Why does mineral fertilization increase soil carbon stocks in temperate grasslands?

Globally, grasslands are a major land cover type and store significant amounts of soil organic carbon (SOC). Fertilization with the major plant nutrients nitrogen (N), phosphorus (P), and potassium (K) may affect SOC stocks, e.g., by altering aboveground and belowground plant productivity, species composition, litter composition and decomposition, and microbial metabolism. However, belowground responses to fertilization in grasslands are not fully understood, hampering accurate predictions of SOC dynamics. In this study, seven different long-term grassland fertilization experiments (16–58 years) in Germany and the Netherlands were sampled to determine the effects of mineral fertilization (i.e., N, P, K, PK, and NPK) compared with unfertilized plots (Control) on SOC stocks and root C stocks. Soils were sampled to a depth of 100 cm or to the maximum depth possible. Potential litter decomposition was assessed using Lipton Rooibos and Lipton Green teas as standardized and well-characterized litter materials. In the topsoil (0–30 cm depth), PK, NPK, and NPK+ (increased NPK) fertilization had significant positive effects on SOC stocks, with annual sequestration rates of 0.28, 0.13, and 0.37 Mg ha⁻¹ yr⁻¹, respectively, within an average time span of 34 (PK, NPK) or 20 (NPK+) years. For NPK fertilization, 1.15 kg of N was needed to sequester 1 kg SOC. Increased SOC stocks could not be explained by altered belowground C inputs or decomposition rates, since root C stocks were not affected by fertilization and potential litter decomposition was unchanged. The highly significant increases in dry matter yield with PK and NPK fertilization and resulting higher aboveground C inputs were also unlikely to explain the observed SOC stock changes, since increases were only observed at soil depths>10 cm. However, significantly narrower root C:N ratios were observed for the N (26.9), PK (36.5), NPK (36.8), and NPK+ (33.2) treatments than for the Control (41.8), which may have caused increased microbial C use efficiency, positively affecting SOC storage. The narrower C:N ratios in PK treatments were explained by significantly increased abundance of legumes. We concluded that the carbon footprint of fertilization-induced SOC sequestration needs to be considered when the latter should be accounted for as a climate mitigation measure. Thereby, PK fertilization has much lower CO₂ costs than NPK fertilization due to N input via legumes.