Radar winds and reflectivity, combined with observations from surface stations and aircraft are assimilated in a non-hydrostatic model at high resolution for the purpose of diagnosing cloud, precipitation, icing, and turbulence in the terminal and enroute airspace in the Great Lakes Region. A real time system was run at our center for two test periods and detailed reanalyses and statistics were collected for two specific cases of significant weather in the analysis region. The assimilation of data into the model yielded improved detection of weather hazards compared to the use of common hazard-detection algorithms and gridded data from the operational Eta model. Additionally intercomparison was done among several analysis techniques, including the ADAS Bratseth method, VVP preprocessor for ADAS and 3DVAR. Results of the comparisons will be presented.