Estimating root: shoot ratio and soil carbon inputs in temperate grasslands with the RothC model

Carbon inputs to soil are mostly site- and management-nonspecific estimates based on measured yield. However, in grasslands most carbon input is root-derived and plant carbon allocation patterns are known to vary strongly across sites and management regimes. The aim here was to estimate carbon inputs by fitting the RothC model to time series of soil organic carbon (SOC) data from field sites and to explain the observed variability in root: shoot ratio (R:S). Time series of SOC stocks in 15 different temperate grasslands were simulated using eight different literature-derived R:S values, which were compared to the optimised, site-specific R:S. The model-derived root inputs were validated with literature-derived root biomass data. A single, static R:S for yield-based carbon input estimation for all grasslands was not appropriate. Nitrogen fertilisation (R-2 = 0.57) significantly reduced the optimised R:S, which can be explained by the higher investment of plants in roots for nitrogen acquisition under nitrogen deficiency. The average R:S derived was 5.9 +/- 1.9 for unfertilised soils and 2.4 +/- 1.5 for fertilised soils. The results enable distinction of unfertilised and fertilised temperate grasslands regarding carbon input parameterisation for the RothC model and highlight the importance of nutrient regime for the carbon cycle.