Agroecosystem diversification increases soil N availability and vegetable yields across a soil fertility gradient in southern Brazil

Anne Elise Stratton and Jennifer Blesh, School for Environment and Sustainability, University of Michigan, Ann Arbor, MI, Jucinei José Comin, Departamento de Engenharia Rural, Centro de Ciências Agrárias, Universidade Federal de Santa Catarina (UFSC), Florianopolis, Santa Catarina, Brazil, Ilyas Siddique, Dep. de Fitotecnia, Centro de Ciências Agrárias, Universidade Federal de Santa Catarina, Florianopolis, Santa Catarina, Brazil, Letícia Dambroz Filipini and Renata Lucas, Cepagro – Centro de Estudos e Promoção da Agricultura de Grupo, Florianopolis, Santa Catarina, Brazil

Background/Question/Methods
Diversified cropping systems can support multiple ecosystem functions while maintaining crop yields. However, the impacts of diversification on ecosystem functioning may vary depending on environmental conditions and management histories on farms. We conducted a two-year experiment focused on two diversification practices designed to increase the presence of legumes in cropping systems—a grass-legume cover crop mixture followed by intercropping of non-legume and legume vegetable crops. We tested the effects of these practices across a soil fertility gradient on 14 farms in southern Brazil with different long-term management histories. We expected that farm management history, especially prior use of agroecological practices, would be an important predictor of soil fertility. In a region with weathered, semi-tropical soils, we tested the hypothesis that the vetch-oat cover crop mixture would increase soil nitrogen (N) availability relative to controls across farms, with the greatest N availability on agroecological farms. In contrast, we predicted that relative yields from cucurbit-pea intercrops would be highest on low-fertility farms due to increased facilitation. Piecewise structural equation modeling was our primary analytical approach, complemented by Principal Components Analysis and two-way mixed-effects ANOVA.
Results/Conclusions
As predicted, farm soils with longer histories of agroecological management had higher levels of soil fertility (marginal R2=0.34, p=0.02), captured by a principal component with high loadings for biochemical measures of fertility: soil organic carbon, potentially mineralizable carbon, and extractable phosphorus. On average, N inputs to soil from cover crop residues were 2.4-fold greater than N from weeds in fallow plots across farms and years. N mineralization (ANOVA, p=0.001) and inorganic N availability (ANOVA, p=0.05) were highest on established agroecological farms. Using piecewise structural equation modeling, we confirmed that soil inorganic N availability increased with cover crop presence (p=0.0003), and with soil fertility (p=0.008) and % clay (p=0.046). Soil pH was the strongest predictor of both aboveground N fixation by the legume (vetch) in the cover crop mixture (p=0.0008) and the Land Equivalent Ratio (LER) from intercropping (p=0.003), with lower pH constraining vetch N fixation and intercrop yields. Overall, intercrops overyielded relative to monocultures across farms and years (t-test, p=0.021; mean LER=1.19). Though interspecific facilitation between crop species can be heightened in less fertile soils, our results suggest that benefits to ecosystem functioning will continue to build as farms in nutrient-limited environments adopt diversification practices during transitions to agroecological management.