We describe development of a legacy precipitation data set for the GEWEX Continental-Scale International Project (GCIP) region in the United States. The data set covers the period 1995-2000 and consists of hourly accumulations with a spatial resolution of 4×4 km2. We are using the national, 15-minute reflectivity composite from the NEXRAD network of WSR-88D radars, daily rain gauge data, and hourly rain gauge data as the principal inputs for the legacy precipitation data set. We have developed GIS-based tools for data storage and browsing. We developed a highly efficient data compression scheme, and we have incorporated it with the other analysis tools.

The final precipitation product is derived from a multi-step precipitation algorithm. First, we threshold the input radar-reflectivity to eliminate areas of unrealistic (as rainfall) reflectivities. Next, we apply a probability matching approach to reduce or eliminate artifacts of the composite images for the areas of overlap of individual radar coverage. This approach is performed for climate zones and on a seasonal temporal scale. The next step in the algorithm is optimization of parameters for the reflectivity (Z) vs. rainfall (R) relationship. We applied the power law Z-R relation (Z=ARb), and we use high quality and high-resolution rain gauge data (e.g. Oklahoma Mesonet) for this optimization. The last step we apply is to account for advection errors when calculating precipitation accumulation.

We also present product evaluation results over several sites across the country using additional rain gauge data sets. The final data set includes the hourly rainfall maps for the entire Mississippi River Basin as well as the input reflectivity data. The data set will be available to modelers of the hydrologic cycle by July 1, 2002. The product will be distributed to the research community through the UCAR/Joint Office for Science Support (JOSS).