GNSS position series contain signals that inform us about the deformation of the Earth’s surface, but also about systematic errors, in any reference frame and on very different time scales: from sub-day tidal deformations to interannual overload deformations or continuous deformations resulting from the movement of tectonic plates.
The SARI tool allows users to view the series in a web browser, remove discontinuities and outliers, interactively fit models and save results. A comprehensive list of functionalities has been implemented to assist the user in extracting the desired information. This list includes, among others, model estimation with time-varying parameters (Kalman filter), spectral and wavelet analysis, series smoothing (Vondrák filter) or correlated noise estimation.
The tool is developed under R and is designed to handle mainly GNSS position series, but it can be used to handle other types of series with or without regular sampling (e.g., tide gauge series or GNSS-R have been used).The advantage for GNSS series is that it is possible to directly display hardware changes recorded in “sitelog” (IGS), “station.info” (GAMIT), University of Nevada discontinuity file (GIPSY) or free format files (e.g., for earthquake dates). It is also possible to compare or correct the series with a model (e.g., overload, post-seismic, etc.) or with the series of a nearby GNSS station.
Alvaro Santamaría (Geosciences Environment Toulouse, OMP)
The tool is available online, with no prerequisites other than an Internet connection.
The tool includes an information tab with a detailed description of the features and how to use them. For more information on SARI and examples, see Alvaro Santamaría-Gómez.
Ref : A. Santamaría-Gómez (2019) SARI : interactive GNSS position time series analysis software. GPS Solutions, in press. DOI : 10.1007/s10291-019-0846-y
This article was published in the newsletter Résif n°16 – June 2019 (pdf in French)
Synthetic series with an annual cycle of variable amplitude superimposed on a variable trend and noise (black dots) adjusted with a Kalman filter (red curve) and smoothed with a Vondrák filter (blue curve). Arbitrary units.