Weather radar network is the only data source which can routinely provide storm-scale observations of the atmosphere at a very high time frequency. Utilizing this kind of precious radar data is vital for successful convective weather predictions. Currently, radar reflectivity data plays an import role in HRRR as it evidently improves forecasts. Its impact has been being examined. Radar radial winds are also used into HRRR analysis. To get best use of radial wind data, lots of fine tuning work are underway. For example, the super-obing of radial winds needs to be compatible with the 3km grid mesh and the shock to the model due to the assimilation of a great amount of radial winds data should be carefully addressed. The background error covariance matrix (BE) should also be correctly specified since at the 3km resolution, the BE is very different from that at a lower-resolution. Tests, evaluations as well as the enhancing development work on the improvement of radial winds data assimilation into HRRR will be reported at the meeting.