全球地面气温和降水定时值数据整合与初步质量控制

Integration and preliminary quality control of global surface hourly temperature and precipitation data

全球地面气温和降水数据是开展气候变化研究的基础,定时值观测数据是气候监测的重要依据。本文基于美国国家气候数据中心的全球陆地小时数据集(Integrated Surface Hourly Database,ISD)和国际通讯系统数据(Global Transmission System,GTS)两种全球地面数据产品,对气温和降水要素定时值数据进行了实时整合,并采用空间一致性检查、内部一致性检查、气候界限值检查及台站极值检查等方法对整合后的数据进行了初步质量控制,最终形成了实时更新的全球地面气温和降水定时值数据产品。结果表明:2015年整合后的地面气温和降水数据共包含全球11 990个气象站点的定时值数据,气象站点分布基本覆盖全球各地区,与整合前相比,整合后各区域气象站点密度和气象要素数据量均明显增加,特别在亚洲、欧洲和北美洲地区数据量增加显著;整合后数据的准确率和疑误率分别为94.4%和5.6%,疑误数据主要分布在欧洲东北、亚洲及中国地区,造成数据可疑和错误的主要原因包括内部一致性矛盾、超过气候界限值等。

Global surface temperature and precipitation data are the foundation of research on climate change, and the hourly observational data are the important basis of climate monitoring. Based on two kinds of global surface data products such as the Integrated Surface Hourly Database （ISD）and the Global Transmission System （GTS） from U. S. National Climatic Data Center, the hourly temperature and precipitation data were integrated in real time. The integrated data were processed with preliminary quality control using methods of spatial consistency checks, internal consistency checks, boundary climate checks, and extreme climate checks, and eventually formed a real-time updated datasets of global surface hourly temperature and precipitation. The results show that the integrat- ed surface temperature and precipitation data since 2015 contain hourly data from 11,990 stations on global that al- most cover the whole world. Compared to the original data without integration, the density of stations and the a- mount of meteorological elements in all regions from the integrated dataset increase significantly, especially in re- gions of Asia ,Europe, and North America. The accuracy and suspected error rate of integrated data is 94. 4% and 5.6%, respectively. The data with suspected error mainly exist in the regions of northeast Europe ,Asia, and China, and the internal consistency conflict and beyond the climate boundary limit value are main reasons for suspect error of data.