New methods and applications for interferometric GNSS reflectometry
The GNSS reflectometry technique has been proven to be usable for measuring several environmental properties, such as soil moisture, snow depth, vegetation, and sea level. As numerous GNSS installations are already installed around the world for geodetic purposes, the technique opens up a large data set for new analyses, complementing other environmental measurement campaigns. However, a main drawback of the technique is that its precision generally is worse than more specialised equipment, and while this is in part compensated for its low cost and maintenance requirements, improved precision is still a main goal of research in the field of GNSS reflectometry.
The first topic of this thesis concerns the development of new methods for analysing GNSS-R data to retrieve precise measurements, especially in the case of sea level.
As GNSS-R measurements are usually done over time spans of around half an hour, the dynamic sea surface has proven to be a challenge to measure. However, using inverse modelling with least squares adjustment, we prove that we can significantly improve the retrieval precision. Developing on the inverse modelling approach, we also prove that high-precision real-time GNSS reflectometry is also feasible using Kalman filtering.
The other main topic of this thesis is finding new applications for the GNSS-R technique. Firstly, we show that when a GNSS-R installation is mounted close to a body of water, it is possible to determine whether the surface is frozen or not. Secondly, while GNSS reflectometry is traditionally performed with high-precision geodetic instruments, we show that everyday devices, such as a mobile phone, can be used instead. We find that the precision of the mobile devices is on a similar level as for geodetic equipment.
Finally, this thesis explores and highlights one of the challenges that are still left in GNSS-R research: absolute referencing of sea level measurements. Past research has mostly focused on precision, leaving out accuracy, and we show that there are unknown effects that cause an offset between GNSS-R measurements and co-located tide gauges.