Monitoring Sea Level in the Coastal Zone with Satellite Altimetry and Tide Gauges
We examine the issue of sustained measurements of sea level in the coastal zone, first by summarizing the long-term observations from tide gauges, then showing how those are now complemented by improved satellite altimetry products in the coastal ocean. We present some of the progresses in coastal altimetry, both from dedicated reprocessing of the radar waveforms and from the development of improved corrections for the atmospheric effects. This trend towards better altimetric data at the coast comes also from technological innovations such as Ka-band altimetry and SAR altimetry, and we discuss the advantages deriving from the AltiKa Ka-band altimeter and the SIRAL altimeter on CryoSat-2 that can be operated in SAR mode. A case study along the UK coast demonstrates the good agreement between coastal altimetry and tide gauge observations, with root mean square differences as low as 4 cm at many stations, allowing the characterization of the annual cycle of sea level along the UK coasts. Finally, we examine the evolution of the sea level trend from the open to the coastal ocean along the western coast of Africa, comparing standard and coastally improved products. Different products give different sea level trend profiles, so the recommendation is that additional efforts are needed to study sea level trends in the coastal zone from past and present satellite altimeters. Further improvements are expected from more refined processing and screening of data, but in particular from the constant improvements in the geophysical corrections.
KeywordsSea level Coastal zone Radar altimetry Coastal altimetry Tide gauge
In the previous paper in this issue, Ablain et al. explain in detail the importance of sea level rise, a measure of the increase in ocean volume, as a clear indicator of climate change and one of its main effects. Satellite altimeter-derived sea level rise is now well quantified both as a global mean and in its geographical distribution (as, respectively, visible in Figs. 2, 5 of Ablain et al., 2016) thanks to multiple efforts that include the sea level climate change initiative (Ablain et al. 2015). Satellite-based measurements of sea level compare well with the measurements from the global tide gauge network (Fig. 8 of Ablain et al., 2016). The requirement prescribed by the Global Climate Observing System (GCOS) of an accuracy better than 0.3 mm/year in the altimeter-derived rate of global mean sea level rise is still not fully met; however, the improvements seen in the last few years make that target a realistic one. The future looks promising for the precise determination of global and regional sea level rise from the integration of altimetry and tide gauges.
All impacts of sea level rise on society and ecosystems are going to be suffered entirely at the coast. As an example, a recent study (Hauer et al. 2016) found that a rise of only 90 cm by 2100 would put 4.2 million people at risk of inundation in the coastal zone for the continental USA alone. When projected globally and considering highly vulnerable areas such as low-lying island and deltas, the number of people that would be flooded if not relocated quickly rises to reach the order of 100 million or more (Hinkel et al. 2014).
Many stretches of the world’s coast still do not possess in situ sea level measuring devices, and those stretches include many vulnerable regions in developing countries. Altimetry is at present the only way of obtaining measurements of sea level variations in those regions and can already offer 24 years of observations from the TOPEX/Jason-1/2/3 ‘reference’ series and from the ERS–Envisat–AltiKa series,1 so it will remain valuable in order to extend the sea level record back in time also when tide gauges will eventually start to be installed in those regions.
In this contribution, we first review the status of the measurements of sea level in the coastal zone from tide gauges (Sect. 2), and then in Sect. 3 we examine the same issue from the point of view of coastal altimetry, with an overview of the improvements in the retrieval of sea surface height from altimetry in the coastal zone which have been made possible by algorithmic improvements and development of better corrections and data editing. This section also presents the datasets available for altimetry and coastal altimetry, for the benefit of the potential users of those products. Some particularly promising prospects for the monitoring of sea level in the coastal zone come from the advent of new technologies, whose coastal performance is discussed in Sect. 4: these technologies are Ka-band altimetry from AltiKa, and SAR altimetry from CryoSat-2 (and now Sentinel-3). We then present two specific examples of coastal sea level monitoring: in Sect. 5, a case study showing how local sea level can be monitored with altimetry and tide gauges around the coast of the UK and, in Sect. 6, some results on the variations of sea level rise rate as a function of distance from shore along the West African coast. The concluding section translates the science reviewed and results presented into recommendations for refinements to the processing, which should lead to further progress of the field.
2 Monitoring Sea Level with Tide Gauges
Tide gauges have been used since ancient times to measure sea level changes at the coast. In Amsterdam, an historical record of observations of sea level changes using a pole provides evidence of sea level rise and variability since 1700 (Van Veen 1945). Several tide gauge records from locations in Europe, for instance Liverpool since 1768 (Woodworth 1999) and Stockholm since 1774 (Ekman 1988) greatly contribute to our understanding of sea level changes over the eighteenth century. Since the 1830s, automatic or self-registering tide gauges were developed. The first automatic tide gauges are often credited to those installed in the Thames Estuary, England (Matthäus 1972). The first automatic tide gauge outside Europe was installed in San Francisco, USA in 1851. By the end of the nineteenth century, automatic tide gauges had been installed at many ports (Baltic and Mediterranean Seas, the USA coast). A network global in scope was starting to appear. Sea level rise during the twentieth century is estimated using almost 1000 tide gauge locations (e.g., Douglas 1997; Jevrejeva et al. 2006, 2014; Church and White 2006, 2011; Hay et al. 2015), with techniques accounting for the fact that most of those tide gauges do not cover the entire twentieth century.
the geographical distribution of tide gauges is naturally confined to the continental margins and some ocean islands, which provides poor sampling of the ocean basins; in addition, most tide gauges are located in the Northern Hemisphere (Europe, Japan and the USA);
available tide gauge records do not all cover the same time period, and their number decreases rapidly as we go back in time, especially prior to 1960;
tide gauges are attached to the land, providing measurements relative to the Earth’s crust, which could move. Vertical land movement is the one of the main difficulties to interpret tide gauge measurements (Wöppelmann and Marcos 2016; Jevrejeva et al. 2014). Some extreme examples (Fig. 1) of the effect that vertical land motion can have on local sea level are Fort Phrachula, near Bangkok, where relative sea level has been rising at a rate of about 15 mm/year since the 1960s due to subsidence caused by increased groundwater extraction, or Stockholm where it is falling at a rate of 3.8 mm/year due to crustal uplift associated with glacial isostatic adjustment.
there is no common reference level for the individual tide gauge records, despite some clear recommendations for it (for instance Woodworth et al. 2013), and this creates a problem of stacking records together.
3 Monitoring Sea Level with Coastal Satellite Altimetry
Satellite altimetry has been one of the workhorses of open-ocean operational oceanography and global sea level monitoring, so efforts are naturally being made to use it also in the coastal zone. The main motivation is its spatial and temporal coverage: altimetry is global in space, covering even the most remote areas of the oceans (and the polar oceans with some satellites), and we have already 24 years of data from missions with accuracies of the order of just a few cm, starting with the ERS-1 launch in 1991.2 Moreover, in addition to sea level, it provides measurements of significant wave height (SWH) and wind.
CNES’ PISTACH (2007–2011) for Jason-2 and ESA’s COASTALT (2008–2012) for Envisat, both developing specialized waveform retracking (see 3.1) and corrections.
The X-TRACK initiative by the Centre of Topography of the Oceans and the Hydrosphere (CTOH) at LEGOS—Laboratoire d’Etudes en Géophysique et Océanographie Spatiales. X-TRACK products are derived for all precise altimeter missions by means of coastal-oriented screening of altimeter data and corrections (see Birol et al. 2016) and are used for the example on variability of sea level trends in the coastal zone in Sect. 6.
Available products for open-ocean and coastal altimetry as of October 2016
e1, tx, e2, en, j1, j2, c2
(LRM/PRLM), sa, h2
L2, L3, L4 also L4
Global + European regions + Arctic + SW Indian
Widely used reference dataset processed with standard techniques. Distribution of global, Mediterranean Sea, Black Sea products is migrating to CMEMS during 2016
e1, tx, e2, en, j1, j2, c2 (LRM/PRLM), sa (s3a to be added soon)
L3 for assimilation
Global + European regions
Marine environment monitoring service of the EC/ESA Copernicus programme, providing products and services for all marine applications
Experimental Jason-2 products for hydrology and coastal studies with specific processing. Will be discontinued at the end of 2016 in favour of PEACHI
sa, (j2 to be added soon)
Experimental SARAL/AltiKa products including dedicated retracking and corrections leading to more accurate products for coastal zones, hydrology and ice. From 2017 expected to generate also j2 products
tx, j1, j2, gfo, en
Specific processing using improved data screening and latest corrections available (see Sect. 6)
EUMETSAT, NOAA, TUDelft
gs, e1, tx, pn, e2, gfo, j1, n1, j2, c2, sa
Widely used dataset, mirrored by tens of sites worldwide, with continuously updated corrections, but no specific coastal processing
j2, n1, (j1, j3 to be added soon)
Global, <50 km from coast
Experimental products from the ALES processor included in SGDR-type files alongside the standard products and corrections
c2 (SAR only)
SAR mode regions
On-demand Processing service for the CryoSat-2 SAR mode data where the user can configure some processing parameters to meet specific requirements (for istance for the coastal zone)
Global products for CryoSat-2 from an ocean processor (output is in PLRM over the SAR mode regions)—but no specific coastal processing
3.1 Strategies for Improving the Coastal Altimetry Data
In the coastal zone, in addition to the refinement of the statistical techniques for screening and filtering of the various data and corrections (such as in X-TRACK), there are two complementary courses of actions for improving the quality of the retrieved data: (1) applying specialized retracking (i.e. improving the estimation of the range term in Eq. 1) and (2) applying improved corrections for the atmospheric, surface or geophysical effects (i.e. improving the corrections term).
Reflected radar pulses returning to the altimeter receiver are recorded against time in the so-called waveforms. These are sent to the receiving station on the ground and ‘retracked’, i.e. fitted with a waveform functional form (waveform model) to yield the fundamental measurements of range (from which sea level is measured), SWH and radar backscatter (in turn related to wind). The fitting is usually carried out via least squares or maximum likelihood algorithms. Over the open ocean, waveforms normally conform well to the Brown model (Brown 1977; Hayne 1980). In a band typically extending ~10 km from the coastline, a significant portion of the radar waveforms depart from the Brown model (this portion gets larger approaching the coast, as shown in Fig. 3 of Halimi et al. 2013), calling for modified retracking strategies. The factors that impact on the waveforms are not only the presence of land in the altimetric footprints, but also the occurrence of ‘bright targets’ in the footprint such as patches of very calm water in sheltered areas (Gómez-Enri et al. 2010). A summary of the various strategies proposed in recent years for coastal retracking is in Passaro et al. (2014); these strategies include the use of a modified functional form (as in Halimi et al. 2013), the pre-classification of waveforms and the retracking of ‘sub-waveforms’, i.e. a portion of the waveform unaffected (or less affected) by coastal artefacts (as in Yang et al. 2012).
The improvements in retracking have been accompanied by equally important improvements in some of the corrections that need to be applied to altimetry data to account for atmospheric path delays and other geophysical effects. The two major improvements are in the correction of the path delay due to tropospheric water vapour (‘wet tropospheric’ correction, see Obligis et al. 2011) and in the tide models that are needed for all those applications where the tidal component is not part of the observed signal and need to be removed (Ray et al. 2011). Here we will briefly summarize the main advances in the wet tropospheric correction, while for the improvement in tidal models we refer to the comprehensive review by Stammer et al. (2014).
The wet tropospheric correction is almost proportional to the integrated water vapour content of the atmosphere. Over the open ocean, it is either directly measured by a 2- or 3-channel passive microwave radiometer on board some altimeters, or can be estimated with good accuracy using meteorological models, which however lack spatial structure. On approaching the coastal zone, the radiometer-based measurements degrade rapidly when land enters the radiometer footprint, which has a 10–50 km diameter depending on the particular channel. Models, on the other hand, are still not particularly able to capture the shorter-scale changes in water vapour in the coastal zone and lack accuracy. The need for an improved correction has been apparent since the inception of coastal altimetry and several solutions have been proposed (Obligis et al. 2011). Notable contributions include the improved algorithm proposed by Brown (2010) and applied to the advanced microwave radiometer on the Jason-2 mission, with an estimated error less than 1.2 cm within 5 km from land. Another successful improvement is the GPD correction by Fernandes et al. (2015), built by combining passive microwave measurements from altimetric missions with path delays measured by a network of coastal GNSS stations, and being extended to include measurements from other imaging microwave radiometers. This has been applied globally to 8 missions in the ESA Sea Level CCI project and has yielded a significant impact on regional sea level trends with particular relevance to the coastal and polar regions, due to an efficient correction for land and ice contamination in the radiometer footprint (Fernandes et al. 2015).
4 The Potential of New Altimetric Technologies in the Coastal Zone
The launch of two satellites—CryoSat-2 and AltiKa with two important technological improvements, i.e. SAR mode altimetry and Ka-band altimetry— has opened new prospects for altimetry. In this section, we describe the potential of these two technologies for monitoring sea level in the coastal zone, where those two missions perform particularly well.
4.1 Ka-Band Altimetry: AltiKa
The SARAL/AltiKa French/Indian satellite altimetry mission was launched on 25 February 2013. The platform embarks a DORIS antenna and GPS receivers for precise orbit determination, a dual frequency radiometer for wet tropospheric path delay retrieval, and is the first mission to carry a Ka-band (36.5 GHz) altimeter providing data at the high posting rate of 40 Hz, corresponding to ~180 m along the ground track of the satellite. Compared to previous altimeters that are using the Ku-band at 13.6 GHz (and 20-Hz posting rate), SARAL/AltiKa is expected to provide better vertical resolution of the range thanks to a larger bandwidth, and improved horizontal resolution thanks to a narrower antenna beam (footprint diameter is only 8 km compared to 20 km on Jason-2), at the cost of higher sensitivity to rain events (Vincent et al. 2006; Steunou et al. 2015). The high precision measurements provided by the altimeter are very valuable for the characterization of coastal sea level and dynamics, which is one of the main scientific objectives of the SARAL/AltiKa mission (Verron et al. 2015). The current geophysical data record (GDR) products (GDR-T patch 2 version) are dedicated to open ocean, but the CNES PEACHI prototype (Valladeau et al. 2015) processes high rate data with up-to-date algorithms for different surfaces (coastal, ice, hydrology). These products are available through the ODES portal (http://odes.altimetry.cnes.fr).
These altimetry-based assessments of instrumental performances can be completed by comparisons to in situ measurements to demonstrate the coastal capabilities of the SARAL/AltiKa mission. Several studies have used SARAL/AltiKa data in coastal zones, and their results tend to confirm what instrumental quality assessment suggests. Troupin et al. (2015) compared ocean currents derived from SARAL/AltiKa altimetry, HF radar and glider measurements and found a good agreement between altimeter and glider currents, as close as 10 km from the coast. Similar results were also found by Pascual et al. (2015), with further improvements foreseen from dedicated near-coast instrumental algorithms and geophysical corrections. Birol and Niño (2015) compared Jason-2 and SARAL/AltiKa data in coastal areas of the north-west Mediterranean Sea. They got a much better sampling from SARAL/AltiKa (more data available), and comparisons with local tide gauges showed a better agreement than with Jason-2, both for correlation (0.7 vs. 0.54) and RMS error (3.3 vs. 4.2 cm). Hareef Baba Shaed et al. (2015) compared significant wave heights (SWH) from SARAL/AltiKa with wave buoy measurements along the coasts of India and found correlations ranging from 0.85 to 0.98 and RMS errors lower than 0.4 m.
All the above examples demonstrate the capabilities of Ka-band altimetry for the monitoring of coastal ocean dynamics. With a complete reprocessing foreseen in 2017, and with its long-term stability validated with comparison with other altimeters and tide gauges as is being done in the Sea Level CCI project, the 3+ years long SARAL/AltiKa record will provide even more valuable data for coastal studies, providing extended observations of sea level over the same 35-day repeat set of orbits sampled by ERS-1, ERS-2 and Envisat between 1992 and 2010.
4.2 SAR Mode Altimetry: CryoSat-2
ESA’s CryoSat-2 satellite was launched on 8 April 2010, with the primary mission role of monitoring the cryosphere by measuring variations in ice thickness, but has been proven of exceptional utility also for the monitoring of the oceans (see for instance Dibarboure et al. 2012). A technological innovation of CryoSat-2 Synthetic Aperture Interferometric Radar Altimeter (SIRAL) is the delay-Doppler mode of measurement (Raney 1998) which we will refer to as ‘SAR mode’ as it involves an unfocused along-track synthetic aperture radar (SAR) processing of the radar echoes. When in SAR mode, SIRAL exploits the Doppler information in the returned pulse to achieve a much finer resolution in the along-track direction (the width of the along-track SAR resolution cell is ~350 m), virtually independent of the sea state. The size of the altimeter footprint in the across-track direction is the same of a conventional altimeter, i.e. 2–20 km depending on the sea state. By averaging independent measurement from adjacent cells, the augmented along-track resolution can be traded off, all or in part, to achieve lower noise on the estimated parameters. Due to power and data downlink constraints, SIRAL on CryoSat-2 can only be operated in SAR mode (and in another experimental mode, SAR interferometry, of primary use over ice surfaces) over a small portion of the Earth’s surface, and is instead in conventional low-resolution mode over most of the surface. In practice, this means that SAR mode data are available over a number of ‘patches’, one of which covers the entire European coastal sector. The importance of CryoSat-2 is magnified by the fact that SIRAL is a precursor of the Synthetic aperture Radar ALtimeter (SRAL) on the Copernicus Sentinel-3 satellites due to provide systematic oceanographic observations for the next 20 years (Donlon et al. 2012). Sentinel-3A was launched on 16 February 2016 and is being operated in SAR mode over the entire global ocean.
Results such those presented are extremely encouraging in terms of demonstrating the low noise level in SAR altimetry data due to the excellent performance of the radar, but for the particular application to long-term monitoring of sea level what is paramount is the stability of the whole measurement system, including the corrections. As for AltiKa, this is currently being investigated in the ESA Sea Level CCI project.
5 A Case Study Around the Coast of the UK
As an example of how the local sea level can be monitored with altimetry and tide gauges, which also allows investigating the link between deep-ocean and coastal sea level variability, we present here the results of an analysis of sea level around the UK coastline, which was conducted within the framework of the Sea Level SpaceWatch project. This project was funded by the UK Space Agency within the Space for Smarter Government Programme to design and prototype an operational service delivering systematically updated sea level observations around the UK, from a combination of satellite altimeter observations and tide gauge measurements (Cotton 2016). The focus of the analysis presented here is on the annual cycle of sea level over the period 2002–2015, but first we also present a comparison between the altimetry and tide gauge observations on interannual timescales, as a form of validation.
Here we use along-track altimetry data from Jason-1 and Jason-2 as reprocessed by the coastally adapted ALES retracker (Passaro et al. 2014) described in Sect. 3.1 and covering the period 2002–2015. We use a total of 58 tide gauge records, of which 46 were obtained from the data archives of the British Oceanographic Data Centre (BODC), 11 from the UK Coastal Channel Observatory (CCO), and 1 from the Port of London Authority (PLA). The temporal resolution of the tide gauge data is 15 min for records stored at the BODC and 10 min for those stored at the CCO and PLA. For consistency with the satellite altimetry data, the atmospheric correction was applied to the tide gauge data. In particular, we used the dynamic atmospheric correction (DAC) provided by AVISO (ftp.aviso.altimetry.fr), which consists of the barotropic response of the ocean to wind forcing and atmospheric pressure as estimated by the Mog2D-G model for periods shorter than 20 days and the inverse barometer (IB) approximation for longer periods. The DAC data are provided in the form of 6-h sea level fields on a 1/4° × 1/4° regular grid covering the global oceans. The atmospheric correction at each tide gauge is taken from the nearest DAC grid point to the tide gauge.
6 Evolution of Sea Level Trend from the Open to the Coastal Ocean
the standard AVISO dataset (version 2014 of delayed-time data,) distributed by the Copernicus Marine Environment Monitoring Service (CMEMS). These are the ‘vfec’ data, i.e. validated, filtered, sub-sampled and corrected for long wavelength errors; the spatial sub-sampling of data results in a spatial resolution of 14 km.
two datasets provided by LEGOS/CTOH, in which along-track altimetric data have been reprocessed using the 2011 and 2016 versions of the X-TRACK algorithms, respectively (Birol et al. 2016). These are more adapted to coastal regions and provide data at a spatial resolution of 7 km. The editing and extrapolation of geophysical corrections has been improved in X-TRACK2016 compared to X-TRACK2011, on two aspects: (1) the removal of land contamination in the radiometer-derived wet tropospheric correction is now performed in X-TRACK2016 with an algorithm computing the proportion p of land within the radiometer footprint of each altimeter mission and rejecting the wet tropospheric correction values for which p > 0. Possible other outliers away from the coast are discarded by an algorithm that detects large differences between the correction values at two consecutive points along the track for each cycle. The resulting data gaps are either interpolated (away from the coast) or filled with the closest values qualified as valid (in the land/sea transition areas). A technique based on the discrete wavelet transform is then used to compute a cleaned and noise-free wet tropospheric correction; (2) the ionospheric correction filtering has been updated to more efficiently detect outliers: the filter is a median absolute deviation (MAD) threshold in X-TRACK2016, used instead of the 3σ-threshold filter used in X-TRACK 2011 (σ being the standard deviation). The sea-state bias (SSB) correction was smoothed in X-TRACK2011 using Bezier curves. In X-TRACK2016, the SSB correction is filtered in the along-track direction with a Loess low-pass filter, and missing values are replaced by the nearest interpolated data. Regarding the other altimeter corrections, usually derived from models, very few values are discarded by the editing process. We choose to replace flagged corrections by their nearest valid neighbours.
The changes in sea level trend from the open to the coastal ocean are shown in terms of percentage in Fig. 11 (left panels). In the AVISO dataset, the trend changes from the open to the coastal ocean mostly range from −20 to +20% for the different coastal sections, with a coastward decrease of the sea level trend equatorward of 10°S/10°N, and a coastward increase of the sea level trend poleward of 25°N/25°S. When sea level trend changes are bin-averaged across all the coastal sections as a function of the distance to the coast (Fig. 11, bottom-right panel), no robust evolution of the sea level trend from the open to the coastal ocean can be seen off Western Africa in the AVISO dataset.
In the X-TRACKv2011 dataset, a substantial increase in the sea level trend can be seen for most coastal sections (Fig. 11, middle-left panel). As more valid data are available in the X-TRACKv2011 dataset than in the AVISO dataset, the averaged coastward evolution of sea level trend can be studied closer to the coast (up to 15 km off the coast). On average across the coastal sections, the sea level trend over 1993–2012 steadily increases from the open to the coastal ocean off Western Africa in this dataset, reaching an increase of 25% 25–30 km off the coast (Fig. 11, middle-right panel). The increase weakens in the last few kilometres off the coast and is less robust across the different coastal sections (grey envelope in Fig. 11, middle-right panel).
Results from the X-TRACKv2016 dataset are in between these from the AVISO and X-TRACKv2011 datasets. The sea level trend increases coastward for coastal sections located south of 15°S and north of 25°N (a result qualitatively robust for the 3 datasets), but less so than in the X-TRACKv2011 dataset. Equatorward of 10°N/10°S, no robust coastward evolution of the sea level trend is seen across the different coastal sections. The greater number of valid data in the X-TRACKv2016 dataset allows quantifying the mean coastward evolution of sea level trend up to 5 km of the coast. On average, the sea level trend only slightly increases coastward in the X-TRACKv2016 dataset (by less than 10%), but this is not robust across the coastal sections (Fig. 11, upper-right panel, grey envelope).
These results show that efforts made to improve satellite nadir altimetry products in the coastal ocean allow recovering more data and obtaining more coherent long-term signals in the coastal zone. In particular, the editing and extrapolation of geophysical corrections have been updated and improved in X-TRACKv2016 compared to X-TRACKv2011. Yet, better geophysical corrections themselves are needed to improve the accuracy and reliability of altimetry data in the coastal zone. The analysis performed here highlights that efforts are still needed on the processing of data and on the geophysical corrections applied to satellite data for studying the sea level trend in the coastal zone over the last two decades more robustly.
It should also be noted that the contribution from waves to sea level variability and trend is not considered here as it is removed from satellite altimetry data since the primary focus of satellite altimetry is to study ocean circulation and dynamics. Yet, wave-induced set-up and run-up can contribute to the sea level trend at the coast (Melet et al. 2016).
7 Summary and Conclusions
In this contribution, we have reviewed the status of the measurements of sea level in the coastal zone. First we have summarized the centennial-scale observations that we get from tide gauges, and then we have aimed at showing how improved satellite nadir altimetry products in the coastal ocean are now complementing those observations by allowing meaningful measurements and the retrieval of coherent long-term signals in the coastal zone. The trend towards better altimetric data at the coast comes not only from improved processing and corrections, but also because of technological innovations such as Ka-band altimetry and SAR altimetry, and we have discussed the main advantages deriving from those two innovations that can be now appreciated thanks to the AltiKa altimeter on SARAL and the SIRAL altimeter on CryoSat-2 (in turn a precursor of the Sentinel-3 altimeter, which has recently been launched and is being operated in SAR mode over the entire ocean).
We have then illustrated the use of altimetry for coastal sea level studies with two examples. First, in a case study conducted along the UK coast, we have found a very good agreement between coastal altimetry and tide gauge observations, with RMSDs as low as 4 cm at many stations. This has given us confidence to use the combination of altimetry and tide gauges to characterize the annual cycle of sea level along the UK coasts. We found amplitudes ranging from 5 to 9 cm, with larger amplitudes found in the northern coasts of the Great Britain, and peaks between early October in the south-east coast and mid-November in most of the west coast. Then, we have examined the evolution of sea level trend from the open to the coastal ocean along the western coast of Africa, comparing standard and coastally improved products. We observed that different products give different answers regarding the coastward evolution of the sea level trend, and we cannot yet robustly deduce the quantitative evolution of sea level trend from the open to the coastal ocean.
The clear recommendation stemming from what we have presented is that further efforts are still needed to study sea level trends in the coastal zone from past and present satellite missions. Further improvements are expected from more refined processing and screening of altimetric data, but in particular from the constant improvements in the geophysical corrections applied to them, such as wet tropospheric, tides and dynamical atmospheric corrections, which all become noisier when coming near shore. It is worth noting that such improvements in corrections should enable the full coastal exploitation of the data now flowing in from Ka- and SAR altimetry, and in particular the global SAR altimetry data now coming from Sentinel-3. This growing coastal altimetry field is going to support the monitoring of sea level in the coastal zone as well as other complementary applications such as the study of extreme events (storm surges—see for example Fenoglio-Marc et al. 2015) and the validation of coastal wave models.
Finally, it is important to remark that the advances in coastal altimetry detailed in this paper prepare the modelling community for the flux of higher resolution data—not only those now starting to flow in from Sentinel-3A (and that will be continued by Sentinel-3B/C/D due for launch over the next 5 years, and then by the two satellites of the Sentinel-6 mission), but also the wide-swath high-resolution observations expected from the surface water and ocean topography (SWOT) mission, due for launch in 2021. The advent of SWOT should hopefully complete the process of ‘closing the gap’ between altimetric observations and tide gauge observations of sea level and hopefully confirm the full consistency of those two sets of measurements.
The TOPEX/Jason-1/2/3 series of satellite altimeters is on a 9.92-day repeat orbit and has continuous measurements since September 1992. The ERS-1/ERS-2/Envisat/AltiKa series is on a 35-day repeat orbit and has measurements since April 1992 with some gaps, the longest of which is between Envisat’s change of orbit in October 2010 and the start of the AltiKa data in March 2013. From July 2016, AltiKa has definitely left the 35-day orbit, and it is in a drifting orbit phase with no more orbit keeping manoeuvres, so the time series on that orbit have ended.
A further extension backwards of the altimetric data record amenable to long-term sea level research could in principle be achieved with ad hoc reprocessing of the GEOSAT mission (1985–1989), and is highly recommended. This will, however, require careful intercalibration with the post-1991 missions, which can be achieved by using carefully selected tide gauges as calibration transfer standards.
A mean sea surface (MSS) is the level of the sea due to all those contributions that can be assumed constant in time and can be computed as the temporal mean of sufficiently long time series. More accurate MSS models (for instance DTU15) are built from combinations of multiple altimetric and gravimetric missions. The choice of a specific MSS over another is critical in the coastal zone, and there may be biases between datasets if they are referred to different MSS.
The AVISO altimeter products were produced by SSALTO/DUACS and distributed by AVISO, with support from CNES (http://www.aviso.altimetry.fr/). The X-TRACK altimetry data used in this study were developed, validated and distributed by the CTOH/LEGOS, France. This work has been partially supported by the Natural Environment Research Council (NERC) National Capability funding. Some results were produced in the framework of the ESA Sea Level CCI project (ESRIN Contract No. 4000109872/13/I-NB) and ESA CryoSat Plus for Oceans (CP4O) project (ESRIN Contract No. 4000106169/12/I-NB). Results presented in Sect. 5 were produced in the framework of the Sea Level SpaceWatch project, funded by the UK Space for Smarter Government Programme through a grant from the UK Space Agency. The tide gauge data used in the Sea Level SpaceWatch study were provided by BODC http://bodc.ac.uk, Channel Coastal Observatory http://www.channelcoast.org/ and by the Port of London Authority. The authors want to thank Florence Birol and Benoit Meyssignac for helpful discussions and Angela Hibbert for the processing of the tide gauge data. This paper represents an outcome of the Workshop on ‘Integrative study of sea level’ at the International Space Science Institute, Bern, Switzerland, 2–6 February 2015.
- Ablain M, Legeais JF, Prandi P, Marcos M, Fenoglio-Marc L, Dieng HB, Benveniste J, Cazenave A (2016) Satellite altimetry-based sea level at global and regional scales. Surv Geophys. doi:10.1007/s10712-016-9389-8
- Birol F, Fuller N, Lyard F, Cancet M, Niño F, Delebecque C et al (2016) Coastal applications from nadir altimetry: example of the X-TRACK regional products. Adv Space Res (in press)Google Scholar
- Cipollini P, Calafat FM (2016) Extended evaluation of CryoSat-2 SAR data in the Coastal Zone, CryoSat Plus 4 Oceans—CCN1 Technical Note, vol. 1.2, pp 30. http://www.satoc.eu/projects/CP4O/docs/CP4O_CCN_Coastal_NOC_v1.2.pdf
- Cotton D, Andersen O, Boy F, Cancet M, Dinardo S, Gommenginger C, Moreau T (2015) Improved oceanographic measurements from SAR altimetry: results and scientific roadmap from the ESA CryoSat Plus for oceans project. Geophysical Research Abstracts, 17, EGU General Assembly, Vienna, Austria, 12–17 April 2015; EGU2015-4613Google Scholar
- Cotton PD, Cipollini P, Calafat FM, Hibbert A, Williams S (2016) Sea level space watch: a service for monitoring UK sea level, wales coastal practitioners conference, Llandudno, April 2016Google Scholar
- Dinardo S (2014) GPOD CryoSat-2 SARvatore Software Prototype User Manual. https://wiki.services.eoportal.org/tiki-index.php?page=GPOD+CryoSat-2+SARvatore+Software+Prototype+User+Manual
- Hauer ME, Evans JM, Mishra DR (2016) Millions projected to be at risk from sea-level rise in the continental United States. Nat Clim Change 6:691–695Google Scholar
- Holgate SJ, Matthews A, Woodworth PL, Rickards LJ, Tamisiea ME, Bradshaw E, Pugh J (2012) New data systems and products at the permanent service for mean sea level. J Coastal Res 29(3):493–504Google Scholar
- Scharroo R, Leuliette E, Miller L (2009) Is sea level rise larger near the coast? 3rd coastal altimetry workshop, Frascati, Italy, 17–18 September 2009Google Scholar
- Shaed KHB, Anand A, Joshi AK, Bhandari SM (2015) Comparison of near coastal significant wave height measurements from SARAL/AltiKa with wave rider buoys in the Indian Ocean region. Mar Geodesy 38(S2):422–436Google Scholar
- Troupin C, Pascual A, Valladeau G, Pujol I, Lana A, Heslop E, Ruiz S, Torner M, Picot N, Tintoré J (2015) Illustration of the emerging capabilities of SARAL/AltiKa in the coastal zone using a multi-platform approach. Adv Space Res 55(1):51–59. doi:10.1016/j.asr.2014.09.011. ISSN 0273-1177 CrossRefGoogle Scholar
- Van Veen J (1945) Bestaat er een geologische bodemdaling te Amsterdam sedert 1700? Tijdschr. K. Ned. Aardrijksk. Genoot LXII:2–36Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.