This research reports on several aspects of real-time monitoring of the spatial and temporal distribution of rainfall from ground-based weather radar. Optimization of the performance of the National Weather Service's Precipitation Processing Subsystem (PPS) is the first objective. This is achieved by developing a calibration procedure which simultaneously estimates the optimal parameter values by providing a global assessment of the system's performance. Evaluation of the system is based on a data set consisting of two months of radar reflectivity measurements, and hourly raingage rainfall accumulations, from the Melbourne, Florida WSR-88D site. Radar-raingage root mean square (RMS) difference reduction up to 20% with respect to the default system parameter values is demonstrated.; Investigation of statistical procedures for real-time adjustment of the mean-field systematic radar rainfall error is the second objective. For this purpose, a data-based Monte Carlo simulation experiment is performed. The study uses an extensive data set of hourly radar rainfall products and raingage accumulations from the Tulsa, Oklahoma WSR-88D site. This intercomparison study concluded to a bias procedure which overall appeared to perform better than the other. The main results from this research are: (1) statistical methods with optimal error model parameters perform significantly better than using only bias observations, and (2) bias adjustment is mostly effective in cold season precipitation measurements.; Final objective of this research is development of a new real-time radar rainfall estimation algorithm. The new processing steps introduced in this algorithm are beam-height effect correction, vertical integration, rain classification, and continuous range effect correction. Additionally, the algorithm applies advection correction at the gridded rainfall rates to minimize the temporal sampling effect, and its calibration is cast in a recursive formulation with parameters adjusted in real-time. A new statistical method has been developed for quantification of radar rainfall products uncertainty. Evaluation of the system and the uncertainty quantification method is based on the data set from Melbourne, Florida WSR-88D site. Radar-raingage RMS difference reduction up to 50% with respect to the default PPS is demonstrated for the proposed algorithm.
展开▼
