With School of Electrical and Computer Engineering.
Lei Lei
28 July 2009, 12:30 PM
National Weather Center, Room 5820
120 David L. Boren Blvd.
University of Oklahoma
Norman, OK
Directions to the NWC (.pdf, 60 kb)
Simulations of radar signals are important in understanding the performance of a radar system and advancing its signal processing. Recently, a weather radar simulator has been developed to generate time series radar signals from numerical weather prediction (NWP) model output. The radar signal is a coherent sum of the echoes from all the individual scatterers while their size and motion are determined by the Advanced Region Prediction System (ARPS). Currently, the simulator generates single polarization radar signals. The operational WSR-88D (Weather Surveillance Radar 88-Doppler) radar network is being upgraded to have dual polarization capability and will be completed within the next few years. Therefore, it is worthwhile to extend the radar simulator to have polarimetric capability. In this thesis, we present the added features of generating dual-polarization radar signals with the propagation effects included. Furthermore, polarimetric radar parameters such as differential reflectivity, differential phase, and copolar cross correlation coefficient are estimated and are shown to be consistent with polarimetric radar measurements.
It has been recognized that the quality of polarimetric radar data degrades as the signal-to-noise ratio (SNR) decreases. This substantially limits the usage of observed polarimetric data to relatively low SNR regions. To improve radar performance and data quality, a multi-lag correlation estimator is introduced and examined in terms of moment estimation accuracy, where lag 0 is excluded to minimize noise effects. In this thesis, we present the theory, simulation results, and experimental validation of the multi-lag estimation technique for radar reflectivity, differential reflectivity, correlation coefficient, and spectrum width. The biases and standard deviations are calculated and compared with those of the conventional method. Theoretical analysis of the estimation errors are provided and verified by numerical simulations. It is found that the multi-lag estimator outperforms the conventional method at low SNR for certain range of spectrum width values.
Back to NWC Seminar Series and Colloquia