The processing of VAD winds began to be assimilated at the NCEP by RUC in July 1995 and by the Eta model in June 1997. However, when the many problems with the data became evident, all operational use of the VAD winds ended in January 1999. Since then, there has been extensive diagnosis at NCEP and elsewhere to determine the source of the problems. As a result of those investigations, a quality control technique was developed that specifically identifies errors with different characteristics, and marks them for non-use. Among the marked winds, the majority of the total were classified as errors due to migrating bird contamination. This classification, however, was based on the differences between the VAD and background winds at preferred altitudes and temperatures during bird migration seasons, without using NEXRAD level-II data nor seeking for ground truth information regarding the presence versus absence of birds. Because of this, discrimination of bird echoes from meteorological scatters remains an unresolved and important problem for both VAD winds and level-II radial winds quality control and assimilation. To this end, polarimetric radar data have been collected during the spring of 2001 to obtain reliable information regarding to the presence versus absence of birds. NCEP model background winds and VAD winds have been collected and compared with radiosonde observations and Level II wind data to identify error characteristics in VAD winds and Level-II radial winds in the presence or absence of birds. Detailed results will be presented at the conference.