Sudden Track Changes of Typhoon Megi (2010) and its Effects on Precipitation over Taiwan - Evaluation of Uncertainty Based on Ensemble Simulations

Abstract Typhoon Megi was the strongest tropical cyclone in the western North Pacific in 2010. While Megi was still distant from Taiwan, its track took a sudden change from westward to poleward over the South China Sea, west to the Luzon Islands, leading to heavy rainfall in the northeastern part of Taiwan during 0000 UTC 19-23 October. The highest amount of the 96-h accumulated rainfall was observed in Yilan, peaking at 1,234 mm. Such heavy rainfall is likely caused by the interaction between Megi's outer circulation and the northeasterly monsoon.

In this study, the uncertainty of 1) the sudden recurvature of Megi; and 2) precipitation over Taiwan under the joint influence of the outer circulation of Megi and the northeasterly monsoon are examined. With valuable aircraft observations collected during the Impact of Typhoons on the Ocean in the Pacific (ITOP) program, ensemble simulations are carried out based on the ensemble Kalman filter data assimilation system and the Weather Research and Forecasting (WRF) Model with 6-km horizontal resolution. Potential vorticity diagnosis is applied to evaluate the relative contribution to the steering flow from individual dynamical systems, such as the Pacific subtropical high, the deep midlatitude trough, and the front system to the northeast of Megi.

For typhoons making landfall in Taiwan, it is well known that the track is an important factor in affecting the rainfall pattern over Taiwan. In contrast to physical intuition and previous understanding, for the remote precipitation case of Megi, the result of the equitable threat score (ETS) shows no clear relationship between track errors and performance of the simulated rainfall over Taiwan. To understand why some members with large (small) track errors could (could not) well simulated the precipitation, the simulated storm size and structure, synoptic-scale flow, and the multi-scale interaction are investigated.

Quantitative precipitation forecast of the typhoon-induced rainfall, especially over the terrain of Taiwan, has always been a challenging task. It is expected that ensemble simulations in this study could provide new insights into the uncertainty and predictability in forecasting tracks and the typhoon-induced remote impact on rainfall.