Evaluating stochastic models for estimating site velocity from daily and weekly GNSS time series in the stable region of the South American plate

Abstract Accurate estimation of GNSS station velocities requires a precise characterization of stochastic noise in coordinate time series. This study evaluates stochastic models for estimating site velocities using weekly network and daily PPP GNSS solutions from 74 stations in the stable South American mid-plate. The functional model incorporated seasonal components, while the stochastic model was based on noise variance estimates. The Non-Negative Least Squares Variance Component Estimation method was applied to estimate noise amplitudes, classifying noise as a combination of white and colored components. Additionally, the spectral index was refined using Maximum Likelihood Estimation. Results show that most time series are best described by a combination of white and flicker noise, with differences in spectral properties between weekly and daily solutions. The impact of these models on velocity uncertainties was assessed, showing that neglecting an appropriate noise model can lead to overestimated uncertainties. These findings contribute to improving GNSS-based velocity estimations for geodynamic studies.