Impact Of Weather Model Error On GPS Positioning Results

The troposphere in GPS (Global Positioning System) refers to from the ground surface upward to about 50-60 kilometers altitude. When transmitting in the troposphere, GPS signals will suffer from transmission delays which contain speed delays and inflecting delays on account of tropospheric refractions. Tropospheric delay can be weakened to some extent when a model with additional parameter estimation algorithm is used. However, on one hand the meteorological conditions along the signals propagation path cannot be known; on the other hand the spatial and temporal distributions of water vapor are very irregular that result in wet delay in random variations. So the tropospheric delay error is still one of the major error sources to limit further improvement of GPS positioning accuracy currently.In high-precision GPS applications, corrections to tropospheric delay depend on the three meteorological elements (pressure, temperature, and humidity) of a station and mapping functions used to map the priori zenith hydrostatic and wet delay to any elevation angle. Thus it has been an important issue to gain accurate station pressure and temperature and perfect the mapping functions to improve GPS positioning precision in the field of GPS technology and application. So far significant progresses have been made, for example, the global meteorological model GPT2 (Global Pressure and Temperature, version 2) which is built on the latest ECMWF (European Centre for Medium-Range Weather Forecasts) Re-Analysis (ERA-Interim) data has the better spatial and temporal variability. Nevertheless, all models have errors, perceiving the influence rules of station meteorological elements and mapping functions on GPS positioning results are significant prerequisites to evaluate GPS positioning precision and apply positioning results correctly.By analyzing the data processing results based on the GPT2 model, comparing and studying the influence of station pressure and temperature on GPS positioning results, this thesis obtains some cognition as follows: 1. The deviations of station pressure result in the deviations of coordinate estimates. In high-precision GPS positioning, although corrections to tropospheric delay usually introduce several zenith delay parameters on the premise of using a model to obtain prior zenith delay, the deviations of priori zenith delay still cause the deviations of coordinate estimates, especially for the vertical component, and the deviations depend on station latitudes and elevation-dependent data weighting used in the analysis. The reason is that the deviations of priori zenith delay which belongs to hydrostatic delay stem mainly from the deviations of station pressure. However, zenith delay parameters are applied mainly for wet delay caused by water vapor, while the hydrostatic mapping function and the wet mapping function have obvious differences at low elevation angles, and the amounts of low angle data observations vary with station latitudes.2. The value of station temperature is one of the reasons that affect the positioning precision. Generally speaking, compared with station pressure, temperature has relatively weak influence on priori zenith delay. But temperature varies disorderly with time, thus zenith delay parameters are able to absorb the deviations of delay caused by the deviations of temperature but are difficult to fully absorb the deviations of delay caused by the disorderly variations of temperature in observation session, and the remaining deviations of delay are in fact equal to casual observation errors and thus affect the positioning precision. As zenith wet delay shows approximately exponential growth with the increase of temperature, positioning results of higher precision will be more easily obtained in seasons or regions of lower temperature compared with higher temperature. This cognition is verified by analyzing the repeatabilities of 4-day observations of 1700 regional stations in the year 2013, and explains the seasonal variations of normalized root-mean-square values of daily solution of continuous observational stations.3. The deviations of pressure of the standard meteorological model and GPT model relative to GPT2 model are one of the sources that lead to aliased seasonal signals. GPT2 is a combined model of GPT and GMF. Compared with GPT/GMF, GPT2 has better spatial and temporal variability. This model is confirmed to bear better’ consistency with meteorological observations by comparing to pressure and temperature observations of stations in mainland China tectonic environment monitoring network and can well simulate the variations of pressure and temperature with time. Therefore GPT2 helps to obtain more accurate and more reliable positioning results in GPS data processing. At the same time, if the pressure values of GPT2 model are used as a reference, the deviations of the pressure values of the standard meteorological model and GPT model exhibit seasonal variations, so using any one of the two models will cause the variations of the coordinate estimates with seasons. This suggests that the previous seasonal variations of the coordinate estimates are not completely real crustal non-tectonic movements, instead the seasonal variations of deviations of station pressure can also induce aliased annual and semi-annual signals.4. The improvements of the mapping function and the temperature gradients of the GPT2 model is very limited to enhance the precision of the positioning results. The GPT2 model also makes some improvements to GMF and introduces temperature gradients which vary spatially and temporally for the first time. However, the difference between the GPT2 and GMF model is tiny and its impact on the positioning results can be ignored. Compared with the temperature gradients provided by the GPT2 model, the invariant temperature gradient (6.5°C/km) can lead to the deviations of station temperature and zenith wet delay and the deviations vary spatially and temporally. As the deviations in zenith wet delay can be absorbed well by the parametes used to correct the the zenith wet delay, the improvements contributed by the refined temperature gradients to positioning accuracy are limited.