Carbon dioxide fertilization of primary production has the potential to offset soil respiration increases caused by warming. However, increased primary production will only be meaningful if additional fixed carbon is stored in slowly cycling pools. We used a three-pool soil-carbon model to compare the effects of ambient and elevated CO2, nitrogen, and biodiversity (1 or 16 species) on soil carbon storage in grassland plots located at Cedar Creek Ecosystem Science Reserve, Minnesota, USA. Soils in this region are nutrient poor, sandy and homogenous.  Fast and slow pool sizes and decay rates were determined using a 391 day soil respiration incubation, while recalcitrant pool sizes were determined using acid hydrolysis digestion. Estimated pool sizes and decomposition rates were compared with a mixed-effects ANOVA, with and without total plant biomass as a covariate.

Results/Conclusions

Averaged across all treatments, the majority of soil carbon was in the slow carbon pool (67%), with 2% in the fast pool and 31% in the recalcitrant pool. Mean residence times for fast and slow pool carbon were 19 days and 9 years respectively. Diverse plots had larger fast carbon pools (P<0.0001, R²=0.18) and slower fast pool decomposition rates (P =0.001, R²=0.10) than one species plots. Species richness also increased the decomposition rate of the slow carbon pool (P=0.0103, R²=0.02), without changing the slow pool size (P=0.9751), suggesting that species richness also increased inputs to the slow pool. Similarly, elevated CO2 increased the decomposition rate of the slow carbon pool (P=0.0385, R²=0.02). Despite this, elevated CO2 also marginally increasing the size of the slow pool (P=0.0604, R²=0.15), suggesting that enhanced input rates outpaced increases in decomposition rates in elevated CO2 plots. Finally, the recalcitrant pool size increased in nitrogen addition plots (P=0.0317, R²=0.10). Species richness effects were rendered insignificant when total plant biomass was taken as a covariate. Total plant biomass left only two significant treatment effects: increased slow pool size in CO2 enriched treatments (P=0.0493, R²=0.15), and marginally increased recalcitrant pool size in nitrogen addition treatments (P=0.0532, R²=0.09). Combined, these results suggest that species richness influences soil carbon through associated changes in total plant biomass. Carbon dioxide enrichment may increase carbon storage in the slow pool, and nitrogen addition may increase carbon storage in the recalcitrant pool, both independent of plant biomass effects. Overall, our results suggest that elevated CO2, nitrogen and plant diversity may increase soil carbon storage in grasslands.