Admittedly, the surface soil layer is the most active soil layer and contains the greatest concentration of soil organic carbon (SOC). However, soil layers deeper than 20 cm contain a greater absolute amount of SOC. Global SOC storage in the first meter of the soil is 1502 Pg, of which only 40 % is present in the top 20 cm. Consequently, it has been suggested that deep soil layers have a great capacity to sequester and that agroecosystems could play a unique role in the sequestration of SOC at depth by introducing deep rooting crop, grass and tree varieties. However, the critical assumption behind these suggestions is that SOC decomposition decreases with depth due to a less favorable environment for decomposition at depth and that extra C input at depth will therefore decompose substantially slower. However, we question the assumption that the environmental conditions favoring decomposition decrease drastically with depth. The observed increase in recalcitrance and turnover time with depth might be due to preferential vertical transport of recalcitrant C (the labile C is decomposed before it gets to lower soil layers). If this is true, the observed increased turnover time with depth is mainly caused by the presence of chemically recalcitrant C rather than by limiting environmental factors. Hence the potential for new deep-soil C storage is more limited than has been suggested. We have integrated experimental work with model development to elucidate and quantify C dynamics at depth. Our model uses a depth explicit version of the CENTURY model, in which soil C can be transported vertically through pedo-turbation. The model was validated on a grassland site in Iowa with profiles of mineral isotope tracers of various turnover times (Be-10, Pb-210, and Cs-137) and carbon isotopes (C-12, C-13, C-14).