Ywₗᵢₘ was estimated using crop simulation models. Yield gaps were calculated as the difference between Ywₗᵢₘ and Yₐ. Average Yₐ and Ywₗᵢₘ were 4571 and 5100 kg ha⁻¹, respectively, with a Yg lower than 30%. High-input treatments promoted a yield response of 131 kg ha⁻¹ across sites (p < 0.001), reducing Yg by 2.6%. Yield response to high-input treatment was positively correlated with Yg and soil pH, but negatively correlated with soil P. No positive economic outcomes were obtained with the high-input treatment. This study showed low benefits of the high-input treatments for reducing Yg in high-yielding soybeans managed by high-productivity farmers. Future yield increases in this type of systems would rely on genetic improvement of soybean yield potential and exploiting genotype x management x environment interactions.