Accurately estimating the amount of plant-available nitrogen (N) from soil and crop residues is important for N management in cropping systems.  The ability to accurately estimate the amount of N from cover crop mulch that will become plant available could enhance N use efficiency in agricultural systems.  Estimating the amount of N that will be mineralized from cover crop residues is challenging because of the complexity of the process and the variety of factors involved, including residue quality, temperature, water content, drying and rewetting events, and soil characteristics.  Because many factors are involved in the mineralization process, simulation models are useful tools for estimating N mineralized from cover crops.  Some of the most widely used models for simulating an entire crop-soil system are CERES models.  Previous work in Georgia showed that the N subroutine of the DSSAT (Decision Support System for Agrotechnology Transfer) family of models, CERES-N, could be calibrated to provide good estimates of N released from rye (Secale cereal L.), wheat (Triticum aestivum L.), oats (Avena sativa L.), and crimson clover (Trifolium incarnatum L.) residues when decomposing on the surface of a Cecil soil.  We compared the CERES-N simulation results with N mineralization results from an in-situ field study using soil cores amended with surface-applied cover crops. For 2011 the model over-predicted plant-available N from crimson clover and rye. The model is driven by soil water content among other factors.  During a dry year the water content of surface-applied cover crops is much lower than that of the underlying soil leading to over-prediction of plant-available N by the model. For 2012, the model more accurately predicted plant available N from crimson clover. More rainfall during the growing season most likely accounts for the increased accuracy of the model for 2012.