Noise Analysis and Velocity Uncertainties of Multi-GNSS Time Series

Noise analysis is carried out for the stand-alone GPS, Glonass and GPS+Glonass daily position time series of continuous mode Global Navigation Satellite System (cGNSS) stations in the Indian subcontinent and adjoining regions during 2015-2021 to characterize the preferred noise model, amplitude and associated velocity uncertainties. Maximum Likelihood Estimation (MLE) and spectral analysis are used to perceive the preferred noise model which showed that colored noise is more dominant in North, East and Up components of all the multi-GNSS time series. Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and spectral indices also confirmed that color noise is the most suitable noise model for all the stations. The Power Spectral Density (PSD) of the residuals revealed flicker noise is the best-fit model compared to random walk noise. Noise amplitudes and velocity estimates with associated uncertainties are calculated by assuming the flicker noise as the preferred model. Understanding the types and sources of noise helps in building robust noise models that can be incorporated in post-processing software to estimate precise velocities with less uncertainties.