Keynote: An Introduction to the QPESUMS Rerun Products for Rainfall Prediction of Landfall Typhoons in the Taiwan Area

ABSTRACT

This paper is to introduce the forecast products for landfall typhoons provided by the rerun system of QPESUMS (Quantitative Precipitation Estimation and Segregation Using Multiple Sensors) currently operated by CWB (Central Weather Bureau) of Taiwan. QPESUMS is a real-time operating system using multiple sensors, primarily including surface rain gauges and Doppler radars, to estimate area-mean rainfall and segregate the type of precipitation. The rerun system is to include the missing data due to communication malfunction in real time and to re-composite the radar reflectivity maps and to generate the revised rainfall products.

Quantitative precipitation forecast of landfall typhoons, especially the location and amount of heavy rain in a very short period of time, is urgent needed by the hydrological application. The rerun data set is used to generate useful historical rainfall products and is an important information source for QPF of landfall typhoons. The major products of the rerun data set include the azimuth-mean and the quadrant- mean rainfall during the period of landfall. The major characteristics of the azimuth- mean precipitation structure changes at landfall are depicted very well by these radius-time sections. This product can be used to identify evolution of the precipitation strength, i.e., the most affected range from the center of the storm and the possible change of the intensity. It has been found in the study of Typhoon Nari (2001), that significant precipitation enhancement can be traced back to 4-5 hours before the storm made landfall and the final hour enhancement (about 50% increase) before landfall is identified.

Quadrant-mean precipitation structure, calculated according to the direction of movement and the vertical wind shear, is also analyzed in a systematical way. The possible effect of the landmass and complex terrain on the transformation of the precipitation structure can be investigated and is helpful for heavy rain location identification. These products are also combined with the products from the CLIPER type rainfall prediction model to improve the QPF accuracy of landfall typhoons. The usage of these products will be demonstrated during the presentation.