Thesis title: Variometric Approach for Real-Time Estimation of Tropospheric Delay: Toward Enhanced Monitoring of Extreme Rainfall Events
As climate change accelerates, extreme rainfall events are becoming more frequent and intense, emphasizing the need for accurate and timely monitoring of atmospheric water vapor. Traditional ground-based observation systems, such as radiosondes and microwave radiometers, provide high-accuracy measurements but are limited by sparse spatial and temporal coverage and high operational costs, making them unsuitable for continuous, wide-area monitoring. In recent decades, GNSS meteorology has emerged as a promising alternative, exploiting the propagation delays introduced by atmospheric water vapor to derive continuous and spatially dense estimates of tropospheric water vapor content.
The most widely adopted technique in this field is Precise Point Positioning (PPP), which enables high-precision estimation of station coordinates and Zenith Tropospheric Delays (ZTDs) from single-receiver observations. However, real-time PPP performance depends on the continuous availability of ultra-rapid satellite orbit and clock products, and current implementations provide only zenith delays. As a result, PPP cannot retrieve slant-path tropospheric delays along individual satellite lines of sight, limiting its ability to resolve horizontal gradients and characterize anisotropic atmospheric structures typically associated with localized and intense precipitation events.
To address these limitations, this work develops a variometric algorithm for complementary estimation of tropospheric delays. The method relies solely on real-time available GNSS products, such as broadcast orbits and clocks, and retrieves delay variations from carrier-phase differences between consecutive epochs. The algorithm was implemented and evaluated in post-processing while strictly using only real-time data and products. This configuration enables the estimation of ZTDs with high temporal resolution while maintaining full autonomy from external precise correction streams.
Validation was performed using both simulated and real GNSS datasets. The simulation campaign confirmed the correct implementation of the algorithmic components and demonstrated sensitivity to directional tropospheric effects through a test in which an anisotropic perturbation was imposed on a single satellite. The algorithm successfully detected the corresponding line-of-sight variation. The evaluation with real data further showed that the variometric reconstruction consistently remained within the PPP uncertainty band. A comparison with ZTD values retrieved from a ground-based microwave radiometer displayed a consistent temporal evolution, demonstrating agreement between the two independent techniques.
The proposed approach advances GNSS-based atmospheric sensing by enabling robust, autonomous, and high-frequency monitoring of tropospheric water vapor. This work lays the foundation for future developments toward real-time slant-delay estimation and improved characterization of tropospheric anisotropy, with direct relevance for nowcasting and severe weather monitoring.