Quantitative Assessment of the Impact of Track Forecast Error on Tropical Cyclone Quantitative Precipitation Forecasts

Heavy rain from tropical cyclones(TC) often claims may lives and causes severe damages. Accurate and timely TC quantitative precipitation forecasts(QPFs) are needed to mitigate such adverse impacts. While significant improvements have been made in TC track forecasts, the skill of TC QPFs is still very limited. One of the most important reasons for slow improvement in TC QPFs by numerical weather prediction(NWP) models is the lack of quantitative research on the main shortcomings and possible sources of errors of model TC QPFs. This paper quantitatively assesses the impact of tropical cyclone track forecast error on TC QPFs and proposes a post-processing method to reduce the impact of model track forecast error on TC QPFs.

The study focuses on a total of 29 TCs over the coastal waters of China in 2015-16, including 12 China landfalling tropical cyclones. GFS forecasts, best track dataset and recently released satellite-based rainfall estimation GPM_3IMERGHH final run dataset are used for validation. Track errors are removed by shifting each 3-h GFS rainfall forecast pattern by a vector pointing from the model forecast storm center at that forecast time to the observed location at the corresponding valid time. Verification results of the shifted forecasts and the untreated forecasts are compared to assess the impact of track error on TC QPFs. Differences between impact on landfalling TC QPFs and on non-landfalling TC QPFs are also evaluated. Results show that the sharply decline of forecast performance for longer lead time is mainly due to the large track errors especially for heavier rain. Track forecast error has less impact on landfalling TC QPFs than on non-landfalling forecasts especially for rainfall over 50mm. Other error sources like land-atmosphere interaction or topographic effect may play a role.

A post-processing method is proposed to reduce the impact of model track forecast error on TC QPFs. The official track forecasts are considered the best track forecast that can obtain when making rainfall forecast. GFS rainfall forecast patterns are shifted according to the latest official track forecast. The results show that the corrected QPFs get fairly good improvements.