Particulate and mineral-associated organic carbon turnover revealed by modelling their long-term dynamics

Abstract

Particulate (POC) and mineral-associated organic carbon (MOC) are measurable carbon pools with distinct function. Their turnover properties have been rarely assessed because of their contrasting stabilization and destabilization processes and the difficulty of in situ monitoring. In this study we used two carbon models (a three-pool and a four-pool model) driven by measured POC, MOC and a more resistant charred organic carbon (COC) pool to obtain insights into the turnover of POC and MOC and to predict long-term soil organic carbon (SOC) dynamics. The models were constrained by measurements of POC, MOC and COC at 159 long-term trials (average trial duration is 22 years) across Australian agricultural regions. Results showed that POC-, MOC- and COC-constrained models, particularly the three-pool model, have less parameter collinearity and predict less uncertainties in most model parameters as well as in SOC dynamics and its vulnerability, compared with the models constrained by total SOC alone. The three-pool model estimated an average decay rate of 0.46 and 0.044 yr−1 for POC and MOC, respectively, across the trials. The four-pool model provided additional insights that on average a large fraction of MOC (28% on average across the trials) is physically protected against decomposition, and the remaining fraction is labile with an average decay rate of 0.068 yr−1. The existence of protected MOC results in prolonged overall residence time of MOC. Additionally, we found that the turnover of POC and MOC are distinctly influenced by climate and soil properties with involvement of non-linear relationships. These results shed new lights on the dynamics of measurable functional SOC pools and demonstrate that models constrained by these pools can provide more accurate predictions of model parameters thereby more accurate projections of SOC dynamics compared to models constrained by total SOC only.