Nitrogen (N) fertilizer inputs to agroecosystems are the largest contributor to reactive N production globally, and N losses from agroecosystems are a persistent source of environmental pollution. An emerging ecological framework for soil nutrient management demonstrates that N losses are reduced when N and carbon (C) cycles are coupled by increasing C-sinks for reactive N in fields. Increasing the functional diversity of crop rotations with cover crops (non-harvested crops) can build soil organic C, supply N through microbial processes, and has the potential to balance N inputs with harvested exports. Cover crop mixtures that combine complementary functional groups – such as legumes and grasses – could increase provisioning of multiple ecosystem functions at once, including N supply and N retention. However, key questions remain unanswered: how do plant traits such as biological N fixation vary with interspecific interactions in mixtures? And, how do ecosystem functions from cover crops vary with soil properties? This study addressed these questions in a two-year, on-farm experiment with a hairy vetch (Vicia villosa) - cereal rye (Secale cereale) mixture across ten working farms in southern Michigan. Participating farms spanned a soil fertility gradient due to differences in management history and soil type.
Results/Conclusions
Baseline soil parameters varied widely across farms: Total organic C varied from 24.8 to 56.8 Mg C ha-1, occluded particulate organic matter N concentration also varied twofold, from 0.8 - 1.7%N, and Bray-1 phosphorus ranged from 6 - 98 ppm. Anaerobic soil incubations identified rates of potentially mineralizable N from 7.2 – 24.7 mg N kg soil-1 week-1. Linear regression models were used to identify which soil parameters predict biological N fixation by vetch in mixture and monoculture. In terms of ecosystem functions of cover crops, in both experimental years the percent of vetch N derived from fixation was significantly greater in mixture than in monoculture (93.3 and 86.5%, respectively); however, on average, total N supply from N fixation was greater in vetch monocultures due to competitive interactions with rye, and greater vetch biomass in monoculture. There was also evidence of facilitation, with vetch transferring approximately 8 – 11 kg of biologically fixed N ha-1 to rye in mixtures, and the C:N of rye in mixture was also significantly lower than monoculture rye. Results of this on-farm experiment advance ecological understanding of the mechanistic role of functional trait diversity within agroecosystems within a context highly relevant to farm management.