通过定义最优配对分段插补法，以NewD66/D66、 D105、 Amdo和BJ为主站，D66/五道梁气象站（WDL）、 D110、 MS3478和MS3608为辅站，对主站缺失的地面气温资料进行插补，以获得完整的气温序列，并以此为基础分析主站近期气温变化。主站和辅站气温一致性分析结果表明，一年中每两个配对站的气温变化均有很好的一致性；温差表现为冬半年大、 夏半年小。插补效果分析表明，插补效果夏季比冬季好；插补效果D66站最好，D105站相对较差；插补误差近似服从正态分布，时间尺度越大插补结果的可用性越强；大幅降温、 降水、 较大风速以及较大风向转变是影响插补效果的主要因素。对主站完整气温序列分析表明，NewD66站的气温年较差最大（26℃），Amdo站最小（19℃）；BJ站多年平均气温最高（-0.3℃），D105站最低（-5℃）；BJ站处于季节冻土区，其余三站处于多年冻土区 ；近十几年主站年平均气温均呈波动上升趋势，BJ站和NewD66站升温明显，D105站和Amdo站升温缓慢。
In order to obtain complete sequences of temperature data, an optimal pairing segmented interpolation method was established in this paper to recover the missing temperature data of objectiive sites (NewD66/D66, D105, Amdo and BJ) according to the temperature relationships between the objectiive sites and reference sites(D66/WDL, D110, MS3478 and MS3608). Furthermore, the recent temperature variations of objective sites were analyzed based on the obtained temperature data of objective sites. Firstly, the analysis shows that the temperature varaitions of objectiive sites are consitent with those of their corresponding reference sites, and that the temperature differences between objective and reference sites are large in winter while small in summer, which is the basis of optimal pairing segmented interpolation method. Then, the analysis indicates that the interpolation result is more acceptable in summer than that in winter, and the result in D66 site is best among all objective sites but that in D105 is relatively poor. The analysis also indicates interpolation values become more and more similar to observation values with time scale becoming larger and laeger because of the fact that interpolation error is approximately normally distributed, and the less acceptable interpolation result is mainly due to the facters such as substantial drop of temperature, precipitation, high wind speed and significant changes of wind direction events. Finally, the analysis based on complete temperature data of the four objective sites shows that the differences between maximun and minimun monthly average temperatue is greatest in NewD66 site while smallest in Amdo, that the average temperature of recent years is highest in BJ site and lowest in D105 site, that BJ site locates in seasonal frozen area, while the other three sites are in permafrost area. Over the last more than ten years, annual averages of temperature appear a growing trend in four obejective sites, and the trend in BJ and NewD66 is, by contrast, more obivious than that in D105 and Amdo sites.
[1]王海军, 涂诗玉, 陈正洪. 日气温数据缺测的插补方法试验与误差分析[J]. 气象, 2008, 34(7): 83-91.
[2]Huth R, Nemeov I. Estimation of missing daily temperature: Can a weather categorization improve its accuracy?[J]. J Climate, 1995, 8: 1901-1916.
[3]DeGaetano A T, Eggleston K L, Knapp W W. A method to estimate missing daily maximun and minimun temperature observations[J]. J Appl Meteor, 1995, 34: 371-380.
[4]江志红, 丁裕国, 屠其璞. 基于 PC-CCA 方法的气象场资料插补试验[J]. 南京气象学院学报, 1999, 22(2): 141-148.
[5]江志红, 屠其璞. 20 世纪全球表面温度场序列的插补试验[J]. 南京气象学院学报, 2001, 24(1): 26-36.
[6]张永领, 丁裕国, 高全洲, 等. 一种基于 SVD 的迭代方法及其用于气候资料场的插补试验[J]. 大气科学, 2006, 30(3): 526-532.
[7]李庆祥, 黄嘉佑, 鞠晓慧. 上海地区最高气温资料的恢复试验[J]. 热带气象学报, 2008, 24(4): 349-353.
[8]涂诗玉, 陈正洪. 武汉和宜昌缺测气温资料的插补方法[J]. 湖北气象, 2001, 3: 11-13.
[9]王智, 师庆东, 常顺利, 等. 新疆地区平均气温空间插值方法研究[J]. 高原气象, 2012, 31(1): 201-208.
[10]张秀芝, 孙安健. 利用车贝雪夫多项式进行资料缺测插补的研究[J]. 应用气象学报, 1996, 7(3): 344-352.
[11]张秀芝, 孙安健. 气候资料缺测插补方法的对比研究[J]. 气象学报, 1996, 54(5): 625-632.
[12]李庆祥, 彭嘉栋, 沈艳. 1900-2009 年中国均一化逐月降水数据集研制[J]. 地理学报, 2012, 67(3): 301-311.
[13]田琳, 王龙, 余航, 等. 基于 BP 神经网络的缺测降水数据插补[J]. 云南农业大学学报: 自然科学版, 2012, 27(2): 281-284.
[14]Young K C. A three-way model of interpolating monthly precipitation values[J]. Mon Wea Rev, 1992, 120(11): 2561-2569.
[15]胡振菊, 何炳文, 高伟, 等. 无资料地质灾害隐患点降水插补方法试验[J]. 贵州气象, 2010, 34(B09): 132-134.
[16]李庆祥, 屠其璞. 近百年北半球陆面降水资料的插补及初步分析[J]. 南京气象学院学报, 2000, 23(4): 528-535.
[17]张志萍, 冉大川, 慕志龙. 大理河流域降水资料插补方法探讨[J]. 人民黄河, 2006, 28(12): 26-27.
[18]王敏, 周才平, 吴良, 等. 遥感估算降水在西藏高原中的应用研究[J]. 高原气象, 2012, 31(5): 1215-1224.
[19]崔林丽, 杨引明, 游然, 等. FY-3A/MWHS数据在定量降水估计中的应用研究[J]. 高原气象, 2012, 31(5): 1439-1445.
[20]庄薇, 刘黎平, 王改利, 等. 青藏高原复杂地形区雷达估测降水方法研究[J]. 高原气象, 2013, 32(5): 1224-1235, doi: 10.7522/j.issn.1000-0534.2012.00118.
[21]Acock M C, Pachepsky Y A. Estimating missing weather data for agricultural simulations using group method of data handling[J]. J Appl Meteor, 2000, 39: 1176-1184.
[22]王伯民. 基本气象资料质量控制综合判别法的研究[J]. 应用气象学报, 2004, 15(B12): 50-59.
[23]Igor Zahumensky. Guidelines on quality control procedures for data from automatic weather stations[R]. WMO-No.305,Geneva Switzerland, 28 June-2 July 2004.
[24]Gandin L. Complex quality control of meteorological observations[J]. Mon Wea Rev, 1988, 116(5): 1137-1156.
[25]王超, 韦志刚, 李振朝. 敦煌戈壁气象塔站资料的质量控制[J]. 干旱气象, 2010, 28(2): 121-127.
[26]徐安伦, 李建, 刘辉志, 等. 大理国家气候观象台地面辐射观测数据的质量控制研究[J]. 高原气象, 2013, 32(5): 1432-1441, doi: 10.7522/j.issn.1000-0534.2012.00133.
