P-42

Impacts of dense and frequent surface observations on sever rainfall forecasts

Y. Maejima (RIKEN), M. Kunii (Meteorological Research Institute), H. Seko (Meteorological Research Institute), R. Maejima (Meisei Electric Co., Ltd), K. Sato (Meisei Electric Co., Ltd) and T. Miyoshi (RIKEN)

To investigate the impacts of observations with high spatial and temporal density on local heavy rainfall prediction, we performed a numerical simulation at a 100-m resolution and a series of Observing System Simulation Experiments (OSSEs) using the Local Ensemble Transform Kalman Filter (LETKF) with the JMA-NHM (NHM-LETKF) at a 1-km resolution. The 100-m simulation is used as the nature run for the OSSEs.

First, we simulated a synthetic radar observations every 1 minute from the nature run, and performed a data assimilation experiment using the NHM-LETKF system. When the radar data are assimilated, the RMSE of water vapor at z=2km became much lower. The spatial precipitation pattern is also improved but the intensity is weaker than the nature run.

We are now investigating potential impacts of surface observations (horizontal wind, temperature, pressure and relative humidity) through OSSEs in this particular case. The results show that the surface data have a significant positive impact on water vapor at the low levels and rainfall intensity. The dense and frequent surface data may contribute to further improvements of the predictability of local heavy rainfalls.