The tropical cyclone ensemble forecast experiments are based on the GRAPES-TCM model, the Lagged Average Forecasting(LAF) and the Breeding Growth Mode(BGM) methods are combined to used for making the initial perturbations, which have been devided into three part to considered: the introduction of initial perturbation, the scaling down of environment region and the vortex region respecctively. The verification technique has been applied on the real time forecsts of 16 tropcial cyclones in 2008, including the ensemble mean, foreacast error, spread analysis and the relative skill score. The results indicate that: The average errors of ensemble mean track forecast are 231.48 km(24 h), 386.4 km(48 h) and 632.05 km(72 h). Both the error of ensemble mean and the ensemble spread are incresed with the forecast leading time. On the contrary, the ratio of these two part is decresed. Most of the samples are well good obey the correlationship between the forecast error and ensemble spread, but there is still a little sample indicate the small spread with large error situation and the large spread with small error situation. The percentage of realative skill score which ensemble forecast performed better or the same is 66.2%, especailly in the later time of model forecas. Futhermore, more attention should be paid on the situation when the ensemble system has different senarios, forecaster or user should aware of the uncertainty of the forecast and the likehood of the occurence, as well as the resluts for the event with small probability.
[1]Kalnay E. Atmospheric Modeling, Data Assimilation and Predictability[M]. Cambridge: Cambridge University Press, 2003.
[2]Chan J C L. Ensemble Forecasting of Tropical Cyclones[M]//The science and forecasting of tropical cyclones. WMO/TD-No.1129, Geneva, Switzerland, 2002: 26-33.
[3]Hamill T M. Ensemble-based Atmospheric Data Assimilation. Predictability of Weather and Climate[M].Cambridge: Cambridge Press, 2006.
[4]Aberson S D, Bender M A, Tuleya R E. Ensemble forecasting of tropical cyclone tracks[R]. Preprints. The 12\+nd Conference on Numerical Weather Prediction, Phoenix, Arizona, Amer Meteor Soc, 1998: 290-292.
[5]Zhang Z, Krishnamurti T N. Ensemble forecasting of hurricane tracks[J]. Bull Amer Meteor Soc, 1997, 78: 2785-2795.
[6]Zhang Z, Krishnamurti T N. A perturbation method for hurricane ensemble predictions[J]. Mon Wea Rev, 1999, 127: 447-469.
[7]Brian P M, Krishnamurti T N. Ensemble forecast of a typhoon flood event[J].Wea Forecasting, 2001, 16(4): 399-415.
[8]Cheung K K W, Chan J C L. Ensemble forecasting of tropical cyclone motion using a barotropical model, Part I: Perturbations of the environment[J]. Mon Wea Rev, 1999, 127: 1229-1243.
[9]Cheung K K W, Chan J C L. Ensemble forecasting of tropical cyclone motion using a barotropical model. Part II: Perturbations of the vertex[J]. Mon Wea Rev, 1999, 127: 2617-2640.
[10]Cheung K K W. Ensemble forecasting of tropical cyclone motion: comparison between regional bred modes and random perturbations[J]. Meteor Atmos Phys, 2001, 78: 23-34.
[11]Buizza R, Leutbecher M, Isaksen L, et al. Combined use of EDA and SV-based perturbations in the EPS[R]. ECMWF Newsletter, No. 123, ECMWF, Reading, United Kingdom, 2010: 22-28.
[12]Puri K, Barkmeijer J, Palmer T N. Ensemble prediction of tropical cyclones using targeted diabatic singular vectors[J]. Quart J Roy Meteor Soc, 2001,127: 709-731.
[13]Torn R D, Hakim G J. Ensemble data assimilation applied to RAINEX observations of Hurricane Katrina (2005)[J]. Mon Wea Rev, 2009, 137: 2817-2829.
[14]Zhang F, Weng Y, Sippel J A, et al. Cloud-resolving hurricane initialization and prediction through assimilation of Doppler radar observations with an ensemble Kalman filter: Humberto (2007)[J]. Mon Wea Rev, 2009, 137: 2105-2125.
[15]Torn R D. Performance of a mesoscale ensemble Kalman filter (ENKF) during the NOAA high-resolution hurricane test[J]. Mon Wea Rev, 2010, 138: 4375-4392.
[16]Wu C C, Lien G Y, Chen J H, et al. Assimilation of tropical cyclone track and structure based on the ensemble Kalman filter (EnKF)[J]. J Atmos Sci, 2010, 67: 3806-3822.
[17]Nutter P, Stensrud D J, Xue M. Effects of coarsely resolved and temporally interpolated lateral boundary conditions on the dispersion of limited-area ensemble forecasts[J]. Mon Wea Rev, 2004, 132: 2358-2377.
[18]Gentry M S, Lackmann G M. Sensitivity of simulated tropical cyclone structure and intensity to horizontal resolution[J]. Mon Wea Rev, 2010, 138: 688-704.
[19]Buizza R, Miller M J, Palmer T N. Stochastic simulation of model uncertainties in the ECMWF ensemble prediction system[J]. Quart J Roy Meteor Soc, 1999, 125: 2887-2908.
[20]Palmer T N, Buizza R, Doblas-Reyes F, et al. Stochastic parametrization and uncertainty[J]. ECMWF Tech Memo, 2009, 598: 42.
[21]Teixeira J, Reynolds C A. Stochastic nature of physical parameterizations in ensemble prediction: A stochastic convection approach[J]. Mon Wea Rev, 2008, 136: 483-496.
[22]Charron M, Spacek L, Houtekamer P L, et al. Toward random sampling of model error in the Canadian ensemble prediction system[J]. Mon Wea Rev, 2010, 138: 1877-1901.
[23]Hacker J P, Ha S Y, Snyder C, et al. The US Air Force Weather Agency's mesoscale ensemble: Scientific description and performance results[J]. Tellus, 2011, 63A(3): 625-641.
[24]Krishnamurti T N, Kishtawal C M, LaRow T, et al. Multi-model superenemble forecasts for weather and seasonal climate[J]. J Climate, 2000, 13: 4196-4216.
[25]Goerss J. Tropical cyclone track forecast using an ensemble of dynamical models[J]. Mon Wea Rev, 2000, 128: 1187-1193.
[26]Russell L E, Lester E C III.Notes and correspondence,consensus of dynamical tropical cyclone track forecast--errors versus spread[J]. Mon Wea Rev, 2000, 128: 4131-4138.
[27]Rappaport E N, Franklin J L, Avila L A, et al, Advances and challenges at the national hurricane center[J]. Wea Forecasting, 2009, 24: 395-419.
[28]张立凤, 于永锋. 增长模叠加方式及其对集合预报初始扰动的影响[J]. 高原气象, 2007, 26(1): 75-82.
[29]张立凤, 牛震宇. 基于LAF思想的一种初始扰动生成方法及集合预报试验[J]. 高原气象, 2008, 27(5):987-992.
[30]李俊, 杜钧, 王明欢, 等. 中尺度暴雨集合预报系统研发中的初值扰动试验[J]. 高原气象, 2009, 28(6): 1365-1375.
[31]高峰, 闵锦忠, 孔凡铀. 基于增长模繁殖法的风暴尺度集合预报试验[J]. 高原气象, 2010, 29(2): 429-436.
[32]肖玉华, 何光碧, 陈静, 等. 区域集合预报增长模繁殖扰动方法研究[J]. 高原气象, 2011, 30(1): 94-102.
[33]王晨稀. 对流参数化影响热带气旋的模拟研究究[J]. 高原气象, 2013, 32(2): 468-480, doi: 10.7522/j.issn.1000-0534.2012.00045.
[34]周霞琼, 端义宏, 朱永褆. 热带气旋路径的集合预报方法研究I——正压模式结果的初步分析[J]. 热带气象学报, 2003, 19(1): 1-8.
[35]Xia Qiongzhou, Zhong Liangchen. Ensemble Forecasting of Tropical Cyclone Motion Using A Baroclinic Model[J]. Adv Atmos Sci, 2006, 3: 22-34.
[36]王晨稀, 梁旭东. 热带气旋路径集合预报试验[J]. 应用气象学报, 2007, 18(5): 586-593.
[37]Tan Yan, Liang Xudong. An ensemble forecast experiment of a landing typhoon[J]. J Trop Meteor, 2012, 18(3): 314-321.
[38]黄小刚, 费建芳, 陆汉城. 基于集合Kalman滤波数据同化的热带气旋路径集合预报研究[J]. 大气科学, 2007, 31(3): 468-478.
[39]黄伟, 端义宏, 薛纪善, 等. 热带气旋路径数值模式业务试验性能分析[J]. 气象学报, 2007, 65(4): 578-587.
[40]Liu Q, Marchork T, Pan H L, et al. Improvements in hurricane initialization and forecasting at NCEP with global and regional (GFDL) models[R]. Maryland: Technical Procedures Bulletin, 2000: 472.
