利用2003—2012年兰州市月均空气污染指数API和季节性时间序列模型SARIMA, 对数据序列进行拟合并对其变化趋势进行了分析, 研究其演变趋势并选取空气质量良好(月平均API<150)的月份(2012年6月2013年2月), 选用时间序列模型ARIMA(1,1,1)进行拟合, 再结合残差控制图, 对2013年3月日平均API预测监控.结果表明, SARIMA(0, 1, 1)×(0, 1, 1)12模型较好的拟合了近10年兰州市API的变化趋势, 兰州市的空气污染大致呈冬、春季污染相对较重, 夏、秋季相对较轻, 空气质量总体在逐渐好转.适合该地区短时间尺度的时间序列模型是ARIMA(1, 1, 1), 结合残差控制图对2013年3月API进行预测监控, 结果显示6, 9, 10, 11, 12和13日超出控制限, 对6日和9日提出预警.预测结果与实际结果相吻合, 证实了将时间序列模型与残差控制图结合预测监控大气污染的有效性.
The mean monthly air pollution index (API) in Lanzhou from 2003 to 2012 was fitted by time series model SARIMA, and seasonal variation and trend of the API was analyzed. Using combined SARIMA model and residual control chart method, a model has been established to forecast daily API in Lanzhou. This model firstly need to select an appropriate stage as a control state from the fitted API data series from 2003 to 2012 using the time series model ARIMA. In the present study, we selected the months from June 2012 to February 2013 as the control state to establish control limits based on the ARIMA (1, 1, 1) fitting for 2003—2012 API data. The model then employed a residual control chart method to forecast and monitor the average daily API in Lanzhou in March 2013. Results indicate that SARIMA (0, 1, 1)×(0, 1, 1)12 model yielded a better-fit to API in Lanzhou City, showing higher air pollution in winter and spring and relatively mild pollution during summer and autumn. Overall both modeled and monitored data revealed the improvement of air quality in Lanzhou for the last decade. Residual control charts model for March 2013 API established in this study predicted that API on 6, 9, 10, 11, 12, 13 exceeded the control limits, and issued a warning on API in 6 and 9 March, respectively. The model forecasting results are consistent with the monitored results, confirming the usefulness and reliability of the model using combined time series model and residual control chart.
[1]王斌, 高会旺. 中国沿海城市空气污染指数的分布特征[J]. 生态环境, 2008, 17(2): 542-548.
[2]Luis A D, Juan C O, Joshua S F, et al. A hybrid ARIMA and artificial neural networks model to forecast particulate matter in urban areas: The case of Temuco, Chile[J]. Atmos Environ, 2008, 42: 8331-8340.
[3]王俭, 胡筱敏, 郑龙熙, 等. 基于BP模型的大气污染预报方法的研究[J]. 环境科学研究, 2002, 15(5): 62-64.
[4]Gardner M W, Dorling S R. Neural network modeling and prediction of hourly NO<sub>X</sub> and NO<sub>2</sub> concentrations in urban air in London[J]. Atmos Environ, 1998, 33: 709-719.
[5]Alwan L C, Roberts H V. Time-series modeling for statistical process control[J]. Journal of Business and Economics Statistics, 1988, 6(1): 87-95.
[6]赵仲莲, 戚登臣, 杨德保, 等. 兰州市三种主要空气污染物(SO<sub>2</sub>, NO<sub>2</sub>, PM<sub>10</sub>)的统计预报方法[J]. 甘肃科技, 2006, 22(12): 102-103.
[7]陶燕, 刘亚梦, 米生权, 等. 大气细颗粒物的污染特征及对人体健康的影响[J]. 环境科学学报, 2014, 34(3): 592-597.
[8]Box G E P, Jenkins G M, MacGregor J F. Some recent advances in forecasting and control: Part II[J]. Applied Statistics, 1974, 23: 158-179.
[9]Box G E P, Jenkins G M. Time Series Analysis: Forecasting and Control[M]. San Fransisco: Holden-daypress, 1976: 575.
[10]Cryer J D, Chan K S. Time Series Analysis with Applications in R[M]. 2nd Edn. NewYork: Springer-Verlay, 2008: 491.
[11]白永飞, 徐丽红, 郭支喜, 等. 山西省布鲁氏菌病时间序列自回归移动平均模型分析[J]. 疾病监测, 2011, 26(8): 647-650.
[12]Lee M H, Rahman N H A, Suhartono M T, et al. Seasonal ARIMA for forecasting air pollution index: A case study[J]. American Journal of Applied Sciences, 2012, 9(4): 570-578.
[13]Wardell D G , Moskowits H , Plante R D. Control charts in the presence of data correlation[J]. Management Science, 1992, 38: 1084-1105.
[14]Lin S W, Adams B M. Combined control charts for forecasted-based monitoring schemes[J]. Journal of Quality Technology, 1996, 28: 289-301.
[15]孙静. 自相关过程的残差控制图[J]. 清华大学学报: 自然科学版, 2002, 42(6): 735-738.
[16]杨位钦, 顾岚. 时间序列分析与动态数据建模[M]. 北京: 北京工业学院出版社, 1987: 350-351.
[17]王振龙, 胡永宏. 应用时间序列分析[M]. 北京: 中国统计出版社, 2000: 21-22.
[18]张蔚, 张彦琦, 杨旭. 时间序列资料ARIMA季节乘积模型及其应用[J]. 第三军医大学学报, 2002, 24(8): 955-957.
[19]孟凡强. ARIMA模型在空气污染指数预测中的应用[J]. 统计与决策, 2009, (7): 33-35.
[20]张强. 地形和逆温层对兰州市污染物输送的影响[J]. 中国环境科学, 2001, 21(3): 230-234.
[21]丁瑞强, 王式功, 尚可政, 等. 中国春季沙尘暴的趋势变化及年代际变化[J]. 高原气象, 2004, 23(5): 660-666.
[22]尚可政, 达存莹, 付有智, 等. 兰州城区稳定能量及其与空气污染的关系[J]. 高原气象, 2001, 20(1): 76-81.
[23]赵敬国, 王式功, 王嘉媛, 等. 兰州市空气污染与气象条件关系分析[J]. 兰州大学学报: 自然科学版, 2013, 49(4): 491-496.
[24]马敏劲, 郭世奇, 王式功, 等. 近11年兰州空气污染特征及其边界层结构影响的分析[J]. 兰州大学学报: 自然科学版, 2012, 48(6): 69-73.
[25]陶健红, 黄玉霞, 陆登荣. 河西走廊沙尘活动对兰州PM<sub>10</sub>浓度的影响及其评估[J]. 中国沙漠, 2007, 27(4): 672-676.
[26]陈晓光, 张存杰, 董安祥, 等. 甘肃省沙尘暴过程的划分及统计分析[J]. 高原气象, 2004, 23(3): 374-381.
[27]张强, 胡隐樵. 浅探兰州市城区大气环境污染与治理的若干问题[J]. 高原气象, 1998, 17(2): 204-210.
[28]陈雷华, 余晔, 陈晋北, 等. 2001-2007年兰州市主要大气污染物污染特征分析[J]. 高原气象, 2010, 29(6): 1627-1633.
[29]陶杰, 张雪芹, 陶建强, 等. 气候变化趋势分析中自相关的检验与去除[J]. 应用气象学报, 2008, 19(1): 47-51.
