Forecasters in the Ohio River Valley region constantly face the challenge of precipitation type determination during the cool season. This paper focuses on the development, description, and demonstration of a scheme for precipitation type forecasting using output from a local implementation of the NCAR/PSU MM5 model. The backbone of this method is an examination of the low-level precipitable hydrometeor mixing ratios that are directly output from the microphysical parameterization and other thermodynamic processes within the model. The final determination of surface precipitation type can then be forecast from a relatively simple set of additional near-surface model output variables. Charts of these quantities were produced operationally at the National Weather Service in Louisville during the 2002-2003 boreal winter season. This method has demonstrated considerable skill at forecasting varying precipitation types on the mesoscale during significant winter storm situations, along with increased forecaster situational awareness regarding the potential for mixed precipitation. Limitations to this particular technique include constraints on the local area model’s solution governed by the initialization and boundary conditions, along with a specific combination of microphysical, convective, and boundary layer parameterizations.