Development of high-resolution archival precipitation data sets.

摘要

The need for high-resolution precipitation data at large scales is growing as earth-atmosphere research at regional, continental, and global scales is increasing. In the United States, there are networks dedicated to the measurement of precipitation, but the incorporation of these data into a single product is still under development. The network of Weather Surveillance Radar-1988 Doppler (WSR-88D) provides information about precipitation at high spatial and temporal resolutions. In addition, some dense rain gauge networks in the U.S. provide measurements of precipitation that can add information for developing data sets that cover geographically vast areas in the U.S. We have incorporated radar reflectivity data that is available over the continental U.S. for use in developing a large scale long term precipitation data set. We present the algorithm used for the development of the precipitation data set. The algorithm includes quality control of input data sources, enhancement of these data using statistical techniques, and production of precipitation accumulations at hourly 4 x 4km2 resolutions. We then present the validation of the precipitation data set by evaluating the uncertainty associated with the estimates. We use data from rain gauge networks, gridded rain gauge products, research quality NEXRAD products, and operational radar products from river forecast centers for the validation of the precipitation data for the Mississippi River Basin.; Comparisons of the radar based precipitation product with rain gauge estimates show good agreement on a basin wide scale. Although there is good agreement between these products, there still exist areas in the basin that suffer from uncertainty in the radar product. We attempt to identify these areas by locating regions that are affected by beam blockage and that are affected by anomalous propagation or other radar artifacts. We further attempt to identify regions in the basin that may suffer from uncertainty due to range dependent errors. We implement a conceptual model to aid in identifying these areas. Finally, we address the issue of sampling differences when comparing radar estimates with rain gauge estimates. We provide a framework for determining the appropriate temporal and spatial scales for comparison of the two different sensors.