春季中国日最高气温延伸期预报误差分析及订正

熊敏诠; 代刊; 唐健

doi:10.16032/j.issn.1004-4965.2020.071

春季中国日最高气温延伸期预报误差分析及订正

doi: 10.16032/j.issn.1004-4965.2020.071

国家气象中心，北京 100081

基金项目:

国家气象中心预报员专项 Y201928

详细信息

通讯作者:
熊敏诠，男，江西省人，博士，主要从事月内尺度气象要素预测。E-mail: minquanxiong@sina.com

中图分类号: P457.3
计量
- 文章访问数: 258
- HTML全文浏览量: 21
- PDF下载量: 21
- 被引次数: 0
出版历程
- 收稿日期: 2020-05-24
- 修回日期: 2020-09-16
- 刊出日期: 2020-12-01

ANALYZING AND CALIBRATING EXTENDED-RANGE FORECAST OF CHINA'S DAILY MAXIMUM TEMPERATURE IN SPRING

National Meteorological Center, Beijing 100081, China

摘要

摘要: 数值模式直接输出和经模式后处理得到的预报误差比较，是延伸期逐日要素预报应用基础。针对中国2 583个站点在2020年春季11~30天的日最高温度预报，根据欧洲数值中心的集合预报输出，首先，使用BP-SM（Back-Propagation - Self memory）法和回归法，进行确定性预报订正效果比较；结果表明BP-SM法和回归法都明显降低了预报绝对误差；在11~14天预报中，BP-SM法得到的平均绝对误差为3.3~3.6 ℃，预报准确率超过35%，订正效果更优。其次，基于模式直接输出和BP-SM法获得的概率预报，使用CRPSS (continuous ranked probability skill score)进行了可预报性分析。结果表明，在地形复杂地区，经过订正，预报准确率明显改善。对于延伸期逐日要素预报，合理的模式后处理方法是降低预报误差和提高预报能力的重要环节。
- 模式后处理过程 /
- 延伸期 /
- 集合预报 /
- 神经网络
Abstract: The comparison between direct model output (DMO) and post-processing product is the key to improving extended-range forecast. Based on the ECMWF ensemble forecasts and the daily maximum temperature forecasts by 2 583 stations in China for 11~30 days in spring 2020, the present study uses back-propagation-self memory (BP-SM) method and regression method to compare the effect of deterministic forecast correction. The result shows that the absolute forecast error through the post-processing method is less than that of the DMO. In the 11~14 d forecast, BP-SM method (absolute error: 3.3~3.6 ℃; accuracy: 35%) performs better than linear regression. Then, using the continuous ranked probability skill score, predictability is discussed. The result shows that in areas with complex terrain, the predictability is improved significantly through the BP-SM method. For the 11~30 d daily maximum temperature forecast, the post-processing method plays an important role in reducing forecast errors and improving forecasting capabilities.
- post-processing /
- extended-range /
- ensemble prediction /
- neural network

HTML全文

图 1 中国区域站点分布

下载: 全尺寸图片幻灯片

图 2 BP-SM方法的示意图

下载: 全尺寸图片幻灯片

图 3 集合平均、BP-SM法、回归法在2020年春季延伸期日最高温度预报绝对误差（a）和预报准确率（b）

下载: 全尺寸图片幻灯片

图 4 2020年春季日最高气温确定预报绝对误差（a~f）和准确率（g~l）空间分布图

a、c、e、g、i、k. DMO；b、d、f、h、j、l.BP-SM法。a、b、g、h.第11天；c、d、i、j.第20天；e、f、k、l.第30天。

下载: 全尺寸图片幻灯片

图 5 2020年春季的日最高温度概率预报CRPSS空间分布图

a、c、e. DMO；b、d、f. BP-SM法。a、b.第11天；c、d.第20天；e、f.第30天。

下载: 全尺寸图片幻灯片

图 6 2020年春季分区域的日最高温度11~30 d概率预报CRPSS趋势图

a. DMO；b.BP-SM法；c.55区；d.51区。

下载: 全尺寸图片幻灯片

表 1 ECMWF集合预报产品概况

产品名称	预报要素	成员数	预报时效	时间间隔	空间分辨率	发布频次	起报时间
中长期预报	定时温度	51	1~15天	12小时	0.5º×0.5º	逐日	20时
月预报	6小时最高温度	51	16~46天	6小时	0.5º×0.5º	一周两次（星期一、星期三）	08时

下载: 导出CSV

参考文献(39)

[1]	LORENZ E N. Deterministic non-periodic flow[J]. J Atmos Sci, 1963, 20: 130-141.
[2]	LORENZ E N. The predictability of a flow which possesses many scales of Motion[J]. Tellus, 1969, 21: 289-307.
[3]	VITART F, ROBERTSON A W, ANDERSON D L T. Subseasonal to seasonal prediction project: bridging the gap between weather and climate[J]. WMO Bull, 2012, 61(2): 23-28.
[4]	GLAHN H R, LOWRY D A. The use of model output statistics (MOS) in objective weather forecasting[J]. J Appl Meteorol, 1972, 11 (8): 1 203-1 211.
[5]	GLAHN B, PEROUTKA M, WIEDENFELD J, et al. MOS uncertainty estimates in an ensemble framework[J]. Mon Wea Rev, 2009, 137 (1): 246-268.
[6]	GNEITING T, RAFTERY A E, WESTVELD Ⅲ A H, et al. Calibrated probabilistic forecasting using ensemble model output statistics and minimum crps estimation[J]. Mon Wea Rev, 2005, 133 (5): 1 098-1 118.
[7]	CLARK M, GANGOPADHYAY S, HAY L, et al. The schaake shuffle: a method for reconstructing space-time variability in forecasted precipitation and temperature fields[J]. J Hydrometeorol, 2004, 5 (1): 243-262.
[8]	SCHEFZIK R. Ensemble calibration with preserved correlations: unifying and comparing ensemble copula coupling and member-bymember postprocessing[J]. Quart J Roy Meteor Soc, 2017, 143 (703): 999-1008.
[9]	VRAC M, FRIEDERICHS P. Multivariate—intervariable, spatial, and temporal—bias correction[J]. J Climate, 2015, 28 (1): 218-237.
[10]	SCHEUERER M, HAMILL T M, WHITIN B, et al. A method for preferential selection of dates in the Schaake shuffle approach to constructing spatiotemporal forecast fields of temperature and precipitation[J]. Water Resour Res, 2017, 53 (4): 3 029-3 046.
[11]	MASTERS T. Practical neural network recipes in C++[M]. Academic Press, 1993: 493.
[12]	BISHOP C M. Neural Networks for Pattern Recognition[M]. Clarendon Press, Oxford. 1996: 482.
[13]	RUMELHART D E, HINTON G E, WILLIAMS R J. Learning internal representations by error propagation[M]. Parallel Distributed Processing, Vol. 1, Rumelhart and McClelland, Eds. Cambridge, MA:M.I.T.Press, 1986.
[14]	SCHALKOFF R. Pattern Recognition: Statistical, Structural and Neural Approaches[M]. New York: wiley, 1992.
[15]	GARDNER M W, DORLING S R. Arti cial neural networks(the multilayer perceptron) --A review of applications in the atmospheric sciences[J]. Atmos Environ, 1998, 32 (14-15): 2 627-2 636.
[16]	MCGOVERN A, ELMORE K L, GAGNE D J, et al. Using artificial intelligence to improve real-time decisionmaking for high-impact weather[J]. Bull Amer Meteor Soc, 2017, 98:2 073-2 090.
[17]	RASP S, LERCH S. Neural networks for postprocessing ensemble weather forecasts[J]. Mon Wea Rev, 2018, 146: 3 885-3 900.
[18]	PATIL K, DEO M C, RAVICHANDRAN M. Prediction of sea surface temperature by combining numerical and neural techniques[J]. J Atmos Ocean Technol, 2016, 33: 1 715-1 726.
[19]	BANKERT R. Cloud classification of AVHRR imagery in maritime regions using a probabilistic neural network[J]. J Applied Meteorology, 1994, 33(8): 909-918.
[20]	MARZBAN C, STUMPF G. A neural network for tornado prediction based on Doppler radar-derived attributes[J]. J Applied Meteor, 1994, 35(5): 617-626.
[21]	MARZBAN C, STUMPF G. A neural network for damaging wind prediction[J]. Wea Forecasting, 1998, 13 (1):151-163.
[22]	SAUTER T, VENEMA V. Natural three-dimensional predictor domains for statistical precipitation down-scaling[J]. J Climate, 2011, 24 (23): 6 132-6 145.
[23]	LAKSHMANAN C K, KRAUSE J, TANG L. Quality control of weather radar data using polarimetric variables[J]. J Atmos Oceanic Technology, 2014, 31 (6): 1 234-1 249.
[24]	TAO, Y, GAO X, HSU K, et al. A deep neural network modeling framework to reduce bias in satellite precipitation products[J]. J Hydrometerorology, 2016, 17(3): 931-945.
[25]	LAKSHMANAN V, RABIN R, DEBRUNNER V. Identifying and tracking storms in satellite images[C]//Second Artificial Intelligence Conference, Long Beach, CA, American Meteorological Society, 2000, 90-95.
[26]	LAKSHMANAN A F, SMITH T, HONDL K, et al. An automated technique to quality control radar reflectivity data[J]. Journal of Applied Meteorology, 2007, 46 (3): 288-305.
[27]	NEWMAN J, LAKSHMANAN V, HEINSELMAN P L, et al. Range-correcting azimuthal shear in doppler radar data[J]. Wea Forecasting, 2013, 28: 194-211.
[28]	PALMER T N. The ECMWF ensemble prediction system: Looking back (more than) 25 years and projecting forward 25 years[J]. Quart J Roy Meteor Soc, 2019, 145(51): 12-24.
[29]	代刊, 朱跃建, 毕宝贵.集合模式定量降水预报的统计后处理技术研究综述[J].气象学报, 2018, 76(4):493-510.
[30]	智协飞, 赵忱.基于集合成员订正的强降水多模式集成预报[J].应用气象学报, 2020, 31(3): 303-314.
[31]	智协飞, 彭婷, 王玉虹.基于BMA方法的地面气温的10~15 d延伸期概率预报研究[J].大气科学学报, 2018, 41(5) : 627-636.
[32]	程文聪, 史小康, 张文军, 等.基于深度学习的数值模式降水产品降尺度方法[J].热带气象学报, 2020, 36(3):307-316.
[33]	熊敏诠.基于集合预报系统的日最高和最低气温预报[J].气象学报, 2017, 75(2): 211-222.
[34]	XIONG M, DAI K. The Optimization algorithm based on neural networks in post-processing ensemble forecasts[C]//in: WANG Y, FU M, XU L, ZOU J (eds) Signal and Information Processing, Networking and Computers. Lecture Notes in Electrical Engineering, 2020, vol 628. Springer, Singapore. 772-780.
[35]	BRADLEY A A, SCHWARTZ S S. Summary verification measures and their interpretation for ensemble forecasts[J]. Mon Wea Rev, 2011, 139: 3 075-3 089.
[36]	ZHU Y, ZHOU X, PEÑA M, et al. Impact of Sea Surface Temperature forcing on weeks 3 and 4 forecast skill in the ncep global ensemble forecasting system[J]. Wea Forecasting, 2017, 32: 2 159-2 174.
[37]	GUAN H, ZHU Y, SINSKY E, et al. Systematic Error Analysis and Calibration of 2-m Temperature for the NCEP GEFS Reforecast of the Subseasonal Experiment (SubX) Project[J]. Wea Forecasting, 2019, 34: 361-376.
[38]	HAGEDORN R, HAMILL T M, WHITAKER J S. Probabilistic forecast calibration using ecmwf and gfs ensemble reforecasts. Part Ⅰ: TwoMeter Temperatures[J]. Mon Wea Rev, 2008, 136: 2 608-2 619.
[39]	GASCON E, LAVERS D, HAMILL T M, et al. Statistical postprocessing of dual-resolution ensemble precipitation forecasts across Europe [J]. Quart J Roy Meteor Soc, 2019, 145 (724): 3 218-3 235.