The improvement of Quantitative Precipitation Forecasts (QPF) in the operational forecasting environment is an endeavor that has recently been given increased priority. The importance of improved QPF's on short time scales (0-24hrs.) reach many sectors of society including public safety, water resource management, and property damage mitigation among others. The utilization of output from numerical weather prediction (NWP) products is a primary tool used by forecasters in determining QPF's. However, the ability of NWP models to predict precipitation can be compromised by incorrect initial moisture and wind structures.
One of the major componets in the observation network for determining the wind and moisture distribution is the RAOB network. The disadvantages to this network are the temporal and spatial aspects of the observations are on the order of 12 hours at a horizontal distance of approximately 400km. Data available from the GOES 8/9 include the determination of clouds/no clouds, cloud top pressures, 3-layer precipitable water, and winds. This allows for a derivation of a number of parameters from the atmosphere at a horizonal spatial/temporal scale of as little as 10km at 30 minute intervals. In addition, measurements can be derived from regions where in situ data is very sparse, such as areas over the oceans. Here, the data is assimilated into the University of Wisconsin Non-Hydrostatic Modeling System (UW-NMS) via the ETA's 3-Dimensional Variational Analysis System (3D-VAR). The data is assimilated at 180 minute intervals, beginning 12 hours before the commencement of a full production run. Results from these experiments will be presented from both a quantitative and qualitative standpoint for both individual case studies and a long term time average