Development of a Seasonal Adjustment Technique STEL and its Application to Daily GNSS Data

Abstract

　Seasonal oscillations, caused by both tectonic and non-tectonic reasons, are one of the major components recorded in the Global Navigation Satellite System （GNSS） position time series. A functional model fitting has been widely used to extract seasonal oscillations, but this method requires assuming the model representing a whole time series. In this study, we develop a seasonal adjustment method based on an existing alternative approach that does not rely on a functional model assumption. In the proposed method, an estimation of an episodic component is added to the original method to express abrupt changes in a GNSS time series due to coseismic and postseismic deformations. In a synthetic test, our method successfully extracts a seasonal component from a time series with episodic changes. In addition, we apply the method to an observed GNSS position time series with a large seasonal variation. The regressed line explains the time series better than a functional model fitting. The source code of our method is now available online. We expect it will be utilized in future studies dealing with various types of datasets and crustal deformation.