A New Methodology for Incorporating Tide Gauge Data in Sea Surface Topography Models

Abstract

As part of the Vertical Offshore Reference Frames (VORF) project sponsored by the U. K. Hydrographic Office, a new model for Sea Surface Topography (SST) around the British Isles has been developed. For offshore areas (greater than 30 km from the coast), this model is largely derived from satellite altimetry. However, its accuracy and level of detail have been enhanced in coastal areas by the inclusion of not only the 60 PSMSL tide gauges with long-term records around the coasts of the United Kingdom and Ireland but also some 385 gauges established at different epochs and for different observation spans by the U. K. Admiralty. All tide gauge data were brought into a common reference frame by a combination of datum models and direct GPS observations, but a more significant challenge was to bring all short-term sea level observations to an unbiased value at a common epoch. This was achieved through developing a spatial-temporal correlation model for the variations in mean sea level around the British Isles, which in turn meant that gauges with long-term observation spans could be used as control points to improve the accuracy of Admiralty gauges. It is demonstrated that the latter can contribute point observations of mean sea level (MSL) with a precision of 0.078 m. A combination of least squares collocation and interpolation was developed to merge the coastal point and offshore gridded data sets, with particular algorithms having to be developed for different configurations of coastal topology. The resulting model of sea surface topography is shown to present a smooth transition from inshore coastal areas to offshore zones. Further benefits of the techniques developed include an enhanced methodology for detecting datum discontinuities at permanent tide gauges.

Keywords:

• Sea surface topography
• tide gauge data
• mean sea level
• satellite altimetry

Acknowledgements

The authors acknowledge the contribution to this study by the United Kingdom Hydrographic Office, through funding and the provision of data. We are also grateful to the Ordnance Surveys of Great Britain, Ireland, and Northern Ireland for permission to use the OSGM02 and OSGM05 models, and to the Permanent Service for Mean Sea Level for provision of tide gauge data. Parallel computing facilities were provided by the UCL Computing Resource Allocation Group. The Institute of Engineering Surveying and Space Geodesy at Nottingham University facility ‘BIGF’ provided archive GPS data and precise ETRF89 heights of a number of PSMSL tide gauges. We are grateful to Casey Construction of Kirkwall for assistance rendered during observation campaigns. The survey authorities on Guernsey, Jersey, and Isle of Man provided invaluable datum information.