The potential for soil carbon (C) sequestration under short-rotation woody crops, like hybrid poplar (Populus spp.), is a significant uncertainty in our understanding of how managed tree plantations might be used to partially offset increasing atmospheric CO2 concentrations. Through development of a multi-compartment model, we reviewed information from studies on hybrid poplar and analyzed the potential impact of changes in plant traits and nitrogen (N) fertilization on soil C storage. For a hypothetical setting in the southeastern U.S.A., and starting from soils that are relatively depleted in organic matter (2.5 kg C m(-2)), the model predicted an increase in mineral soil C stocks (1.7 kg C m(-2)) over four 7-year rotations of hybrid poplar. However, at the end of the fourth rotation, both cumulative soil C gains and annual rates of soil C accrual (23-93 g C m(-2) yr(-1)) varied widely depending on fertilization rate, biomass yield, and rates of dead root decomposition (three factors that were examined in a factorial model-based experiment). Our analysis indicated that processes linked to genetically modifiable poplar traits (aboveground biomass production, belowground C allocation, root decomposition) are potential controls on soil C sequestration. Key measures of model performance were sensitive to how aboveground biomass production responded to N fertilization. Site specific properties that were independent of plant traits were also important to predicted soil C accumulation and point to possible genotype x site interactions that may explain contradictory data from both empirical and theoretical studies of C sequestration under hybrid poplar plantations. (C) 2010 Elsevier Ltd. All rights reserved.