33. Meteorological Data Processing and Forecasting – A Regional Perspective

Numerical weather prediction (NWP) is an indispensable tool for forecasting, from daily weather to the monthly outlook. Rapid progress has been made in the last 50 years with the advancement of atmospheric science, information and data-processing technology. The resolution of NWP models has been enhanced and physical parameterization schemes have become more sophisticated as computing power has increased. The accuracy of three-day forecasts at present is twice as good as 20 years ago (Simmons and Hollingsworth, 2002; Bengtsoon, 1991). The 48-hour tropical cyclone track errors have been reduced to half of that at 20 years ago (Typhoon Committee, 2002). Data-assimila-tion techniques have evolved rapidly in recent years at advanced NWP centres, Satellite radiance is directly assimilated with four-dimension variational data assimilation (Bouttier and Kelly, 2001). Radar reflectivity and Doppler wind vectors are assimilated in regional analysis as a research mode (Wilson et al., 1998). There has been a wide application of ensemble techniques, including the consensus approach, from short- to longrange forecasting (Fritsch et. al., 2000), which are further extended to tropical cyclone tracks. Concern about air quality is high, and environmental technology has advanced for the monitoring and numerical modelling of trajectory of radioactive pollutants and chemical spills. Commercialization is in progress and weather services have been transformed toward customers’ demand in developed countries (Tennekes, 1988). It is, however, important to note that quite different interests and concerns about weather services are found in different parts of the world and Region II is no exception.