The applicant of this proposal has recently developed a new multiple-Doppler radar wind synthesis algorithm named WISSDOM (WInd Synthesis System using DOppler Measurements, see Liou and Chang 2009; Liou et al. 2012, Liou et al. 2015). This method uses continuity condition and vertical vorticity equation as two major constraints, and is capable of using data simultaneously from any number of radars to recover the three-dimensional wind fields along/near the radar baseline as well as immediately above terrain. This is so far the only algorithm in the radar meteorology community that possesses all advantages mentioned above. It is particularly suitable to recover the wind field in an environment such as Taiwan with complex topography, and can optimally uses the data collected by Taiwan’s island-wide radar network.This research will apply the WISSDOM-synthesized wind field, the retrieved three-dimensional thermodynamic (i.e., pressure and temperature) field to initialize a high resolution numerical model, and study the performance of the model in forecasting the rainfall in a mountainous region. This proposal is the third part of a three-year project, and the goal of each year is listed in the following:1st year(2016/8-2017/7): To better use the wind field following the terrain obtained from WISSDOM, we will design a new thermodynamic retrieval scheme to recover the pressure and temperature fields above the complex topography.2nd year(2017/8-2018/7): Using the Observation System Simulation Experiment (OSSE) framework, we will apply the radar observed and retrieved three-dimensional wind, temperature, pressure, rain water, snow water, all are in a domain with complex terrain, to initialize a high resolution numerical prediction model. The purpose is to examine that under this idealized scenario, the improvement of the model’s capability in terms of predicting heavy rainfall in a mountainous region. Modification of the entire algorithm will be performed if necessary, based on the results from the OSSE tests. 3rd year(2018/8-2019/7): Based on the results from the previous two years, this method will be applied to real case studies. Currently, the selected targets can be the heavy rainfall cases from 2008 SOWMEX/TiMREX, or the intensive rainfall cases occurring in a metropolitan area such as the Taipei city.