Importance of Environmental Variability to Storm-scale Radar Data Assimilation
Jidong Gao, CAPS/Univ. of Oklahoma, Norman, OK; and D. J. Stensrud and M. Xue
The assimilation of WSR-88D radar data into storm-scale weather prediction models for short-range forecasting (warn-on-forecast) represents a significant scientific and technological challenge. Numerical experiments over the past few years indicate that the quality of the resulting storm-scale analyses and forecasts are sensitive to the storm environment, parameterized microphysics, the details of the assimilation methodology, the quality of the radar data, and the method used to generate the initial ensemble members. In this study, we investigate the importance of environmental variability to storm-scale radar data assimilation by including explicit environmental information from a 30-member mesoscale ensemble data assimilation system (WRF-DART).
A three-dimensional variational data assimilation system (3DVAR) developed at the Center for Analysis and Prediction of Storms (CAPS) is used to assimilate observations from the WSR-88D radar network into a storm scale nonhydrostatic NWP model—the Advanced Regional Prediction System (ARPS) for the 4-5 May 2007 Greensburg, KS, tornadic thunderstorm case. Both reflectivity and radial velocity data are assimilated. A tiered approach of increasing complexity is used to examine the importance of mesoscale environmental variability on storm-scale radar data assimilation. For the first experiment, a horizontally homogeneous environment provided by the mesoscale WRF-DART mean analysis sounding at the grid point closest to the observed initiation of the supercell thunderstorm will be used to provide the storm environment. For the second experiment, a 30-member ARPS ensemble will be created that starts from the ARPS 3DVAR analyses using each of the vertical soundings from the 30-member WRF-DART analyses as background and 88D radar data as observations. It is expected that these analyses will provide a reasonable estimate of the environmental variability for the stormscale forecasts. Finally, for the third experiment, 30 three-dimensional grids will be extracted from the mesoscale WRF-DART analyses centered over the location of the observed Greensburg tornadic event and provided to the ARPS 3DVAR to create an ensemble with heterogeneous environments. By using this tiered approach, we can assess the relative importance of vertical and horizontal environmental variability to storm-scale radar data assimilation. The purpose of these experiments is to quantify the value of an ensemble of 3DVAR analyses, and the assimilation of 88D radar network to the accuracy of stomscale forecast, in both deterministic and probabilistic frameworks. Some preliminary experiments have been performed. Several new numerical experiments will be conducted and their performance will be reported at the conference.
Extended Abstract (1.1M)
Session 9B, Data Assimilation
Tuesday, 28 October 2008, 4:30 PM-6:00 PM, South Ballroom
Previous paper Next paper
Browse or search entire meeting
AMS Home Page