Several fundamental but important factors controlling the feedback of boreal organic carbon (OC) to climate change were examined using a mechanistic model of soil OC dynamics, including the combined effects of temperature and moisture on the decomposition of OC and the factors controlling carbon quality and decomposition with depth. To estimate decomposition rates and evaluate their variations with depth, the model was inverted using a global optimization algorithm. Three sites with different drainage conditions that represent a broad diversity of boreal black spruce ecosystems were modeled. The comparison among the models with different depth patterns of decomposition rates (i.e., constant, linear, and exponential decrease) revealed that the model with constant inherent decomposition rates through the soil profile was able to fit the observed data in the most efficient way. There were also lower turnover times in the wettest site compared to the drier site even after accounting for moisture and temperature differences. Taken together, these results indicate that decomposition (especially for the wetter site) was not accurately represented with standard moisture and temperature controls and that other important protection mechanisms (e.g., limitation of O₂, redox conditions, and permafrost) rather than low inherent decomposition rates are responsible for the recalcitrance of deep OC. The simulation results also showed that most of the soil CO₂ efflux is generated from subsurface layers of OC because of the large OC stocks and optimal moisture conditions, suggesting that these deeper soil OC stocks are likely to be critically important to the future carbon dynamics.