An improved QPE algorithm for the South African Weather Service's S-Band Doppler Radar at Irene

Weather radar systems can be used to estimate rainfall in areas where no gauges are available. Although its measurements are indirect, it remains the best alternative to estimate rainfall on high temporal and spatial resolution. With the implementation of the South African Flash Flood Guidance system (SAFFG) and the installation of the new S-band Doppler radar systems, much emphasis has been placed on the performance from radar based Quantitative Precipitation Estimates (QPE). A methodology used by the Bureau of Meteorology (BoM) was selected to upgrade and improve the existing radar precipitation field. As a starting point, a rainfall classification algorithm has been introduced which separates convective from stratiform precipitation with the intent to assign dual Z-R relations to the two different rainfall types. A technique for smoothing the accumulation of precipitation, which is based upon Optical Flow Constraints (OFC), has also been introduced into the QPE algorithm. The newly calculated precipitation field will be compared against hourly and daily ARS and AWS data, as well as the existing Thunderstorm Identification Tracking and Nowcasting (TITAN) precipitation algorithm. The TITAN system is the current operational system used by SAWS forecasters and the SAFFG model. Results show that there is a decrease in the overall error as well as a decrease in the over-estimation of precipitation for the S-band radar at Irene when using dual Z-R relationships. The results are further improved when the rainfall field is morphed with the assistance of OFC vectors. Results for a selected period during the summer rainfall season of 2010/11 will be presented.