利用非差法(准)实时估计GNSS天顶对流层延迟

GNSS zenith tropospheric delay (ZTD) contains rich information of atmospheric water vapor content. The assimilation of ZTDs into numerical weather prediction models can greatly improve forecasting accuracy. However, the application of ZTDs in weather nowcasting is restricted due to the current estimation method is difficult to achieve ZTD (near) real-time and high precision estimating because of low efficiency parameters computation, predicted orbit errors, imperfect stochastic model of multi-GNSS combined and other factors. In this respect, the project intends to conduct in-depth theoretical and experimental research on this scientific issue about (near) real-time estimating GNSS ZTD using undifferenced modeling. Key issues for this project are as follows: (1) Least square collocation of parameters elimination and square root information filtering is promoted to improve computation efficiency. (2) The method of relaxing orbital parameters and data detection is used to compensate predicted orbit errors. (3) Robust Helmert variance component estimation is promoted to improve stochastic model of multi-GNSS combined. The project attempts to explore optimization methods of ZTD (near) real-time estimating. The research results can provide (near) real-time and high precision ZTDs for numerical weather prediction model assimilation, to improve the capability for severe weather instant monitoring and forecasting, which has important scientific value and social value.

GNSS天顶对流层延迟（ZTD）蕴含丰富的大气水汽含量信息，将其估计值同化到数值天气预报模式中，可以极大提高预报准确度。然而，由于受到低效参数解算、预报轨道粗差以及Multi-GNSS融合随机模型欠完善等因素影响，目前的ZTD估计方法难以实现对其（准）实时和高精度估计，从而制约了ZTD在即时天气预报中的应用。为此，针对利用非差法（准）实时估计GNSS ZTD这一科学问题开展深入的理论和实验研究，并重点研究：（1）最小二乘配置消参法和平方根信息滤波，以提高ZTD参数解算效率；（2）松弛轨道参数法和数据探测法，以补偿轨道粗差；（3）抗差Helmert方差分量估计，以改善Multi-GNSS融合随机模型。项目研究试图优化ZTD（准）实时估计方法。研究成果可为数值天气预报模式同化提供（准）实时和高精度的ZTD估值，以提高灾害性天气即时监测和预报能力，具有重要的科学价值和社会价值。

GNSS天顶对流层延迟（ZTD）蕴含丰富的大气水汽含量信息，将其估计值同化到数值天气预报模式中，可以极大提高预报准确度。然而，由于受到低效参数解算、预报轨道粗差以及Multi-GNSS融合随机模型欠完善等因素影响，目前的ZTD估计方法难以实现对其（准）实时和高精度估计，从而制约了ZTD在即时天气预报中的应用。为此，优化了利用非差法估计GNSS ZTD相关算法。在非差PPP法中，提出了一套适用于PPP的实时质量控制方法，能有效探测并修复钟跳，探测周跳和剔除粗差；利用验后残差建立了等价权函数合理确定观测值随机模型，消除数据预处理后残余粗差或周跳的影响；利用平方根信息滤波单测站数据处理时间为5.2 s,GPS ZTD估计精度为7.5 mm。在非差网解法中，利用最小二乘配置消参法提高了数据处理效率；采用松弛轨道补偿预报轨道粗差显著提高了ZTD估计精度。研究成果可为数值天气预报模式同化提供（准）实时和高精度的ZTD估值，以提高灾害性天气即时监测和预报能力，具有重要的科学价值和社会价值。