On landfall in 2002, Typhoon Rusa caused record-breaking rainfall (870.5 mm/d) in Gangneung at the foot of the Taebaek Mountain range, Korea. In this study, the predictability of the torrential rainfall associated with the typhoon and the detailed mesoscale precipitation distribution were investigated through numerical simulation. The weather research and forecasting (WRF) model successfully simulates the mesoscale rainfall distibution and timing. With a 10 km (3.3 km) horizontal grid, the model predicted 830.8 mm (1307.9 mm) of rain with some uncertainty in the observed local maximum in coverage. This study shows that the complexity of forecasting is associated with the interactions between environmental flows, typhoon flow, and topography. In bogus and 1-day-earlier initial-time experiments, the typhoon track and intensity are substantially improved, although this has little impact on successful rainfall simulation. The interaction between the large-scale features and the proper vertical structure of the typhoon is key. The simulated rainfall at the cloud-resolving spacing of 3.3 km is greater than that at 10 km because of the increased terrain height and not the greater resolution. Analysis of observed and simulated data shows that the torrential rainfall had two different causes with two peaks in the hourly rainfall at Gangneung. The first peak links to a mesoscale frontal structure, characterized by strong moisture and thermal gradients formed by the intrusion of cold, dry northerly air from a midlatitude trough and moist southerly air from the typhoon. The second peak results from the direct effect of the typhoon and the lifting of the moisture-laden typhoon winds. Copyright 2010 by the American Geophysical Union.
CITATION STYLE
Lee, D. K., & Choi, S. J. (2010). Observation and numerical prediction of torrential rainfall over Korea caused by Typhoon Rusa (2002). Journal of Geophysical Research Atmospheres, 115(12). https://doi.org/10.1029/2009JD012581
Mendeley helps you to discover research relevant for your work.