Estimation of offsets in GPS time-series and application to the detection of earthquake deformation in the far-field

Montillet, J.-P.; Williams, S.D.P. ORCID:; Koulali, A.; McClusky, S.C.. 2015 Estimation of offsets in GPS time-series and application to the detection of earthquake deformation in the far-field. Geophysical Journal International, 200 (2). 1205-1219.

Before downloading, please read NORA policies.
Text (.)
This article has been accepted for publication in Geophysical Journal International ©: 2015, The Authors. Published by Oxford University Press on behalf of the Royal Astronomical Society. All rights reserved.
Geophys. J. Int.-2015-Montillet-1205-19.pdf - Published Version

Download (1MB) | Preview


Extracting geophysical signals from Global Positioning System (GPS) coordinate time-series is a well-established practice that has led to great insights into how the Earth deforms. Often small discontinuities are found in such time-series and are traceable to either broad-scale deformation (i.e. earthquakes) or discontinuities due to equipment changes and/or failures. Estimating these offsets accurately enables the identification of coseismic deformation estimates in the former case, and the removal of unwanted signals in the latter case which then allows tectonic rates to be estimated more accurately. We develop a method to estimate accurately discontinuities in time series of GPS positions at specified epochs, based on a so-called ‘offset series’. The offset series are obtained by varying the amount of GPS data before and after an event while estimating the offset. Two methods, a mean and a weighted mean method, are then investigated to produce the estimated discontinuity from the offset series. The mean method estimates coseismic offsets without making assumptions about geophysical processes that may be present in the data (i.e. tectonic rate, seasonal variations), whereas the weighted mean method includes estimating coseismic offsets with a model of these processes. We investigate which approach is the most appropriate given certain lengths of available data and noise within the time-series themselves. For the Sumatra–Andaman event, with 4.5 yr of pre-event data, we show that between 2 and 3 yr of post-event data are required to produce accurate offset estimates with the weighted mean method. With less data, the mean method should be used, but the uncertainties of the estimated discontinuity are larger.

Item Type: Publication - Article
Digital Object Identifier (DOI):
ISSN: 0956-540X
Additional Keywords: Time-series analysis, Satellite geodesy, Plate motions, Earthquake source observations
NORA Subject Terms: Marine Sciences
Date made live: 27 Mar 2015 11:03 +0 (UTC)

Actions (login required)

View Item View Item

Document Downloads

Downloads for past 30 days

Downloads per month over past year

More statistics for this item...