This work is part of a project to assess the role of synoptic-scale processes in controlling the occurrence of tornado outbreaks. Two of the primary goals of the project are to determine the extent by which synoptic-scale processes are responsible for the occurrence of tornado outbreaks and to pinpoint which synoptic-scale features (if any) at particular lead times are significant in the production of tornado outbreaks. Because even the most sophisticated operational numerical models do not explicitly forecast tornadoes, a method must be derived to distinguish between tornadic and primarily nontornadic severe weather outbreaks from model output. The use of meteorological covariates is necessary in order to determine if the model run is indeed making this distinction. To find these covariates, NCEP/NCAR Reanalysis data are used to initialize and provide boundary conditions for mesoscale model runs of 50 tornado outbreaks and 50 primarily nontornadic severe weather outbreaks. A 24-hour, 48-hour, and 72-hour forecast is run for each case. Each run has four nested domains. The mother domain uses 150 km grid spacing. The innermost domain has 2 km grid spacing and is positioned in the same outbreak-relative position for each case. The results of these simulations will be used to evaluate various candidate covariates. Three-dimensional composite fields of the case studies will then be initialized by the models to evaluate the covariates further, as well as to determine if a distinction between tornadic outbreaks and primarily nontornadic severe weather outbreaks can be made by the models.