Soil C storage in productive grassland mixtures: the role of species traits and mixture composition

Productive perennial grassland systems can increase soil carbon (C) storage compared to annual cropping systems, but the effect of species mixture composition as a means to optimize soil C input and stabilization and aboveground biomass yield at low nitrogen (N) fertilizer inputs remains unexplored.

In a field experiment, we measured aboveground yield and soil C inputs in 2-species mixtures with grasses or forbs combined with red or white clover, and in multi-species mixtures with 6 and 18 species including all 3 plant functional groups (grass, forb, legumes). All mixtures were fertilized with 75 kg N ha^-1 yr^-1. Monoculture perennial ryegrass plots were established at low and high N application rates (75 and 300 kg N ha^-1 yr^-1). We assessed aboveground yield and the input to belowground C pools (i.e., root C and rhizodeposited C) using isotopic labelling with ¹³C and a tracer mass balance approach. Further, we measured the allocation of the rhizodeposited C into newly formed mineral-associated organic C (MAOC) and particulate organic C (POC) fractions and related these to root traits. Lastly, we quantified selected amino sugars as proxies for bacterial and fungal necromass along a species richness gradient (1, 2, 6, 18).

The mixtures with red clover (including the multi-species mixtures) had an aboveground yield between 19.0 and 20.8 t DM ha^-1 (830 – 890 g C m^-2), matching the yield of the high-fertilized monoculture perennial ryegrass. The 2-species mixtures with white clover yielded on average 27% lower than mixtures with red clover. Mixtures with higher species richness than 2 yielded similar aboveground biomass as the 2-species mixtures with red clover. The multi-species mixture, consisting of 6 productive, resource-acquisitive species, resulted in a total soil C input to 1 m depth of 425 ± 30 g root C m^-2 and 70 ± 10 g rhizodeposited C m^-2, which was higher than all other treatments. Red clover, tall fescue and chicory secured high root C, while white clover, perennial ryegrass and plantain contributed with high C rhizodeposition. The 18-species mixture had lower total C input to soil compared to the 6-species mixture, likely due to several extra species diluting the effect of the 6 productive species used in the 6-species mixture. Legumes (low C:N ratio in root biomass) increased the proportion of MAOC of total rhizodeposition, while grasses (high root length density and root surface area) increased total C rhizodeposition and the proportion of POC. Further, the mixtures with legumes had a higher content of fungal and bacterial microbial necromass in the soil at the end of the growing season compared to monoculture perennial ryegrass. This indicates that the stabilization potential of rhizodeposted C can be enhanced by mixtures with legumes compared to monoculture grasslands.

Our results showed how grassland mixture composition can 1) increase total C input to soil without compromising high aboveground yield, 2) regulate the relative proportion of MAOC and POC from root-derived C, and 3) increase microbial necromass formation and the potential persistence of newly formed soil C.