一种最优多模式集成方法在我国重污染区域PM_(2.5)浓度预报中的应用
|
摘要
为了提高我国重污染区域PM_(2. 5)浓度预报准确率,基于4套国家级以及区域环境气象业务中心发展和维护的空气质量数值预报模式,通过均值集成、权重集成、多元线性回归集成和BP-ANNs集成分别建立集成预报,在实时预报效果评估基础上,建立了最优多模式集成预报。对2015—2016年预报效果进行评估,结果表明:相对于单个空气质量数值预报模式,均值和权重集成对预报偏差的改进幅度有限,但多元线性回归、BP-ANNs和最优集成能较大幅度降低预报偏差;最优集成预报与观测值间的归一化平均偏差(NMB)和均方根误差(RMSE)分别为-10%～10%和10～70μg/m3,且在更多的站点表现出强相关性,但依然低估了高污染等级的PM_(2. 5)浓度。对2018年2月25日—3月4日京津冀地区污染过程进行评估,结果表明:最优集成能较好预报出该过程中PM_(2. 5)浓度的变化趋势和量级;在北京、石家庄和郑州3个代表城市中,预报和观测值间的NMB和相关系数(R)分别为-26%～-4%和0. 49～0. 77;最优集成对轻度污染及中度污染的TS评分为0. 39～0. 73,重度污染及以上TS评分为0. 13～0. 30,能为预报员提供客观参考,但对污染峰值的预报能力还需进一步改进。
To improve the forecast accuracy of PM_(2. 5) concentration in heavily polluted regions of China, ensemble forecasts were built by mean ensemble, weighted ensemble, multiple linear regression ensemble and back propagation artificial neural networks ensemble, respectively, based on four numerical air quality models developed and maintained by national or regional environmental metrological service centers. A best multi-model ensemble forecast was established based on real-time evaluations of performances of single numerical models and ensemble methods. Through evaluation of the forecast results during 2015-2016, compared with single numerical air quality forecast models, improvements on forecast biases due to mean and weighted ensembles were limited, but multiple linear regression, back propagation artificial neural networks and best ensembles could largely reduce the forecast biases. The NMB and RMSE values between best ensemble forecast and observation were from-10% to 10% and from 10 to 70 μg/m3, respectively. Best ensemble showed strong correlation with observations at more sites compared with other ensemble methods, but also underestimated PM_(2. 5) concentrations in high pollution level.During the pollution process occurred in Jing-Jin-Ji region from February 25 to March 4, 2018, best ensemble had the ability to forecast the trend and magnitude of PM_(2. 5) concentrations. In three representative cities of Beijing,Shijiazhuang and Zhengzhou, the NMB and R values between best ensemble and observations varied from 26% to4% and from 0. 49 to 0. 77, respectively. The TS scores of best ensemble for mild and moderate pollution ranged from 0. 39 to 0. 73, and that of severe and above pollution ranged from 0. 13 to 0. 30. These indicate that best ensemble can provide a strong objective reference to forecaster, but its forecast ability of peak values needs to be further improved.
To improve the forecast accuracy of PM_(2. 5) concentration in heavily polluted regions of China, ensemble forecasts were built by mean ensemble, weighted ensemble, multiple linear regression ensemble and back propagation artificial neural networks ensemble, respectively, based on four numerical air quality models developed and maintained by national or regional environmental metrological service centers. A best multi-model ensemble forecast was established based on real-time evaluations of performances of single numerical models and ensemble methods. Through evaluation of the forecast results during 2015-2016, compared with single numerical air quality forecast models, improvements on forecast biases due to mean and weighted ensembles were limited, but multiple linear regression, back propagation artificial neural networks and best ensembles could largely reduce the forecast biases. The NMB and RMSE values between best ensemble forecast and observation were from-10% to 10% and from 10 to 70 μg/m3, respectively. Best ensemble showed strong correlation with observations at more sites compared with other ensemble methods, but also underestimated PM_(2. 5) concentrations in high pollution level.During the pollution process occurred in Jing-Jin-Ji region from February 25 to March 4, 2018, best ensemble had the ability to forecast the trend and magnitude of PM_(2. 5) concentrations. In three representative cities of Beijing,Shijiazhuang and Zhengzhou, the NMB and R values between best ensemble and observations varied from 26% to4% and from 0. 49 to 0. 77, respectively. The TS scores of best ensemble for mild and moderate pollution ranged from 0. 39 to 0. 73, and that of severe and above pollution ranged from 0. 13 to 0. 30. These indicate that best ensemble can provide a strong objective reference to forecaster, but its forecast ability of peak values needs to be further improved.
引文
[1] JIANG J K,ZHOU W,CHENG Z,et al. Particulate matter distributions in China during a winter period with frequent pollution episodes(January 2013)[J]. Aerosol and Air Quality Research,2014,15(2):494-451.
[2] WANG K,DICKINSON R,LIANG S. Clear sky visibility has decreased over land globally from 1973 to 2007[J]. Science,2009,323:1468-1470.
[3] STOCKER T F,QIN D,PLATTNER G K,et al. IPCC,the physical science basis of climate change:clouds and aerosols[M]. Cambridge:Cambridge University Press,2013.
[4] LEE P,MCQUEEN J,STAJNER I,et al. NAQFC developmental forecast guidance for fine particulate matter(PM2. 5)[J].Weather and Forecasting,2017,32(1):343-360.
[5] MENUT L,BESSAGNET B,KHVOROSTYANOV D,et al.CHIMERE 2013:a model for regional atmospheric composition modelling[J]. Geoscientific Model Development,2013,6(4):981-1028.
[6] MONTEIRO A,RIBEIRO I,TCHEPEL O,et al. Bias correction techniques to improve air quality ensemble predictions:focus on O3and PM over Portugal[J]. Environmental Modeling and Assessment,2013,18(5):533-546.
[7] SAVAGE N,AGNEW P,DAVIS L,et al. Air quality modelling using the met office unified model(AQUM OS 24-26):model description and initial evaluation[J]. Geoscientific Model Development,2013,6(2):353-372.
[8]李曼,张载勇,李淑娟,等. CUACE系统在乌鲁木齐空气质量预报中的效果检验[J].沙漠与绿洲气象,2014,8(5):63-68.LI M,ZHANG Z Y,LI S J,et al. Verification of CUACE air quality forecast in Urumqi[J]. Desert and Oasis Meteorology,2014,8(5):63-68.
[9]李晓岚,马雁军,王扬锋,等.基于CUACE系统沈阳地区春季空气质量预报的校验及修正[J].气象与环境学报,2016,32(6):10-18.LI X L,MA Y J,WANG Y F,et al. Verification and modification to spring air quality forecasted by CUACE system in Shenyang[J]. Journal of Meteorology and Environment,2016,32(6):10-18.
[10]杨关盈,邓学良,吴必文,等.基于CUACE模式的合肥地区空气质量预报效果检验[J].气象与环境学报,2017,33(1):51-57.YANG G Y,DENG X L,WU B W,et al. Verification of CUACE model in Hefei,Anhui Province[J]. Journal of Meteorology and Environment,2017,33(1):51-57.
[11] ZHOU G Q,XU J M,XIE Y,et al. Numerical air quality forecasting over eastern China:an operational application of WRFChem[J]. Atmospheric Environment,2017,153:94-108.
[12]赵秀娟,徐敬,张自银,等.北京区域环境气象数值预报系统及PM2. 5预报检验[J].应用气象学报,2016,27(2):160-172.ZHAO X J,XU J,ZHANG Z Y,et al. Beijing regional environmental meteorology prediction system and its performance test of PM2. 5concentration[J]. Quarterly Journal of Applied Meteorology,2016,27(2):160-172.
[13]邓涛,吴兑,邓雪娇,等.珠三角空气质量暨光化学烟雾数值预报系统[J].环境科学与技术,2014,36(4):62-68.DENG T,WU D,DENG X J,et al. Numerical forecast system of air quality photochemical smog over Pearl River Delta Region[J]. Environmental Science and Technology,2014,36(4):62-68.
[14] DJALALOVA I,WILCZAK J,MCKEEN S,et al. Ensemble and bias-correction techniques for air quality model forecasts of surface O3and PM2. 5during the TEXAQS-II experiment of 2006[J]. Atmospheric Environment,2010,44(4):455-467.
[15] MARECAL V,PEUCH V H,ANDERSSON C,et al. A regional air quality forecasting system over Europe:the MACC-Ⅱdaily ensemble production[J]. Geoscientific Model Development,2015,8(9):2777-2813.
[16]王自发,吴其重,ALEX G,等.北京空气质量多模式集成预报系统的建立及初步应用[J].南京信息工程大学学报,2009,1(1):19-26.WANG Z F,WU Q Z,ALEX G,et al. Ensemble air quality multimodel forecast system for Beijing(EMS-Beijing):model description and preliminary application[J]. Journal of Nanjing University of Information Science and Technology,2009,1(1):19-26.
[17]黄思,唐晓,徐文帅,等.利用多模式集合和多元线性回归改进北京PM10预报[J].环境科学学报,2015,35(1):56-64.HUANG S,TANG X,XU W S,et al. Application of ensemble forecast and linear regression method in improving PM10forecast over Beijing area[J]. Acta Scientiae Circumstantiae,2015,35(1):56-64.
[18] LIU D J,LI L. Application study of comprehensive forecasting model based on entropy weighting method on trend of PM2. 5concentration in Guangzhou,China[J]. International Journal of Environmental Research and Public Health, 2015, 12:7085-7099.
[19] OPREA M,MIHALACHE S F,POPESCU M. Computational intelligence-based PM2. 5air pollution forecasting[J].International Journal of Computers Communications&Control,2017,12(3):365-380.
[20] POPESCU M,MIHALACHE S F,OPREA M. Air pollutants and meteorological parameters influence on PM2. 5forecasting and performance assessment of the developed artificial intelligencebased forecasting model[J]. Revista De Chimie,2017,68(4):864-868.
[21] FENG X,LI Q,ZHU Y J,et al. Artificial neural networks forecasting of PM2. 5pollution using air mass trajectory based geographic model and wavelet transformation[J]. Atmospheric Environment,2015,107:118-128.
[22] QIN S S,LIU F,WANG J Z,et al. Analysis and forecasting of the particulate matter(PM)concentration levels over four major cities of China using hybrid models[J]. Atmospheric Environment,2014,98:665-675.
[23] ZHOU Q P,JIANG H Y,WANG J Z,et al. A hybrid model for PM2. 5forecasting based on ensemble empirical mode decomposition and a general regression neural network[J].Science of the Total Environment,2014,496:264-274.
[24]张伟,王自发,安俊岭,等.利用BP神经网络提高奥运会空气质量实时预报系统预报效果[J].气候与环境研究,2010,15(5):595-601.ZHANG W,WANG Z F,AN J L,et al. Update the ensemble air quality modeling system with BP model during Beijing Olympics[J]. Climatic and Environmental Research,2010,15(5):595-601.
[25]张恒德,张庭玉,李涛,等.基于BP神经网络的污染物浓度多模式集成预报[J].中国环境科学,2018,38(4):1242-1256.ZHANG H D,ZHANG T Y,LI T,et al. Forecast of air quality pollutants’concentrations based on BP neural network multimodel ensemble method[J]. China Environmental Science,2018,38(4):1242-1256.
[2] WANG K,DICKINSON R,LIANG S. Clear sky visibility has decreased over land globally from 1973 to 2007[J]. Science,2009,323:1468-1470.
[3] STOCKER T F,QIN D,PLATTNER G K,et al. IPCC,the physical science basis of climate change:clouds and aerosols[M]. Cambridge:Cambridge University Press,2013.
[4] LEE P,MCQUEEN J,STAJNER I,et al. NAQFC developmental forecast guidance for fine particulate matter(PM2. 5)[J].Weather and Forecasting,2017,32(1):343-360.
[5] MENUT L,BESSAGNET B,KHVOROSTYANOV D,et al.CHIMERE 2013:a model for regional atmospheric composition modelling[J]. Geoscientific Model Development,2013,6(4):981-1028.
[6] MONTEIRO A,RIBEIRO I,TCHEPEL O,et al. Bias correction techniques to improve air quality ensemble predictions:focus on O3and PM over Portugal[J]. Environmental Modeling and Assessment,2013,18(5):533-546.
[7] SAVAGE N,AGNEW P,DAVIS L,et al. Air quality modelling using the met office unified model(AQUM OS 24-26):model description and initial evaluation[J]. Geoscientific Model Development,2013,6(2):353-372.
[8]李曼,张载勇,李淑娟,等. CUACE系统在乌鲁木齐空气质量预报中的效果检验[J].沙漠与绿洲气象,2014,8(5):63-68.LI M,ZHANG Z Y,LI S J,et al. Verification of CUACE air quality forecast in Urumqi[J]. Desert and Oasis Meteorology,2014,8(5):63-68.
[9]李晓岚,马雁军,王扬锋,等.基于CUACE系统沈阳地区春季空气质量预报的校验及修正[J].气象与环境学报,2016,32(6):10-18.LI X L,MA Y J,WANG Y F,et al. Verification and modification to spring air quality forecasted by CUACE system in Shenyang[J]. Journal of Meteorology and Environment,2016,32(6):10-18.
[10]杨关盈,邓学良,吴必文,等.基于CUACE模式的合肥地区空气质量预报效果检验[J].气象与环境学报,2017,33(1):51-57.YANG G Y,DENG X L,WU B W,et al. Verification of CUACE model in Hefei,Anhui Province[J]. Journal of Meteorology and Environment,2017,33(1):51-57.
[11] ZHOU G Q,XU J M,XIE Y,et al. Numerical air quality forecasting over eastern China:an operational application of WRFChem[J]. Atmospheric Environment,2017,153:94-108.
[12]赵秀娟,徐敬,张自银,等.北京区域环境气象数值预报系统及PM2. 5预报检验[J].应用气象学报,2016,27(2):160-172.ZHAO X J,XU J,ZHANG Z Y,et al. Beijing regional environmental meteorology prediction system and its performance test of PM2. 5concentration[J]. Quarterly Journal of Applied Meteorology,2016,27(2):160-172.
[13]邓涛,吴兑,邓雪娇,等.珠三角空气质量暨光化学烟雾数值预报系统[J].环境科学与技术,2014,36(4):62-68.DENG T,WU D,DENG X J,et al. Numerical forecast system of air quality photochemical smog over Pearl River Delta Region[J]. Environmental Science and Technology,2014,36(4):62-68.
[14] DJALALOVA I,WILCZAK J,MCKEEN S,et al. Ensemble and bias-correction techniques for air quality model forecasts of surface O3and PM2. 5during the TEXAQS-II experiment of 2006[J]. Atmospheric Environment,2010,44(4):455-467.
[15] MARECAL V,PEUCH V H,ANDERSSON C,et al. A regional air quality forecasting system over Europe:the MACC-Ⅱdaily ensemble production[J]. Geoscientific Model Development,2015,8(9):2777-2813.
[16]王自发,吴其重,ALEX G,等.北京空气质量多模式集成预报系统的建立及初步应用[J].南京信息工程大学学报,2009,1(1):19-26.WANG Z F,WU Q Z,ALEX G,et al. Ensemble air quality multimodel forecast system for Beijing(EMS-Beijing):model description and preliminary application[J]. Journal of Nanjing University of Information Science and Technology,2009,1(1):19-26.
[17]黄思,唐晓,徐文帅,等.利用多模式集合和多元线性回归改进北京PM10预报[J].环境科学学报,2015,35(1):56-64.HUANG S,TANG X,XU W S,et al. Application of ensemble forecast and linear regression method in improving PM10forecast over Beijing area[J]. Acta Scientiae Circumstantiae,2015,35(1):56-64.
[18] LIU D J,LI L. Application study of comprehensive forecasting model based on entropy weighting method on trend of PM2. 5concentration in Guangzhou,China[J]. International Journal of Environmental Research and Public Health, 2015, 12:7085-7099.
[19] OPREA M,MIHALACHE S F,POPESCU M. Computational intelligence-based PM2. 5air pollution forecasting[J].International Journal of Computers Communications&Control,2017,12(3):365-380.
[20] POPESCU M,MIHALACHE S F,OPREA M. Air pollutants and meteorological parameters influence on PM2. 5forecasting and performance assessment of the developed artificial intelligencebased forecasting model[J]. Revista De Chimie,2017,68(4):864-868.
[21] FENG X,LI Q,ZHU Y J,et al. Artificial neural networks forecasting of PM2. 5pollution using air mass trajectory based geographic model and wavelet transformation[J]. Atmospheric Environment,2015,107:118-128.
[22] QIN S S,LIU F,WANG J Z,et al. Analysis and forecasting of the particulate matter(PM)concentration levels over four major cities of China using hybrid models[J]. Atmospheric Environment,2014,98:665-675.
[23] ZHOU Q P,JIANG H Y,WANG J Z,et al. A hybrid model for PM2. 5forecasting based on ensemble empirical mode decomposition and a general regression neural network[J].Science of the Total Environment,2014,496:264-274.
[24]张伟,王自发,安俊岭,等.利用BP神经网络提高奥运会空气质量实时预报系统预报效果[J].气候与环境研究,2010,15(5):595-601.ZHANG W,WANG Z F,AN J L,et al. Update the ensemble air quality modeling system with BP model during Beijing Olympics[J]. Climatic and Environmental Research,2010,15(5):595-601.
[25]张恒德,张庭玉,李涛,等.基于BP神经网络的污染物浓度多模式集成预报[J].中国环境科学,2018,38(4):1242-1256.ZHANG H D,ZHANG T Y,LI T,et al. Forecast of air quality pollutants’concentrations based on BP neural network multimodel ensemble method[J]. China Environmental Science,2018,38(4):1242-1256.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |