| Since the advent of radar in the 1940s, it has been well known that water drops composing precipitation scatter microwaves in a predictable manner. This characteristic of early radar has lead to the present day Weather Surveillance Radar (WSR-88D) or NEXRAD systems, operated by the National Weather Service (NWS). In parallel to the evolution of weather radar for measuring precipitation over large areas, remote networks of rain gauges have been deployed and managed by agencies such as the Florida Water Management Districts. Since the recent deployment of the NWS network of WSR-88D, as well as the recent launch of the NASA Tropical Rainfall Measurement Mission (TRMM) satellite, significant attention has been placed upon the merging of these diverse sources of rainfall measurement. The main focus of this dissertation research has been to develop and analyze methods of rain gauge and radar correlation for the purpose of optimizing rainfall estimates. The techniques presented in this dissertation observe that the physical link between rain gauge and radar reflectivity data is the drop size distribution (DSD). Using various numerical algorithms, as well as methods common to image and signal processing such as median filtering, two-dimensional cross-con-elation, and adaptive signal processing, methods of analysis are presented which attempt to correlate radar reflectivity, rain gauge, and disdrometer data. Particular attention is given to the subjects of rain gauge and radar interpolation; disdrometer calibration; microscale radar rainfall estimation; and a convolution model of DSD evolution, which attempts to model the convective-like properties of rainfall. |
