The aim of this study is to examine the impact of using five different microphysics parameterization schemes, including single-, double-, and triple-moment microphysics, in an efficient high-resolution data assimilation system suitable for nowcasting and short-term forecasting with low latencies. In addition to testing the sensitivity to microphysics, the impact of gap-filling radars and variations in analysis cycling and incremental analysis updating (IAU) techniques are explored using a variety of verification methods. On 24 May 2011, Oklahoma experienced an outbreak of tornadoes, including one rated EF-5 and two rated EF-4. The extensive observation network in this area, including the WSR-88D radars, Collaborative Adaptive Sensing of the Atmosphere (CASA) IP-1 X-band radars, Oklahoma Mesonet, and standard surface data, makes this an ideal case for these tests. Additionally, the real-time configuration of the 1-km ARPS, which used 3DVAR with cloud analysis via IAU, had success providing a good baseline forecast. ARPS forecasts of 0-2h are verified using point-to-point, neighborhood, and object-based verification techniques. The object-based verification technique uses updraft helicity fields to represent mesocyclone centers, which are verified against tornado locations from three supercells of interest. Varying levels of success in the forecasts are found and appear to be dependent on the complexity of storm interaction, with early forecasts of isolated storms exhibiting the most success. Verification scores indicate the multi-moment schemes tend to produce better forecasts, assimilating CASA radar data can improve forecasts for storms within the CASA radar network, and analysis cycling and modified IAU techniques generally contribute to better forecasts.