Vertical land motion as a key to understanding sea level change and variability

Abstract

Vertical land motions are a key element in understanding how sea levels have changed over the past century and how future sea levels may impact coastal areas. Ideally, to be useful in long-term sea level studies, vertical land motion should be determined with standard errors that are 1 order of magnitude lower than the contemporary climate signals of 1 to 3 mm/yr observed on average in sea level records, either using tide gauges or satellites. This metrological requirement constitutes a challenge in geodesy. Here we review the most successful instrumental methods that have been used to determine vertical displacements at the Earth's surface, so that the objectives of understanding and anticipating sea levels can be addressed adequately in terms of accuracy. In this respect, the required level of uncertainty is examined in two case studies (global and local). A special focus is given to the use of the Global Positioning System (GPS) and to the combination of satellite radar altimetry with tide gauge data. We update previous data analyses and assess the quality of global satellite altimetry products available to the users for coastal applications. Despite recent advances, a near-plateau level of accuracy has been reached. The major limitation is the realization of the terrestrial reference frame, whose physical parameters, the origin and the scale factor, are beyond the scope of a unique technique such as the GPS. Additional practical but nonetheless important issues are associated with the installation of GPS antennas, such as ensuring that there is no unknown differential vertical motion with the tide gauge.

A 10-Year Comparison of Water Levels Measured with a Geodetic GPS Receiver versus a Conventional Tide Gauge

“…Moreover, the difficult task of tying the sea level measurements into a well-defined terrestrial reference frame becomes automatic. Indeed, for studies of global mean sea level, the problem of vertical land motion at tide gauges is a critical one (e.g., Wöppelmann and Marcos 2016). This has motivated an international campaign to deploy GPS receivers (or similar geodetic instrumentation) at a large global network of tide gauges (Schöne et al 2009).…”

“…Inadequately known land motion is a problem that routinely plagues studies of long-term trends in mean sea level (e.g., Wöppelmann and Marcos 2016), and addressing that problem is automatically an integral part of the system. In addition, GPS reflections require none of the traditional infrastructure like stilling wells, which are susceptible to storm damage and biological fouling and need regular maintenance (modern tide gauges based on microwave radar sensors also dispense with stilling wells; e.g., see The data analyzed here were collected at Friday Harbor (48.5468N, 123.0138W), located about 130 km northwest of Seattle, Washington, on San Juan Island, which sits between the Strait of Juan de Fuca and the Strait of Georgia.…”

Reassessment of 20th century global mean sea level rise

“…Stations for which a continuous GPS station is available are adjusted using the rates and uncertainties provided by Universite de La Rochelle (ULR)6a (note that this is an update of ref. 16, which used ULR5). If GPS is not available at a particular station, VLM is alternatively determined by differencing altimetry and tide gauge time series for their common period.…”

“…If GPS is not available at a particular station, VLM is alternatively determined by differencing altimetry and tide gauge time series for their common period. Uncertainties are computed considering the noise content in the differenced time series as a combination of white noise and power law noise of an a priori unknown spectral index (16). The accuracy of the VLM correction is used to derive 11 different subsets of tide gauges; namely, only those for which VLM is known with an uncertainty smaller than 0.5 mm·y −1 (228 stations), 0.6 mm·y −1 (283 stations), or 1.5 mm·y −1 (448 stations).…”

“…Our tide gauge selection is based on the data set described in ref. 16, consisting of 322 stations (Fig. S1A), for which VLM corrections with uncertainties of less than 0.7 mm·y −1 are available (see Materials and Methods and Fig.…”