The impact of five hydrometeor species provided by the fifth generation of ECMWF ReAnalysis (ERA5) on the skill of precipitation forecasts is investigated in very-short-range numerical weather prediction modeling using WRF. Three sets of experiments using different numbers of hydrometeors as input in the initial and lateral boundary conditions were designed for 6-h forecasts run over the region of Korea in East Asia during two one-month periods of the summer and winter seasons: specific humidity only, additional specific cloud liquid water and cloud ice water, and additional specific rain water and snow water. The 6-class Thompson microphysics parameterization scheme that uses prognostic condensate variables was adopted for forecasts to represent the effect of using these hydrometeors in the initial conditions. Simulated precipitation was verified against surface observations in terms of both categorical scores based on contingency tables and object-oriented measures. The evaluation of the precipitation skill reveals that the forecasts using multiple hydrometeors as input generally show increased skill of precipitation forecasts compared to forecasts using only specific humidity as input. It is found that the increase of precipitation forecasting skill appears both in winter and in summer. Using more hydrometeors in the initial conditions reduces the spin-up time and accelerates the occurrence of resolved precipitation in the model. This effect is clearly shown in the first one to two hours of the forecast. Afterwards, the precipitation forecast skill is similar to that of forecasts that use only specific humidity as input in the initial conditions. The more hydrometeor species used in the initial conditions of the forecast model, the more skillful becomes its precipitation forecast in the initial hours. Another set of experiments using liquid- and ice-phase hydrometeors separately shows that initialization with ice-phase hydrometers would be slightly more important to improving precipitation forecasts. Note that the results reported here should be understood as valid under the specific experimental configuration in this paper.
All Science Journal Classification (ASJC) codes
- Atmospheric Science