Radar data assimilation increases the accuracy of precipitation prediction in a numerical model by assimilating hydrometeor information and by producing analysis fields similar to the actual atmosphere. Single polarization (single-pol) radar data assimilation has limits because it has hydrometeor classification method from error prone model background temperature field and radar operators only consist of reflectivity. Dual polarization (dual-pol) radar can obtain information on the characteristics and types of precipitation particles in the atmosphere. Dual-pol radar data has the potential to improve the accuracy model forecasted precipitation compared to single-pol radar.
This study conducted radar data assimilation through hydrometeor operators using dual-pol radar variables and direct classification algorithms to improve the accuracy of precipitation prediction not obtained by single-pol radar data assimilation.
In cases 1 and 3, dual-pol radar data assimilation experiments improved the accuracy of precipitation prediction compared to CTRL and single-pol radar data assimilation experiments by applying the dual-pol radar hydrometeor classification fields and hydrometeor operators with dual-pol radar variables. However, case 2 showed that the hydrometeors increased by dual-pol radar variables increased the temperature of the atmosphere due to latent heating, resulting in less accurate prediction than the CTRL. The model background temperature adjustment must be accompanied when hydrometeors are injected in the model atmosphere. Overall, the dual-pol radar hydrometeor operators contribute to an increase of the absolute amount of assimilated hydrometeors and the dual-pol radar hydrometeor classification method contributes to the redistribution of vertical hydrometeor profiles. The accuracy of precipitation forecast experiment show that when the dual-pol radar operators and hydrometeor classification fields are both applied at the same time, the performance is maximized.