Efforts have been underway since MM5v3 was released in July 1999 to set up air quality simulations using Eta analyses as background fields. Previous simulations used a one-way quadruple-nested set of domains with horizontal grid spacing of 108, 36, 12, and 4 km. With Eta analyses, it is anticipated that the 108-km domain, and possibly the 36-km domain, can be eliminated due to the better horizontal and vertical resolution of Eta compared to the 2.5 deg NCEP and ECMWF global analyses. In addition, it is anticipated that RAWINS can be eliminated from the processing pipeline since the three-dimensional variational analysis used in the Eta Data Assimilation System (EDAS) is widely considered to be superior to the Barnes technique used in RAWINS. That is, there may be little benefit to running RAWINS when the EDAS theoretically includes all of the observations that RAWINS could incorporate, and EDAS also ingests satellite information. Since the operational and archived Eta domains and our 36-km simulation domain have considerable resolutions, it is hypothesized that the EDAS analyses would be appropriate to initialize our horizontal scale of interest.