Characterization of vegetation cover is a key input for several regional air quality modeling processes. One example is the calculation of heat, moisture, and momentum fluxes with the Mesoscale Meteorological Model (MM5). Another example is the estimate of biogenic volatile organic compound and nitric oxide emissions with the Biogenic Emissions Inventory System (BEIS3). This paper compares vegetation cover estimated from three contemporary databases: (1) the North American Land Cover Characteristics (NALCC) database, (2) the National Land Cover Database (NLCD), and (3) the Biogenic Emissions Landuse Database (BELD3). The NALCC database, which is released by USGS and supported by NCAR for use with the MM5 model, consists of 1-km resolved land cover classes derived from the AVHRR satellite. The NLCD was released recently by a multi-federal agency group called the Multi-Resolution Land Cover Consortium (MRLCC). NLCD is based on LANDSAT-TM data and is available at a 30-m resolution. BELD3 is a hybrid vegetation cover database developed by EPA to provide inputs to the Biogenic Emissions Inventory System. BELD3 attempts to resolve tree and crop cover (by species) to a 1-km resolution. This paper will examine vegetation cover distributions from these three databases for a 984,064 km2 MM5 modeling domain centered over Nashville, Tennessee. Comparisons indicate significant differences between the three databases. For a 3 degree x 3 degree area, agricultural area estimates range from 29 to 47%, while forested area estimates range from 46 to 68%. The commonly-used NALCC database indicates significantly more forest area and less crop area than the other two databases. Overall, forest and crop cover percentages agreed reasonably well between NLCD and BELD3. These differences and their impact on biogenic emission estimates will be considered in this paper.