北京地区自动气象站气温观测资料的质量评估

Evaluating the Quality of Temperature Measured at Automatic Weather Stations in Beijing

该文从完整性、准确性、可靠性3个方面设计了自动气象站数据质量评估的流程,并对北京地区187个自动气象站1998—2009年逐时气温资料进行了质量评估。结果显示:北京地区自动气象站建设初期重点兼顾城区和山区的布设策略,为北京区域气候研究提供了较好的基础;北京地区正常运行的自动气象站发生中度和重度缺测的站点相对较少,离散和轻度连续的缺测较集中,具有较好的区域一致性;错误发生率最高为3.8%,大多数年份错误发生率均在1%以下,可见自动气象站错误数据相对较少:尽管可疑数据涉及的站点相对较多,但经空间一致性检查后,有超过50%可归并为正确数据。评估分析结果表明:北京地区自动气象站数据具有一定的准确性和可靠性,具有较强的应用潜力和前景。

Data quality is a basic assurance for meteorological researches and data applications.Considering data integrality,veracity and confidence,the standard for AWS（automatic weather station） data quality assessment is defined band a set of index is established for AWS data quality assessment and a feasible observation data quality control flow is designed.Following the definition and principia,AWS data can be categorized as correct,mistake and dubious data.The spatial and temporal consistent detections are employed in the data quality control flow.Based on the quality control flow and assessment index,hourly data measured by 187 AWS in Beijing from 1998 to 2009 is evaluated.The results show that the AWS net of Beijing is set up following a fine layout.Although the distribution of AWS is uneven,most AWS is located in urban area such as Haidian and Chaoyang districts,especially in early time of the AWS construction process, some AWS are set up in mountainous area according to the need of different region representative.It is beneficial for urban and rural climate comparing,data sequence reconstruction and regional climate research. The severe missing data is rare,and discrete and slight continuous missing data is just found in concentrated regions and with regional consistent characteristic.The highest error rate of AWS temperature data is 3.8%and it is below 1%in most years.The amount of dubious data is much more than the mistaking data.But more than half of dubious data can be got back after space consistency check.It means that the AWS data in Beijing is reliable.The rates of mistaken data are above 20%after 2004.It shows that the amount of dubious data is not related with the rate of mistake data and on the other hand the uncertainty of AWS data set is reduced after 2004.In conclusion,the assessment result reveals that the AWS data in Beijing are accurate,reliable and show great potential in the future application.