Main content area

Modelling soil organic carbon dynamics in two long-term experiments of north-eastern Italy

Lugato, E., Paustian, K., Giardini, L.
Agriculture, ecosystems & environment 2007 v.120 no.2-4 pp. 423-432
agroecosystems, soil organic matter, land management, carbon, biogeochemical cycles, long term experiments, nutrient management, sandy soils, clay soils, peat soils, crop rotation, crops, crop residues, organic fertilizers, animal manures, Italy
Simulation models are widely used to assess the impacts of management and environmental variables on soil organic matter dynamics, to address questions on ecosystem sustainability and carbon cycling under global change. We tested the Century ecosystem model for two long-term experiments in north-eastern Italy: one (SF) comparing nutrient management treatments in small confined plots containing widely contrasting soil types (i.e., sandy, clay and peat) and the other (CR) involving a field study with crop rotation, nutrient, and management intensity variables. The organic matter changes in the SF experiment, showed a strong, linear relationship with C inputs from crop residues and added manures in the sand and clay soils, which was closely mimicked by the model. There was a net loss of soil C for all treatments in the peat soil, but the rate and overall magnitude of C losses were accurately simulated by the model, which suggested that treatment effects on soil C inputs was the major determinant of SOC dynamics in all three soils. In the CR experiment the model reasonably simulated the large initial decline (averaging about 30% of initial levels) in SOC observed in all treatments, as well as mean treatment effects over the course of the experiment. The model predicted a general pattern of higher SOC in the high management intensity, high fertility treatment combinations and lower SOC in the low management intensity, low fertility treatments; however, observed soil C did not show a clear pattern related to the treatments. Simulated soil C contents were linearly related to C input levels in the different treatments while there was no significant relationship between measured soil C and C inputs based on observed data.