规模GNSS网数据处理一体化方案与中国大陆水平格网速度场模型构建研究

针对当前大规模GNSS网在数据预处理、组织管理、基准站选取、处理策略、分区方案、质量检核与评估以及现有研究中速度场模型数据量小、连续性不佳、现势性不强等问题，首先基于2011~2017年陆态网260多个连续运行基准站的观测数据探讨大网解算的一体化数据处理流程，给出适用于大规模GNSS网数据处理的优化原则与质量评估方案，并着重分析不同的数据解算成果评估策略，其对大规模GNSS站网的高精度数据解算具有借鉴和参考意义；然后基于解算得到的ITRF2014框架下高精度速度场成果，采用克里金插值构建中国大陆水平格网速度场模型，并利用国内IGS站和陆态网区域站的精确速度值，对格网内插结果进行外部检验，结果表明，东、北方向速度残差的平均误差和中误差均小于2 mm/a，验证了模型的准确性、可靠性、简便性与现势性。

Research on an Integrational Scheme of Large-Scale GNSS Network Data Processing and Establishment of the Horizontal Grid Velocity Field Model in Chinese Mainland

In view, not only of the challenges facing the large-scale GNSS network in data preprocessing, organization and management, selection of reference stations, processing strategies, partition schemes and quality evaluation, but also the limitation of the existing velocity field model such small data, poor continuity and timeliness, we present an integrational scheme for large-scale GNSS network data processing and establish the horizontal grid velocity field model in Chinese mainland. Based on observation data of more than 260 continuous operation reference stations in crustal movement observation network of China(CMONOC) from 2011 to 2017, we first discuss the integrated scheme of large network data processing, then give the optimization principle and quality evaluation mode for large-scale GNSS network data processing, and analyze the different evaluation strategies of the computed results, which plays an important role in high-precision data processing of a large-scale GNSS network. Finally, based on the velocity field of high-precision ITRF2014, the Kriging interpolation model is applied to construct the horizontal grid velocity field model of Chinese mainland. We utilize the accurate velocity field of all IGS stations and the regional stations to assess our results of the horizontal grid velocity field in Chinese mainland. The average and medium error of the horizontal residual velocity of the two results are less than 2 mm/a, which indicates besides good timeliness and simplicity. Therefore, the accuracy and reliability of the model can meet the requirements of users.