Comparisons were performed using several different analysis techniques. Diurnal hourly-mean rainfall rates were used to test how well the model simulated the early morning and late afternoon rainfall maxima. Percent frequency distributions of rainfall rate were computed to determine how well the model simulated the distribution of rainfall rates throughout the period. Model mixing ratio values were used to derive reflectivities for a more direct comparison to the observed reflectivity values. These derived reflectivity fields were then compared using cumulative frequency by altitude diagrams (CFAD). After a convective and stratiform partitioning scheme was applied to the reflectivity data, an analysis of the modality of the maximum convective column height of the 30 dBZ contour was performed.
Results indicate that the ice microphysical parameterizations were too simple and thus unusable to accurately simulate the ice phase processes. Model-derived reflectivity values near and below the melting layer were much too high as a result of the erroneous simulation of the ice phase processes and liquid precipitation in the warmer region below. The CRM did simulate the general patterns observed in the radar data rather well. In addition, results of the comparison showed good agreement in the diurnal cycle of rainfall rates over the entire period.