Addition of Analyses and Forecasts to a Global NWP Cycle System

The KMA (Korea Meteorological Administration) operational global NWP system is based on a 6-hour analysis cycle, with 6-hour data assimilation windows. Early-observation-cutoff analyses are used to initialize the main forecasts, and late-observation-cutoff analyses are used to generate background forecasts for the next analysis update. In this system, the most up-to-date forecasts can sometimes be based on observational data that is at least 6 hours old. By producing analyses and forecasts more frequently, we can reduce this time gap, and potentially improve the quality of the most up-to-date forecast products at certain times of day. In this study, we added extra analyses to the current system, midway between the existing 00, 06, 12 and 18 UTC analyses, without changing the underlying data assimilation cycle. This method required us to overlap the extra data assimilation windows with the existing windows, so we took care to ensure that observations were not re-used in the extra analyses. To avoid delays, the background forecasts for the extra analyses come from early-observation-cutoff rather the late-observation-cutoff forecasts.
To evaluate the performance of forecasts from the new 03, 09, 15 and 21 UTC analyses, we compared the new forecasts with the existing early-cutoff forecasts from the cycles 3-hour earlier. The results showed small improvements to temperature and geopotential height forecasts, and more significant improvements to wind forecasts. As well as improving the quality of the best-available forecasts at certain times of the day, the new analyses provide more regular updates on the evolution of high-impact meteorological phenomena, such as typhoon tracks and intensity, which is of particular value to forecasters needing to make critical decisions in real-time.