Improved models are needed to predict the fate of carbon in forest soils under changing environmental conditions. Within a temperate sugar maple forest, soil CO 2 efflux averaged 3.58 lmol m -2 s -1 but ranged from 0.02 to 25.35 lmol m -2 s -1 . Soil CO 2 efflux models based on temperature and moisture explained approximately the same amount of variance on gentle and steep hillslopes (r 2 = 0.506, p \ 0.05 and r 2 = 0.470, p \ 0.05 respectively). When soil carbon content and sorption capacity were added to the models, the amount of explanation increased slightly on a gentle hillslope (r 2 = 0.567, p \ 0.05) and substantially on a steep hillslope (r 2 = 0.803, p \ 0.05). Within the organic-rich surface of the mineral soil, carbon content was positively related and sorption capacity was negatively related to soil CO 2 efflux rates. There were general patterns of smaller carbon pools and lower sorption capacity in the upland positions than in the lowland and wetland positions, likely a result of hydrological transport of particulate and dissolved substances downslope, leading to higher soil CO 2 efflux in the upland positions. However, the magnitude of the soil CO 2 efflux was mitigated by the higher sorption capacity of the organic-rich surface layer of the mineral soils, which was negatively correlated to soil CO 2 efflux. More accurate estimates of forest soil CO 2 efflux must take into account topographic influences on the carbon pool, the environmental factors that affect rates of carbon transformation, as well as the physicochemical factors that determine the fraction of the carbon pool that can be transformed.