GNSS Coordinates Time Series Assessment using Functional and Stochastical Models.

Abstract:

Global Navigation Satellite System (GNSS) has been largely used by the civil community, notably by Geodesists. The position and velocity of terrestrial stations have been estimated using GNSS data series based on daily solutions. So, currently it is possible to analyse the GNSS coordinates time series aiming to improve the functional and stochastic models what can help better understand geodynamical phenomenon. Several sources of errors are mathematically modelled or estimated in the GNSS data processing to obtain accurate coordinates what in general is carried out by using scientific software. However, due to impossibility to model all errors some kind of noises can remain contaminating the coordinate time series, especially those related with seasonal effects. The noise affecting GNSS coordinate time series can be composed by white and coloured noises what can be characterized from Variance Component Estimation technique through Least Squares Method. Experiments were carried out by using GNSS time series for few Brazilian stations considering almost ten years of daily solutions. The results showed that the adoption of an adequate stochastic model considering the noises variances of time series can produce more realistic and reliable functional model for GNSS coordinate time series. Applications that will be based on such serie of coordinates will have more reliable assessments.