Simulation and validation of long-term changes in soil organic carbon under permanent grassland using the DNDC model

摘要

Long-term changes in soil organic carbon (SOC) are difficult to quantify experimentally because of measurement errors and high spatial and temporal variability. Modelling can help to provide a more robust assessment by reducing these uncertainties and reproducing greenhouse gas (GHG) and C exchange processes in an ecosystem by identifying key drivers. In this study, the Denitrification-Decomposition (DNDC95) model was used to evaluate SOC density (SOC rho) and its annual changes (Delta SOC rho) in temperate grassland soils, which received different forms of nitrogen (N) (i.e. inorganic and organic) and at different application rates for 45 years. We found that simulated values for SOC rho (0-15 cm depth) in unfertilized (54 t C ha(-1)) and fertilized soils (55 t C ha(-1)) were lower than measured values (73 and 77 t C ha(-1), respectively). Despite some variations, measured and simulated SOC rho was higher under cattle (88-99 vs. 66-116 t C ha(-1)) than pig slurry (75-78 vs. 55-69) applications, and increased with increasing rates of added C. Irrespective of nutrient treatment, overall mean sequestration rates were 0.46 +/- 0.06 (observed) and 0.37 +/- 0.01 (simulated) t C ha(-1) yr(-1). Simulated values explained 66% of the variability between years and treatments (slope: 1.41; intercept: - 34.58 t C ha(-1)) with reasonably good prediction efficiency The variations in simulated values could be explained by differences in applied N (63%), which were linked to differences in C (62%), rainfall (15%) and air temperature (11%). The model (R-2 0.77-0.99/-0.99; p < 0.05- < 0.0001) was sensitive to soil variables, and SOC rho increased with increasing clay fraction, bulk density and inherent SOC concentration. Finally, simulated (and measured) values suggest that a new SOC equilibrium was not reached even after 45 years of intensive management. The study demonstrated how DNDC provides reasonably accurate representation of how organic N applications, as well as key soil and climatic variables may affect SOC density changes over time.