Assimilation of NEXRAD-VAD winds in summertime meteorological simulations over the northeastern United States

Next-Generation Radar (NEXRAD) velocity azimuth display (VAD) winds were available at 10 sites in the northeastern United States during intensive observing periods of the North American Research Strategy on Tropospheric Ozone-Northeast field study conducted in the summer of 1995. These VAD winds represent a potentially valuable routine source of upper-air data suitable for mesoscale four-dimensional data assimilation and other mesometeorological applications. The objectives of this paper are to develop appropriate quality-checking methods for these data during a period with weak dynamic forcing and to learn if their assimilation into a nonhydrostatic mesoscale model, the Fifth-Generation Pennsylvania State University-National Center for Atmospheric Research Mesoscale Model (MM5), can reduce wind errors in lengthy numerical integrations. Two types of quality checking were applied: 1) a standard internal vertical and temporal consistency check, and 2) a new filter that uses bias-corrected model predictions as a first guess. After unreliable data were removed, the VAD winds were assimilated into MM5. Experiment evaluation using independent data demonstrated that the VAD winds significantly reduced model wind errors, especially below 2.0 km, where the wind data are most numerous in this case. Independent verification also indicated that the filter presented in this paper contributed to the improvement of the data-assimilated model results. Although the application in this case is designed to generate wind fields to drive an air qualify model, the techniques developed, with some generalization and testing, also should be adaptable for forecast-initialization applications.