Vegetation plays a significant role in determining the partition of surface sensible and latent heat flux through transpiration, interception of precipitation, dew formation, evaporation, and radiation. Therefore, vegetation must be represented adequately in numerical weather prediction models. The NCEP Noah Land Surface Operational Model uses the NESDIS 0.144 degree (about 14km) monthly Green Vegetation Fraction database, which is based upon a 5-year NDVI climatology from AVHRR data observed from the NOAA polar-orbiting satellites (1986 to 1991). In reality, vegetation fraction changes from year to year and can have a significantly different seasonal evolution from climatology, e.g., due to an advance or delay in spring 'green-up', changes in temperature and precipitation patterns, different irrigation or crop harvesting, forest fires, deforestation, desertification, draining wet land, building dams, hailstorm, flooding or drought affect vegetation distribution.

Recently, NESDIS has produced 'real-time' weekly green vegetation fraction from the remote sensor AVHRR. In this study, this new data set is tested in the Noah land-surface model coupled with the 32-km mesoscale Eta model to make many 3-day simulations covering the growing season. Resulting surface water and energy fluxes, 2-meter dew-point and air temperature, and 850-700mb thickness will be compared to those predicted by the operational Eta model, and to observations including data from various mesoscale observation networks.