Premium
Selective ensemble‐mean technique for tropical cyclone track forecast by using ensemble prediction systems
Author(s) -
Qi Liangbo,
Yu Hui,
Chen Peiyan
Publication year - 2014
Publication title -
quarterly journal of the royal meteorological society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.744
H-Index - 143
eISSN - 1477-870X
pISSN - 0035-9009
DOI - 10.1002/qj.2196
Subject(s) - tropical cyclone , meteorology , environmental science , ensemble average , track (disk drive) , ensemble forecasting , tropical cyclone forecast model , climatology , quantitative precipitation forecast , numerical weather prediction , forecast error , computer science , precipitation , mathematics , geography , econometrics , geology , operating system
Abstract This article proposes a selective ensemble‐mean technique for tropical cyclone (TC) track forecast based on the errors of ensemble prediction system (EPS) members at short lead times (SLTs, 12 h in this study). The means (SEAV) and weighted means (SEWE) of selected EPS members are applied to EPS products from the European Centre for Medium‐range Weather Forecasts (ECMWF), Japan Meteorological Agency (JMA), National Centers for Environmental Prediction (NCEP), and China Meteorological Administration for 35 TCs in the western North Pacific in 2010 and 2011. Verification results show that SEAV behaves better than SEWE, with a skill of 5% to 30% over relevant ensemble means of EPS within 72 h. The SEAV method is the most effective for the JMA EPS, with a skill of 10% even at 96 h. SEAV predictions are compared with the high‐resolution deterministic model predictions of ECMWF and several official forecasts, with special consideration given to the time delay associated with numerical model products in operation. The SEAV for the ECMWF EPS can overcome the high‐resolution ECMWF deterministic model at 24 h. Case analyses and sensitivity tests on the error thresholds of member selection and SLTs are also presented in this article.