9419103 Krajewski The main goal of the proposed work is to develop a radar database for use by the hydrological and meteorological community. The data are from digital RADAP-II radars operated by the National Weather Service and which are being phased out and replaced by the new NEXRAD technology, complemented by the raingage networks. The database will consist of high quality radar-based rainfall estimates based on several WSR-57 and WSR-74S radars equipped with RADAP-II digital processors, operated by the National Weather Service (NWS). The data base will include rainfall estimates at temporal scale that range from instantaneous to daily accumulations, and a spatial scales of 4 km by 4 km, 15 km by 15 km, and 50 by 50 km. In order to obtain rainfall estimates from radar data, several processing steps are required. First, removal of anomalous propagation (AP) echoes will be done with a modified Gaussian classification approach. Range correction is required to remove systematic errors due to range effects. Radar reflectivity to rainfall rate (Z-R) conversion based on raindrop size distributions is applied to achieve physically-based data, and polar to rectangular coordinate conversion is carried out followed by removal of advection effects. Accumulation calculations and bias removal complete the estimation procedure. The final step is to perform an error propagation analysis of the preceding steps. The work planned under the proposal includes three main elements: (1) development of a quality-controlled radar reflectivity database; (2) providing estimates of rainfall based on the radar data; and (3) developing procedures for uncertainly propagation and providing estimates of error variance for the above rainfall products. The quality of the proposed products would be greatly enhanced if a combined radar-raingage analysis was conducted. This is due to significant gaps in the temporal coverage by the radar data caused by various technical difficulties as well as the potential rad ar calibration problems. On the other hand, the radar data will be helpful in clarifying certain quality control problems in the raingage data, and in estimating the error structure of the raingage-based estimates, particularly the error component due to a spare spatial sampling by the raingage network. The main product of our research will be grided analysis of rainfall based on combined radar and raingage observations, complemented by a quantitative assessment of its uncertainty. This analysis will be performed for the Arkansas River Basin with a particular focus on the ARM/CART experimental site in Oklahoma and Kansas. Essential elements of the project will be development of conceptual design for data analysis, simulation studies of error propagation, and independent cross-validation of our products.
