A new software package development for the analysis of GNSS time series based on the non-classical theory of measurement errors
Students Name: Kerker Vladyslav Vitaliiovych
Qualification Level: magister
Speciality: Space Geodesy
Institute: Institute of Geodesy
Mode of Study: full
Academic Year: 2023-2024 н.р.
Language of Defence: англійська
Abstract: Time-series data derived from GNSS coordinates exhibit signals arising from various sources, including the slow movement of tectonic plates due to geological processes, deformations in the Earth’s surface, and errors manifesting at different temporal scales. The methods employed for visual interpretation, preprocessing, and statistical analysis of these time-series data depend on the nature of the signals and other factors influencing the dataset’s evolution. Traditionally, these approaches do not delve into the intrinsic nature of the residual errors. Instead, they assume these errors to be random and adhering to the classical normal distribution. It is essential to recognize that these residual errors primarily stem from the specific metrological conditions surrounding individual GNSS observation stations. To address this issue, one effective method for analyzing time-series coordinates with residual systematic errors is through mathematical modeling. This modeling approach often leverages the Pearson-Jeffreys error law, also known as the non-classical modeling procedure. This technique combines probabilistic and mathematical-statistical approaches to assess whether the statistical distribution of residual errors deviates from the classical Gaussian distribution. The parameters within this approach serve as critical metrological indicators of the observation method. For the analysis of residual errors at a particular GNSS station, a novel software package, termed PS-NETM, has been developed. This software, programmed in Python, is specifically designed for scrutinizing the residual errors in processed GNSS coordinate time series spanning a year and a half or longer. Its primary purpose is to ascertain whether residual errors adhere to the normal Gaussian distribution. In doing so, it can reveal subtle and previously unaccounted sources of systematic errors. In this work, we present the outcomes of our analysis of processed GNSS coordinate time series data from several permanent stations within the IGS/EPN network. This demonstration showcases the capabilities and performance of the PS-NETM software in assessing the nature of residual errors and the impact of unaccounted sources on the GNSS data quality.