University of Edinburgh logo British Antarctic Survey logo National Oceanographic Centre logo University of Leeds logo


Development and exploitation of a high-resolution sea level product in the coastal ocean

Correlation maps from satellite altimetry and ocean models, showing strong correlation along coastlines
Correlation maps from satellite altimetry and ocean models, showing strong correlation along coastlines

Scientific background and motivation
As oceans respond to local, regional and global changes a key parameter to monitor is sea level particularly
along the coasts, as this is where changes impact on populations and a better understanding of measurable
values is required. Currently, tide gauges provide point measurements with high temporal resolution whereas
satellite altimeters can provide synoptic information but with poor temporal resolution. The quality of
measuring sea level from satellite altimeters has improved markedly over recent years in particular in the
critical zone close to the coast. For example, through coastal retracking ( or the
introduction of SAR altimetry (CryoSat-2, Sentinel-3 and Sentinel-6). The studentship will include the
possibility of a placement at NASA JPL and/or participation in a research cruise.

Aims and objectives
In order to address some key issues in the coastal zone it is proposed to develop a methodology for a gridded
coastal sea level product (for 2011 onwards). By building a statistical model linking sea level data (altimetry
and tide gauges) with satellite sea surface temperature and/or ocean colour products as a proxy for changes
in sea level the student will be able to create a high-resolution product for scientific analyses. Associated
case studies could be global or regional in nature, with a number of key areas of interest already identified
(e.g. Severn Estuary).

Conventional geostatistical methods will face several challenges when applied to reconstructing coastal sea
level. One challenge is that the correlation length scales of sea level are much longer along the coast than
across the shelf and so isotropic covariance functions, which are predominantly used in geostatistics, will not
work well. New geostatistical models based on Gaussian processes with local anisotropy will be investigated to
address this problem. The approach would incorporate multiple satellite altimeters (with different sampling
so simple methods result in aliasing), tide gauges and other satellite/in situ/model parameters (e.g. surface
temperature and/or ocean colour) via some form of spatio-temporal model. There is the potential in the
latter part of the studentship for comparison of results with results from the innovative SWOT mission that
will provide the same information we are looking to produce.

Data sources
The key data sources to be exploited in this PhD are:
• satellite altimetry sea level data (e.g. CryoSat-2, Jason series, Sentinel 3, Sentinel 6 and possibly
• satellite sea surface temperature and ocean colour
• tide gauge data