Observing System Simulation Experiment: Development of the system and preliminary results

An Observing System Simulation Experiment (OSSE) was developed and conducted to assess the potential impact of different observing strategies of field experiments on analysis and short-term forecasts. A Mei-Yu front rainfall case study, occurring in 2003 in southeastern Asia, was utilized to demonstrate the OSSE application to field experiments. Data sampling strategies from the Taiwan Island Monsoon Rainfall Experiment (TIMREX), with some modifications, were used. The nature run was produced by the Weather Research and Forecasting (WRF) model with a resolution of 1 km. The observational operators, which were developed to simulate observations, included radar, radiosonde, dropsonde, wind profiler, and surface stations. The verification of simulated observations, such as radar echo and radial velocity, and preliminary results from data assimilation experiments, demonstrated that the developed OSSE system performed reasonably well. The use of more observations, such as radar data, dropsondes, and extra radiosondes, was able to significantly improve analysis of winds and, to a lesser extent, moisture, and short-term rainfall forecasts. While more observations helped improve simulations, the use of higher-frequency observations (e.g., 3 h radiosondes) launched at the same locations for this case study did not substantially influence results. Thus, the information may be potentially saturated and more studies to tackle this problem are required.