Strategies for long-term monitoring of tide gauges using GPS

Teferle, Felix Norman (2003) Strategies for long-term monitoring of tide gauges using GPS. PhD thesis, University of Nottingham.

[img]
Preview
PDF
21Mb

Abstract

Changes in mean sea (MSL) level recorded relative to tide gauge benchmarks (TGBM) are corrupted by vertical land movements. Accurate estimates of changes in absolute sea level, require these MSL records to be corrected for ground level changes at tide gauge sites. For more than a decade, the Global Positioning System (GPS) has been used to determine positions of TGBMs and to monitor their position changes, i.e. station velocities, over time in the International Terrestrial Reference System (ITRS). This was initially carried out by episodic GPS campaigns and later on by continuous GPS (CGPS) or a combination of both.

Highly accurate realizations of the ITRS, satellite orbits and models for the mitigation of systematic effects currently enable the determination of station positions using GPS at the centimetre or even millimetre level. It is however argued that accurate long--term estimates of changes in the vertical component at the 1mm/yr level cannot be achieved, making intercomparisons between GPS estimates and other techniques necessary.

Daily processing and analysis of continuous GPS networks requires automated procedures. The modifications and improvements to the existing procedures at the IESSG are described. The newly developed tools include the monitoring and quality control of daily archived GPS observations and of processing results. A special focus is on the coordinate time series analysis and methodologies used to obtain the best possible estimates of vertical station velocities and associated uncertainties.

The coordinate time series of 21 CGPS stations in the UK and France are analysed. Eight of these stations are co-located with tide gauges. The effects of two processing strategies and two realizations of the ITRS on the coordinate time series are investigated. Filtered coordinate time series are obtained by application of a regional filtering technique. Station velocity estimates are obtained by fitting a model including a linear and annual term, and offsets to the unfiltered and filtered coordinate time series. Realistic uncertainties for these velocities are obtained from the application of two empirical methods which account for coloured noise in the coordinate time series. Results from these are compared to the Maximum Likelihood Estimation (MLE), which allows for more rigorous and accurate, simultaneous estimation of the model parameters and their uncertainties. Strategies for coordinate time series analysis on a daily or monthly, and annual or bi-annual basis are defined.

At two CGPS stations the dual-CGPS station concept is tested and compared to the single baseline analysis and the application of an adaptive filter. An empirical method to obtain coordinate time series specific filter parameters is described. This investigation shows that reliable relative vertical station velocity estimates can be obtained after much shorter observation spans than absolute vertical station velocity estimates. The availability of dual-CGPS station pairs allows a simplified processing strategy and a multitude of coordinate time series analysis methods, all contributing to a better understanding of the variations in the positions of CGPS stations.

Vertical station velocity estimates for the unfiltered and filtered coordinate time series and different analysis strategies are compared for 17 of the CGPS stations and show disagreements of up to 2mm/yr. At the eight CGPS stations co-located with or close to tide gauges alternative estimates of vertical land/crustal movements from absolute gravimetry, geological information and glacial isostatic adjustment models are compared to the GPS estimates, and it is suggested that the latter are systematically offset. An alignment procedure is demonstrated, correcting the vertical station velocity estimates of all 17 CGPS stations for this offset. The correlation of the geology-aligned vertical station velocity estimates and the MSL records from eight tide gauges suggests changes in absolute sea level of approximately +1mm/yr around the UK.

Item Type:Thesis (PhD)
Supervisors:Dodson, A.H.
Bingley, R.M.
Uncontrolled Keywords:sea level change, climate change, British Isles, Global Positioning System, tide gauge, vertical land movements, time series analysis, noise, error analysis, station velocity
Faculties/Schools:UK Campuses > Faculty of Engineering > School of Civil Engineering
ID Code:27
Deposited By:Felix Norman Teferle
Deposited On:09 Sep 2003
Last Modified:06 Feb 2009 14:43

Repository Staff Only: item control page