BACK

 

Technical issues and recommendations related to the installation of continuous GPS stations at tide gauges

Case study - Marseille, France (GLOSS-LTT 205)

Guy Wöppelmann
Centre Littoral de Géophysique (CLDG)
Université de La Rochelle, France

1. Motivation

Marseille has a secular history of sea level records beginning in 1885. The belief, about a century ago, that the average level of the sea was constant over long periods of time led to define the concept of geoid and, subsequently, to establish the origin of the levelling networks on "mean sea level". To determine this quantity in France, the sea level observatory of Marseille was built on the Mediterrean coast in 1884. The French datum (NGF) was then computed over an arbitrary time period from continuous tide gauge records covering the period of February 1885 to December 1896.

However, mean sea level varies from place to place and at one specific place over time. Today, the mean sea level at Marseille is about 10 cm above the local 1885-1896 datum. Thus, the datum no longer represents the "real" average sea level at this site (Click on the figure to enlarge the view).

Today, secular sea level time series are of utmost importance for monitoring long term trends and accelerations in global sea level. Marseille contributes to the dedicated set of GLOSS Long Term Trend stations (GLOSS-LTT). However, in order to provide an estimate of 'absolute' rather than a 'land relative' sea level secular trend, continuous geodetic positioning of tide gauge benchmarks is required.

A permanent GPS station has been installed in June 1998 by the French agency IGN (Institut Geographique National) in Marseille. The main objective of this GPS station is to determine the vertical land motion that 'contaminates' the tide gauge record. If the motion proves to be essentially linear in the next decade(s), it may be subtracted with confidence to the historical tide gauge record collected over the last 100 years.

Tide gauge data analysis show that the observed sea-level trend at Marseille can be estimated with a 'statistical' error lower than about 0.1 mm/yr. GPS measurements should therefore strive to achieve similar sub-millimiter accuracy for land motion estimate. However, getting such an accuracy on the vertical component is quite challenging today and rises open questions about the time-dependent systematic errors of some technical and environnemental GPS aspects, for instance : antenna phase center variations, troposphere corrections, reference frame realisations, atmosphere and ocean tide loading. This scientific argument as well as practical cost-effective considerations explain why IGN chose a continuous GPS observing strategy instead of frequently repeated campaigns.

 

2. General background

Marseille is located on the French Mediterranean coast. The tide gauge site was carefully chosen for its long term stability and representativity of open sea conditions (away from any source of fresh water). At the end of the 19th century, these criteria were considered of the highest importance to establish the origin of the French national height system.

The sea-level observatory is under IGN's charge. It is a concrete structure built in 1884 directly on hard bedrock, obviously highly representative of the surrounding area motion. The above picture (left) shows the tide gauge building in the foreground. The former tide gauge guardian house is also visible in the background of the picture.

 

3. Handling constraints for GPS installation

3.1. Science

The main objective of the GPS installation is to estimate and remove the vertical motion of the tide gauge from the relative trend in mean sea level with an accuracy better than 1 mm/yr. This vertical motion may be an addition of (i) the crustal motion of the whole surrounding region and (ii) the local subsurface and monument instabilities.

Locating the GPS antenna upon the tide gauge ground base monument was a priority, as this ideal situation ensures (i) no subsequent relative motion of the antenna and the tide gauge to be monitored by frequent levelling or GPS ties, (ii) no subsequent increase of the total error budget (which is already a very challenging agenda : less than 1 mm/yr !).

Considering these constraints, the standards have been slightly lowered on the technical and environmental aspects of the GPS installation. Next section develops these aspects. It turns out that none of them were as severe as to preclude the GPS installation upon the tide gauge monument, that is, on the roof of the tide gauge building (Click on the figure to enlarge the view).

3.2. Technique and Environment

Elevation masks - The sketch on the left shows an almost clear view of the sky in all directions down to 10 degrees elevation for a GPS antenna located on the sea-side border of the tide gauge roof, except for a 40 degrees azimuth area in the Eastwards direction. There, an abrupt slop topography block the GPS satellite signals up to a 20 degrees elevation angle.

Multipath is hardly noticeable. If any it might arise from the former tide gauge guardian house. No radio interference has been noticed either.

GPS receiver and antenna - Following the standards of the International GPS Service (IGS) aiming at precise geodynamical applications, a dual frequency code and phase measuring GPS receiver and a high quality Dorne-Margolin choke ring antenna were purchased and installed in June 1998.

Power supply and communication facilities -The tide gauge building was already provided with standard electrical power (220 Volts AC). New power points were installed to supply the scheduled instruments (tide gauge, GPS receiver...). Nevertheless, the maximum available length of GPS antenna cable was needed (30 meters). As there were no data communication facilities and no local staff, the place has been equipped with several phone lines and modems in order to allow GPS data downloading from afar.

Monumentation - The tide gauge building is a rigid ancient concrete monument, directly built on the underlying solid crust of the earth. Moreover, by examining the relative long height history of the Marseille tide gauge benchmarks, one can confirm and conclude that the tide gauge building has not undergone any significant local relative subsidence and that it can be viewed as stable at the sub-millimeter level with respect to the bedrock.

To limit monument instabilities no pillar and no mast were built : the GPS antenna was set on a centering aluminium plate anchored in the solid rock of the tide gauge roof. Aluminum is generally immune from corrosion when exposed to sun and rain. The shape of the plate is triangular. Its size is smaller than the base of the antenna and its vertical axis of symmetry coincides with the center of the antenna.

Local ties - An hemispheric bronze bolt mark was placed below the aluminium triangle plate, which actually has a hollow center. The vertical offsets between the antenna reference point, the hemispheric bronze mark and the tide gauge benchmark were determined at the time of construction. Although the vertical offsets can be assumed to be constant, levelling ties are scheduled each time a tide gauge calibration takes place. The levelling tie is not as difficult as one would imagine for an antenna located on the roof : the tide gauge building is at the level of the adjacent road. The levelling ties between all the TGBMs are often performed in one single day.

 

4. GPS data collection and processing

Marseille is presently equipped with a Trimble 4000 SSI GPS receiver. Its internal memory capability is on the low side (about 10 Mbytes). There is no computer whatsoever on site. Automatic download routines are run at the EUREF Local Data Center 'IGN' in Paris (also called RGP). The GPS data files are downloaded on a daily basis over the phone line.

GPS measurements are then quality checked and written in RINEX format. The RINEX files are available at the IGN/RGP web and FTP servers with about one day delay :

At least five analysis centers process Marseille GPS data :

 

BACK