新疆1993—2012年辐射观测资料质量分析

Analysis of Radiation Observational Data Quality from 1993 to 2012 in Xinjiang

利用新疆11个辐射站最近20 a(1993—2012年)共2 640个辐射信息化数据文件,从原始数据表-33及月辐射数据(R)文件中遴选出疑误数据共计7 025条,其中更改错误数据323个,对疑误、错误数据进行了质量检查及评估,并应用数理统计方法分析了其时空分布特征以及总辐射、散射辐射、直接辐射、反射辐射、净全辐射出现疑误数据的概率,同时对出现疑误、错误数据的原因进行了分析,结果表明:新疆最近20 a辐射疑误数据主要发生并集中在一、二级辐射观测站,其中数据文件部分出现疑误较多,记录格式部分出现错误相对较少;疑误数据出现的主要原因是由于辐射参数的设置只能考虑正常天气状况,而异常天气情况比较复杂,很难完全涵盖。此外直接辐射表误差和自动跟踪架误差也是出现疑误数据较多的原因。整体而言,自2003年辐射数据采集质控实现自动化后,经质量控制软件检查的R文件出现疑误、错误数据的概率显著减少,资料可靠,可为科研工作提供有利的技术支持。

2 640 data files of radiation informatization over last 20 years （1993-2012）from 11 radiation stations in Xinjiang were examined. With original data sheet-33 and monthly radiation data （R）files, 7 025 doubted data were selected, including 323 erroneous data to be corre-cted. Such doubted and erroneous data were treated with quality inspection and evaluation, and their probabilities with respect to spatial-temporal distribution characteristics, total radiation, scattered radiation, direct radiation, reflected radiation and net total radiation were analyzed with mathematical statistical method; and the reasons were also analyzed. The results indicated that the doubted data occurs and concentrates mainly at the primary and secondary radiation observation stations, mostly in the form of data files and record format. The major reason was that only normal weather conditions were considered for the setting of radiation parameters, but abnormal weather conditions are too complicated to be covered fully. Besides, direct radiation meter error and auto-tracking mount error also account for many doubted data. On the whole, the probability of doubted data and erroneous data declined significantly in R files examined with quality control software since radiation data acquisition quality control was automated in 2003. The data from R files are reliable to provide technical support to scientific research.