Chaoqian Xu

Real-Time GPS Precise Point Positioning-Based Precipitable Water Vapor Estimation for Rainfall Monitoring and Forecasting

GPS-based precipitable water vapor (PWV) estimation has been proven as a cost-effective approach for numerical weather prediction. Most previous efforts focus on the performance evaluation of post-processed GPS-derived PWV estimates using International GNSS Service (IGS) satellite products with at least 3-9-h latency. However, the suggested timeliness for meteorological nowcasting is 5-30 min. Therefore, the latency has limited the GPS-based PWV estimation in real-time meteorological nowcasting. The limitation has been overcome since April 2013 when IGS released real-time GPS orbit and clock products. This becomes the focus of this paper, which investigates real-time GPS precise point positioning (PPP)-based PWV estimation and its potential for rainfall monitoring and forecasting. This paper first evaluates the accuracy of IGS CLK90 real-time orbit and clock products. Root-mean-square (RMS) errors of <; 5 cm and ~0.6 ns are revealed for real-time orbit and clock products, respectively, during July 4-10, 2013. Second, the real-time GPS PPP-derived PWV values obtained at IGS station WUHN are compared with the post-processed counterparts. The RMS difference of 2.4 mm has been identified with a correlation coefficient of 0.99. Third, two case studies, including a severe rainfall event and a series of moderate rainfall events, have been presented. The agreement between the real-time GPS PPP-derived PWV and ground rainfall records indicates the feasibility of real-time GPS PPP-derived PWV for rainfall monitoring. Moreover, the significantly reduced latency demonstrates a promising perspective of real-time GPS PPP-based PWV estimation as an enhancement to existing forecasting systems for rainfall forecasting.

ITG: A New Global GNSS Tropospheric Correction Model

Tropospheric correction models are receiving increasing attentions, as they play a crucial role in Global Navigation Satellite System (GNSS). Most commonly used models to date include the GPT2 series and the TropGrid2. In this study, we analyzed the advantages and disadvantages of existing models and developed a new model called the Improved Tropospheric Grid (ITG). ITG considers annual, semi-annual and diurnal variations, and includes multiple tropospheric parameters. The amplitude and initial phase of diurnal variation are estimated as a periodic function. ITG provides temperature, pressure, the weighted mean temperature (Tm) and Zenith Wet Delay (ZWD). We conducted a performance comparison among the proposed ITG model and previous ones, in terms of meteorological measurements from 698 observation stations, Zenith Total Delay (ZTD) products from 280 International GNSS Service (IGS) station and Tm from Global Geodetic Observing System (GGOS) products. Results indicate that ITG offers the best performance on the whole.

An Optimal Weighting Method of Global Positioning System (GPS) Troposphere Tomography

The functional model of Global Positioning System (GPS) troposphere tomography consists of three types of equations including the observation equation, the horizontal constraint equation, and the vertical constraint equation. The prerequisite for ensuring the accuracy of troposphere tomography modeling is to determine the optimal weights for the three types of equations. In order to reasonably determine the weights among these equations, this paper proposes an optimal weighting method. Compared to the conventional equal weighting scheme and constant weighting scheme, the method proposed in this paper can adaptively adjust the weights for various equations and enable the posterior unit weight variances for the three types of equations that achieve statistically equal. Numerical results under various weather conditions showed that the proposed method can significantly improve the accuracy of GPS tomography modeling with the GPS PPP-estimated slant tropospheric delay as reference when compared to the other two conventional methods.

Analysis of GPS satellite clock prediction performance with different update intervals and application to real-time PPP

The GPS satellite clock offset prediction is investigated and applied to a real-time PPP system. First, the current situation of GPS satellite clock is introduced and analysed with respect to their stability. Then the satellite clock prediction with different update intervals is presented, in which the satellite clock day boundary jump is addressed. Afterwards, the investigation of the satellite clock prediction model for GPS satellite IIF clocks is carried out and the effects of periodic terms are discussed. After that, the verification of the satellite clock offset prediction will be carried out both in the time and positioning domain. Positioning accuracy at 0.021, 0.049, and 0.017 m in the east, north, and vertical directions can be obtained for 6-h static positioning using the predicted clock offset updating every hour, while the 3D RMS for kinematic real-time PPP is 0.360 m, with 28% improvement over that utilising the IGU predicted products.

A global empirical model for mapping zenith wet delays onto precipitable water vapor using GGOS Atmosphere data

The importance of water vapor in research of global climate change and weather forecast cannot be over emphasized; therefore substantial efforts have been made in exploring the optimal methods to measure water vapor. It is well-established that with a conversion factor, zenith wet delays can be mapped onto precipitable water vapor (PWV). However, the determination of the exact conversion factor depends heavily on the accurate calculation of a key variable, weighted mean temperature of the troposphere (Tm). As a critical parameter in Global Positioning System (GPS) meteorology, Tm has recently been modeled into a global grid known as GWMT. The GWMT model only requires the location and the day of year to calculate Tm. Despite the advantages that the GWMT model offers, anomalies still exist in oceanic areas due to low sampling resolution. In this study, we refine the GWMT model by incorporating the global Tm grid from Global Geodetic Observing System (GGOS) and obtain an improved model, GWMT-G. The results indicate that the GWMT-G model successfully addresses the anomaly in oceanic areas in the GWMT model and significantly improves the accuracy of Tm in other regions.

A global empirical model for estimating zenith tropospheric delay

Tropospheric delay acts as a systematic error source in the Global Navigation Satellite Systems (GNSS) positioning. Empirical models UNB3, UNB3m, UNB4 and EGNOS have been developed for use in Satellite-Based Augmentation Systems (SBAS). Model performance, however, is limited due to the low spatial resolution of the look-up tables for meteorological parameters. A new design has been established in this study for improving performance of the tropospheric delay model by more effectively eliminating the error produced by tropospheric delay. The spatiotemporal characteristics of the Zenith Tropospheric Delay (ZTD) were analyzed with findings that ZTD exhibits different annual variations at different locations and decreases exponentially with height increasing. Spherical harmonics are utilized based on the findings to fit the annual mean and amplitude of the ZTD on a global scale and the exponential function is utilized for height corrections, yielding the ZTrop model. On a global scale, the ZTrop features an average deviation of -1.0 cm and Root Mean Square (RMS) of 4.7 cm compared with the International GNSS Service (IGS) ZTD products, an average deviation of 0.0 cm and RMS of 4.5 cm compared with the Global Geodetic Observing System (GGOS) ZTD data, and an average deviation of -1.3 cm and RMS of 5.2 cm compared with the ZTD data from the Constellation Observing System of Meteorology, Ionosphere, and Climate (COSMIC). The RMS of the ZTrop model is 14.5% smaller than that of UNB3, 6.0% smaller than that of UNB3m, 16% smaller than that of UNB4, 14.5% smaller than that of EGNOS and equivalent to the sophisticated GPT2+Saas model in comparison with the IGS ZTD products. The ZTrop, UNB3m and GPT2+Saas models are finally evaluated in GPS-based Precise Point Positioning (PPP), as the models act to aid in obtaining PPP position error less than 1.5 cm in north and east components and relative large error (>5 cm) in up component with respect to the random walk approach.

The Research on Four-Dimensional Water Vapor Tomography Based on Real-Time PPP Technique

With the development of International GNSS Service (IGS) real-time pilot project (RTPP) acquiring precipitable water vapor (PWV) with high accuracy has become a reality based on the real-time precise point pointing (RT-PPP) technique. The accuracy of zenith total delay (ZTD) and PWV derived from RT-PPP have been validated using observed global positioning system (GPS) data and meteorology data from Satellite Positioning Reference Station Network (SatRef) in 2014. The ZTD comparison with that from afterwards PPP and GAMIT software shows that the relative coefficients are 0.9786 and 0.9687, respectively. The PWV comparison with that from radiosonde shows that the relative coefficient and RMS are 0.9512 and 2.13 mm, respectively. It is a clear evidence that the RT-PPP technique has a similar accuracy with the result calculated using afterwards IGS products. However, PWV is mean of water vapor information of many GNSS signal rays during a period of time over the station, which cannot reflect the three-dimensional water vapor distribution. Slant water vapor (SWV) can be obtained by mapping PWV at different elevation and azimuth angles. The tomographic experiment has been performed using SWVs of twelve stations from SatRef as tomographic observation and compared with result from radiosonde. The comparison shows a good agreement and the RMS, SD, Bias, and MAE of integrated water vapor (IWV) are 3.60, 2.78, 2.29, and 2.92 mm, respectively, the root mean square (RMS), standard deviation (SD), Bias, and mean absolute error (MAE) of calculated water vapor density are 1.08, 1.03, −0.21, and 0.77 g/m3, respectively. The above result makes it possible that acquiring the real-time three-dimensional water vapor distribution using tomography approach with SWVs derived from RT-PPP technique, which has an important influence on short-term disastrous weather and now-casting precipitation forecasting.