Short-Term Wind Prediction over Dallas Fort Worth (DFW) Urban Network

The main objective of this paper is to develop short term wind forecast system over the DFW region. In order to do that we utilize the HRRR model forecast as the initial and boundary condition for the WRF model and build the wind forecast from that, at 1 Km spatial and 5 min temporal resolution. The model domain covers the DFW area (300 x 300 km) and configured with 301×301 horizontal grid points with 1 km grid spacing. The Collaborative and Adaptive Sensing of the Atmosphere (CASA) radar network data and Next Generation Weather Radar (NEXRAD) data (both radial velocity and reflectivity) at DFW are assimilated through 3DVAR and 4DVAR techniques in WRF model. The next objective here is to investigate the role and impact of assimilating the high resolution radar observations.

The rapid 3DVAR cycling is used to assimilate radar data every 5 minutes whereas 15 minutes time window is used in 4DVAR. One month WRF model forecasts (60 ensembles) are used in National Meteorological Center (NMC) method to generate a domain based background Error (BE) covariance matrix. Subsequently experiments were conducted with and without assimilating the high resolution observations. It was used to assess the impact of high resolution observations, especially over short lead time scales. The radar retrieved horizontal, wind speed and direction (vector winds) is considered as truth, and the Model Evaluation Tool (MET) is used to examine the simulations. The detailed comparison between 3DVAR, and 4DVAR simulations are presented.