An algorithm for retrieving polarimetric variables from numerical model fields is developed. By using this technique, radar reflectivity at horizontal polarization, differential reflectivity, specific differential phase shift and correlation coefficients between the horizontal and vertical polarization signals at zero lag can be derived from rain, snow and hail contents of numerical model outputs. Effects of environmental temperature and the melting process on polarimetric variables are considered in the algorithm. The algorithm is applied to the Advanced Regional Prediction System (ARPS) model simulation results for a hail storm. The spatial distributions of the derived parameters are reasonable when compared with observational knowledge. This work provides a forward model for assimilation of dual linear polarization radar data into a mesoscale model.