Wide Area Neutral Atmosphere Models for GNSS Applications

In this work we discuss developments related to neutral atmosphere delay prediction models at UNB. We are introducing a new model, which was designed to provide better predictions for different regions inside a delimited wide area. The goal of this new development is to have a more reliable model for wide area augmentation system users, with some homogeneity in terms of performance over the area of interest. The approach to create the new wide area neutral atmosphere model for North America (UNBw.na) is comprehensively described and discussed. All result analyses took into consideration the most recent version of UNB models, until now, UNB3m. Results for meteorological parameters prediction showed that the new grid-based model could perform better than a latitude (only) based model (such as UNB3m). The general results do not show a spectacular improvement for the new model, however it is consistently better than its predecessor, and, the improvement for certain regions is more significant than others. Regions where the performance of the old model was not satisfactory had results significantly improved with the new model. A validation of UNBw.na predicted zenith delays was realized using radiosonde-derived delays as reference. This analysis showed that different regions of the continent manifested improvement for the estimations with the new model. Investigation of the performance of both models (UNBw.na and UNB3m) with radiosonde ray-raced delays at a few sample stations showed that UNBw.na generally has a better fit to the yearly behavior of the zenith delays. It was also possible to notice that results from UNBw.na are more consistent between stations at different locations than when using UNB3m. UNBw.na was shown to be consistently better than UNB3m in several aspects, and the adopted procedure for the grid calibration works in an adequate way, resulting in a reliable model. INTRODUCTION Mitigating the neutral atmosphere refraction is a crucial step in GNSS positioning. Also often called tropospheric delays, the neutral atmosphere delays are one of the main sources of measurement errors in GNSS. One usual way to account for these effects is using prediction models. There are also other alternatives for neutral atmosphere delay mitigation, such as the parameterization of the zenith delay in the positioning model, when dual frequency carrier-phase measurements are available. However, even in this case, the parameter is commonly a