Main content area

Modelling the dynamic physical protection of soil organic carbon: Insights into carbon predictions and explanation of the priming effect

Luo, Zhongkui, Baldock, Jeff, Wang, Enli
Global change biology 2017 v.23 no.12 pp. 5273-5283
data collection, models, prediction, soil, soil organic carbon
The role and significance of physically protected soil organic carbon (SOC) in regulating SOC dynamics remains unclear. Here, we developed a simple theoretical model (DP model) considering dynamic physical protection to simulate the dynamics of protected (Cₚ) and unprotected SOC (Cᵤ), and compared the modelling results with a conventional two‐pool (fast vs. slow) model considering chemical recalcitrance. The two models were first constrained using extensive SOC data collected from soils with and without fresh carbon (C) inputs under incubation conditions, and then applied to project SOC dynamics and explore mechanisms underpinning the priming effect (PE). Overall, both models explained more than 99% of the variances in observed SOC dynamics. The DP model predicted that Cₚ accounted for the majority of total SOC. As decomposition proceeds, the proportion of Cₚ reached >90% and kept relatively constant. Although the similar performance of the two models in simulating observed total SOC dynamics, their predictions of future SOC dynamics were divergent, challenging the predictions of widely used pool‐based models. The DP model also suggested alternative mechanisms underpinning the priming of SOC decomposition by fresh C inputs. The two‐pool model suggested that the PE was caused by the stimulated decomposition rates, especially for the slow recalcitrant pool, while the DP model suggested that the PE might be the combined consequence of stimulated Cᵤ decomposition, the liberation of Cₚ to decomposition and the inhibition of the protection of unprotected SOC. The model‐data integration provided a new explanation for the PE, highlighting the importance of liberation of initially physically protected SOC to decomposition by new C inputs. Our model‐data integration demonstrated the importance of simulating physical protection processes for reliable SOC predictions, and provided new insights into mechanistic understanding of the priming effect.