Open AccessEnsemble Typhoon Quantitative Precipitation Forecasts Model in Taiwan
Author(s) -
Jimin Hong,
Chin–Tzu Fong,
Ling-Feng Hsiao,
Yi-Chiang Yu,
Chian-You Tzeng
Publication year - 2015
Publication title -
weather and forecasting
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.393
H-Index - 106
eISSN - 1520-0434
pISSN - 0882-8156
DOI - 10.1175/waf-d-14-00037.1
Subject(s) - typhoon , quantitative precipitation forecast , precipitation , environmental science , meteorology , ensemble forecasting , climatology , forecast skill , lead time , computer science , geography , marketing , business , geology
In this study, an ensemble typhoon quantitative precipitation forecast (ETQPF) model was developed to provide typhoon rainfall forecasts for Taiwan. The ETQPF rainfall forecast is obtained by averaging the pick-out cases, which are screened using certain criterion based on given typhoon tracks from an ensemble prediction system (EPS). Therefore, the ETQPF model resembles a climatology model. However, the ETQPF model uses the quantitative precipitation forecasts (QPFs) from an EPS instead of historical rainfall observations. Two typhoon cases, Fanapi (2010) and Megi (2010), are used to evaluate the ETQPF model performance. The results show that the rainfall forecast from the ETQPF model, which is qualitatively compared and quantitatively verified, provides reasonable typhoon rainfall forecasts and is valuable for real-time operational applications. By applying the forecast track to the ETQPF model, better track forecasts lead to better ETQPF rainfall forecasts. Moreover, the ETQPF model provides the “scenario” of the typhoon QPFs according to the uncertainty of the forecast tracks. Such a scenario analysis can provide valuable information for risk assessment and decision making in disaster prevention and reduction. Deficiencies of the ETQPF model are also presented, including that the average over the pick-out case usually offsets the extremes and reduces the maximum ETQPF rainfall, the underprediction is especially noticeable for weak phase-locked rainfall systems, and the ETQPF rainfall error is related to the model bias. Therefore, reducing model bias is an important issue in further improving the ETQPF model performance.