Crop functional diversity alters soil fertility within two years of low-input successional agroforestry

Successional agroforestry systems (SAFS), diversified and dynamic intercrops of herbs and woody species inspired in forest succession, have been promoted for their expected positive biodiversity effect on restoring biogeochemical cycling in low-input, oligotrophic agroecosystems. Using a trait-based approach during the initial establishment of SAFS, we aimed to assess the contribution of trait dissimilarity (Functional Diversity, FD) and dominant trait values (community-weighted mean, CWM) to crop aboveground biomass production and soil fertility. We implemented an experiment in humid subtropical Brazil (soil: 10 % clay; pH 5.4; 3.7 % organic matter) with three treatments consisted of three agroforestry mixtures with constant species richness but contrasting functional composition of crops and trees based on the functional trait Leaf Nitrogen Concentration (LNC). Structural Equation Models confirmed our general hypothetical model in which crop functional composition affects aboveground biomass and soil chemical properties. FD was positively related to soil C:N ratio (estimate β ± SE: 3.97 ± 0.94), soil Total Organic Carbon (estimate β ± SE: 6.49 ± 1.88) and mineral Nitrogen (estimate β ± SE: 0.01 ± 0.002). Crop biomass was positively related to the Sum of Bases (K+Mg+Ca; estimate β ± SE: 0.33 ± 0.15). We conclude that niche complementarity, along with the vegetation quantity hypothesis, better explained the changes in our soil data; and suggest that by maintaining a manageable level of species richness, it is possible to optimize agroecosystem functions related to soil fertility by selecting a functional structure with sufficient diversity, based on plant leaf traits such as LNC.