高级检索
司鹏, 徐文慧. 利用RHtestsV4软件包对天津1951~2012年逐日气温序列的均一性分析[J]. 气候与环境研究, 2015, 20(6): 663-674. DOI: 10.3878/j.issn.1006-9585.2015.15014
引用本文: 司鹏, 徐文慧. 利用RHtestsV4软件包对天津1951~2012年逐日气温序列的均一性分析[J]. 气候与环境研究, 2015, 20(6): 663-674. DOI: 10.3878/j.issn.1006-9585.2015.15014
SI Peng, XU Wenhui. Homogenization of Tianjin Daily Surface Air Temperatures by Software Pack RHtestsV4 during 1951-2012[J]. Climatic and Environmental Research, 2015, 20(6): 663-674. DOI: 10.3878/j.issn.1006-9585.2015.15014
Citation: SI Peng, XU Wenhui. Homogenization of Tianjin Daily Surface Air Temperatures by Software Pack RHtestsV4 during 1951-2012[J]. Climatic and Environmental Research, 2015, 20(6): 663-674. DOI: 10.3878/j.issn.1006-9585.2015.15014

利用RHtestsV4软件包对天津1951~2012年逐日气温序列的均一性分析

Homogenization of Tianjin Daily Surface Air Temperatures by Software Pack RHtestsV4 during 1951-2012

  • 摘要: 基于国家气象信息中心“基础气象资料建设专项”研制的中国地面历史基础气象资料及台站元数据,利用RHtestV4软件包对天津1951~2012年的历史气温序列进行了均一性分析。结果显示,通过惩罚最大t检验(PMT)方法对12个地面站逐日平均、最高、最低气温序列检验得到,迁站是导致平均气温和最低气温序列突变的主要原因,同类型仪器更换则是导致最高气温序列突变的主要原因,而2005年以后的自动站业务化并没有对天津地区气温序列的均一性造成很大影响。同时,研究中也检验出少部分未知原因的显著间断点,可能是由于观测员的误判或观测仪器翘变等因素造成的历史数据疑误。从订正量来看,逐日平均气温和最高气温序列主要以正偏差订正为主,而最低气温则主要以负偏差订正为主。其中,最高气温序列的分位数匹配(QM)订正量均值最大,90%以上集中在0.1~1.0 ℃,平均气温序列的QM订正量均值相对最小,90%以上的订正量在-0.7~0.7 ℃,而最低气温的订正量90%以上集中在-1.5~1.5 ℃范围中。另外,以Xu et al.(2013)研制的数据集为参照,通过误差分析,发现两类研究得到的年(季节)尺度气温数据具有较高的相符率和一致性,从而可以说明本研究订正后的天津地区1951~2012年逐日气温序列具有一定的可靠性。

     

    Abstract: Based on historical ground-level meteorological data and their metadata in Tianjin, China, produced by the National Meteorological Information Center, China Meteorological Administration, the homogenization of the daily surface air temperature data for Tianjin during the last 60 years, by the software package RHtestV4, is tested. Results indicate that the main cause of the discontinuities in daily mean and minimum temperatures is station relocation, as tested by the penalized maximal t-test; while for daily maximum temperature, it is instrument change. Furthermore, the operation of automatic stations after 2005 does not have a great impact on the homogeneity of temperature at Tianjin. Moreover, some significant discontinuities for unknown reasons are also detected, which we think may be due to reasons such as observer misjudgment or instrument warping. For the adjustment amount, the positive bias corrections are mainly expressed in daily mean and maximum temperature series. Meanwhile, the negative bias corrections are mainly in daily minimum temperature series, in which the largest average quantile-matching (QM) adjustments are for daily maximum temperature (more than 90% are from 0.1 ℃ to 1.0 ℃), and the smallest are for daily mean temperature (more than 90% are from -0.7 ℃ to 0.7 ℃). For daily minimum temperature, the range is -1.5 ℃ to 1.5 ℃. In addition, error analysis reveals that there is much higher consistency between the two datasets of the present paper and those of a previous study, on annual and seasonal scales, by setting the datasets from the previous study as the reference. This can explain why the daily temperature series for the period 1951-2012 in Tianjin, adjusted in this study, show a certain reliability.

     

/

返回文章
返回