COMBINING ERS ALTIMETRY AND SURFACE GRAVIMETRY IN THE AZORES REGION FOR LOCAL GEOID DETERMINATION

Abstract

This study is part of a project which aims the determination of a regional and precise geoid model for the Azores-Gibraltar region, by combining radar altimetry from ERS satellites and surface gravity data. The study area is the North Atlantic region between the Azores Archipelago and the Portuguese mainland (36°N = latitude = 43°N, 32°W = longitude = 10°W). At this stage two independent solutions have been derived: an altimetric mean surface from ERS multi-mission radar altimeter data and a gravimetric geoid from all available surface gravity data. This paper presents the results of the comparison of these surfaces for simultaneous validation of the altimetric and the gravimetric solution and addresses the steps to follow in order to integrate these data to obtain a better regional geoid model for this area.

1. INTRODUCTION

Regional geoid solutions computed solely from gravity measurements are known to be very sensitive to data coverage. In most parts of the world’s ocean the data are sparse with an irregular spatial distribution and variable accuracy. The problem will continue until the launch of a dedicated gravity mission.

The quantity of data available from the most recent altimetric missions such as GEOSAT, ERS-1, ERS-2 and TOPEX/Poseidon, with increasing accuracy and diversity in data coverage, have made possible the use of these data for a variety of applications, including the determination of the marine geoid. The main limitation here is the lack of an accurate knowledge of the dynamic sea surface topography, determined by independent hydrographic measurements. In the absence of such a model alternative methods have been attempted and several approaches to this problem have been published in the dedicated literature. They mainly concentrate on the computation of combined solutions where gravity anomalies derived from altimetric sea surface heights are used to fill the gaps in the gravimetric data coverage. The main limitation of this procedure is also the inability of properly correcting the altimetric sea surface heights for the dynamic sea surface topography. Our approach aims to integrate, in an optimal way, a gravimetric solution with an altimetric surface by developing a methodology which uses the spectral information from both data sets, in particular the fact that existing altimetric data, spanning a variety of different spatial and temporal resolutions, depict different features of the geoid and the ocean circulation.

2. methodology

The altimetric data used are three data sets from ERS-1 and ERS-2 OPR02 altimeter data as summarised in Table 1. To reduce the errors at the edges, an altimetric solution has been computed selecting data from a larger area containing our study region.

Data processing started with the computation of residual sea surface heights relative to a reference field model (OSU91A). The reference ellipsoid used in all computations was GRS80. The satellite ephemerides used were the precise DGM-E04 orbits made available by Delft University of Technology (TUDELFT), (Scharroo, 1996).

Table 1 - Altimetric data used

The altimetric measurements were corrected for all instrument and geophysical effects including solid Earth and ocean tides, tidal loading, dry and wet tropospheric corrections (ESA, 1994 and ESA, 1996), inverse barometer effect and sea state bias (Gaspar and Ogor,1994). Altimeter height bias were applied (-40.9 cm for ERS-1 and 0 cm for ERS-2) as well as time tag bias according to TUDELFT recommendation (1.5 ms for ERS-1 and 1.3 ms for ERS-2). For each data set a crossover adjustment was performed using tilt and bias parameters (Rummel, 1993). The rms of crossover differences was 13 cm for both ERS-1 data sets and 16 cm for the ERS-2 set. After the adjustment the rms of crossover differences was 8 cm for all data sets.

After the crossover adjustment, mean tracks were computed for the 35-day cycle, by averaging all ERS-1 and ERS-2 coincident tracks. First a regular grid of points at 2 seconds interval (D l=0.02°, average D j=0.06°) was created, by fitting a polynomial of degree 2 to all ERS-1 coincident tracks. Then all coincident tracks from both satellites were interpolated to the grid points, using 2-D interpolation.

The final surface (35-day mean and 168-day data) was adjusted to OSU91A. The result of this adjustment is roughly the removal of a plane through the observations with an west-east trend. From this final mean altimetric surface the subset covering the study area was selected. This subset contains 74270 points from 690 tracks. This mean surface is represented in Figure 1.

The gravimetric solution (Figure 2) is a regional model computed in the scope of project GEOMED (GEOid in the MEDiterranean), (Catalão, 1994). The data used are free air gravimetric anomalies provided by international institutions. The computation model utilised was the remove-restore technique with the central step performed by least squares collocation. The geopotential model OSU91A and the bathymetric model ETOPO5U (with a shift of 5’ west ) were used to model the long wavelength part of the field and to compute the residual terrain corrections respectively. The global accuracy of this model over the whole area is about 20cm.

3. RESULTS

The two independently computed solutions were analysed either globally and along satellite tracks (35-day means only). Figure 3 represents the difference between the two surfaces, after applying a low pass filter to remove data noise. Figures 5, 6 and 7 represent differences along satellite tracks located over regions of particular interest. The location of these tracks is also shown is Figure 3.

The adjustment of the altimetric surface to OSU91A removes the bias and tilt between the two surfaces due to differences in the corresponding reference systems and serves our purpose which is to analyse in detail the relative differences between the altimetric and the gravimetric solution. Before the adjustment, the mean (m) and standard deviation (s) of the differences between the altimetric and the gravimetric solutions were m=-0.619m and s=0.243 m respectively. After the adjustment the corresponding values were m=-0.085 m and s=0.210 m.

The average difference of about 60 cm found between the altimetric mean sea surface heights and both the regional gravimetric geoid and OSU91A heights above GRS80 ellipsoid confirms that the best fitting ellipsoid has an equatorial radius shorter than the corresponding value adopted by GRS80 (a=6378137 m) by about the same amount. The adjustment of the altimetric surface to OSU91A resmoves this bias from the results. It also removes the average mean dynamic sea surface topography over the whole area.

Results obtained with ERS-1 and ERS-2 data are very similar, and consistent with the value adopted for the relative height bias between these satellites.

The spectral information of the altimetric and gravimetric solutions is very similar. When compared along the satellite tracks, they reveal the same wavelengths, showing different amplitudes, due to sea surface topography features or lack of gravimetric data. Large discrepancies are detected mainly near tectonic plate boundaries.

The study region has very interesting features from the geodetic point of view such as the Azores triple junction (in the west) and areas of very rough bathymetry and high geoid gradients such as the Goringe ridge (in the east). Apart from some restricted zones, there is a good overall agreement between the altimetric and gravimetric solutions, even in areas with very irregular bottom topography and large height gradients, such as the Goringe ridge (Figure 6). The largest discrepancies may be explained by the lack of sea gravimetric data or by the inaccuracy of the bathymetric model used. The ETOPO5U model is known to be inadequate close to plate boundaries.

Figure 1 - Mean Altimetric Surface

Figure 2 - Gravimetric geoid

Figure 3 - Mean Altimetric Surface - Gravimetric geoid

Figure 4 - Track across the Azores Triple Junction

Figure 5 - Track across S. Miguel Island

Figure 6 - Track across the Goringe ridge

This might explain the large discrepancies shown near the Azores microplate (Figures 4 and 5).

4. FUTURE WORK

These results clearly point out the need to integrate the information contained on the altimetric and gravimetric solutions in order to determine a combined geoid model. The different spatial and temporal information contained on the 35-day and 168 day data sets will be used to help separate the dynamic topography features from the geoid. Hydrographic models determined by independent data, will be used to validate the resulting dynamic sea surface topography.

This study is one step towards the determination of a precise geoid model for this region by combining heterogeneous data: altimetry (sea and airborne), gravimetry, tide gauge data and GPS measurements. An airborne gravimetric and altimetric campaign is planned within the MAST III project AGMASCO in the Azores region, for October 1997. This will also allow interesting comparisons between the satellite and airborne altimetric data.

5. REFERENCES

Catalão, J., 1994, Geoid Computation in the North-East Atlantic (Açores - Portugal), First Results - Mare Nostrum, No 4, 77-90

ESA (European Space Agency), 1994, Altimeter Products User Manual, C1-EX-MUT-A21-01-CN

ESA (European Space Agency), 1996, Altimeter & Microwave Radiometer ERS Products User Manual, C2-MUT-A-01-IF

Gaspar P. and F. Ogor, 1994, Estimation and analysis of the sea state bias of ERS-1 altimeter., Report of task B1-B2 of IFREMER contract Nº 94/2.426 016/C

Rummel, R., 1993, Principle of Satellite Altimetry and elimination of Radial Orbit Errors, Lecture Notes in Earth Sciences, 50, Springer Verlag, 190-241.

Scharroo, R., 1996, Gravity Field and Orbit Improvement for the ERS Satellites, IV European Conference on Satellite Altimetry, 2-4 October 1996, Porto, Portugal

Keywords: ESA European Space Agency - Agence spatiale europeenne, observation de la terre, earth observation, satellite remote sensing, teledetection, geophysique, altimetrie, radar, chimique atmospherique, geophysics, altimetry, radar, atmospheric chemistry