The effect of stochastic factors on soil carbon makes the quantity of carbon generated under a sequestration project uncertain. Hence, the quantity of sequestered carbon may need to be discounted to avoid liability from shortfalls. We present a potentially applicable uncertainty discount and discuss difficulties that might arise in empirical use. We insist that the variance in historical crop yields across geographical areas is used to derive a proxy variance for forming an uncertainty discount for carbon projects. Application of our approach suggests that project level uncertainty discounts would be 15-20% for the East Texas region.