Long range geoid control through the European GPS traverse: Final results

The European north-south Global Positioning System (GPS)-traverse proposed by IAG SSG 3.88, should control and improve the European geoid. This traverse follows first order leveling lines, included in the United European Leveling Network. From May to August 1986 and in July 1987, the central and northern part of this traverse (approx. 3000 km) was observed using up to four TI 4100 receivers, covering Austria, Federal Republic of Germany, Denmark, Sweden and Norway. Both traverse parts contain 71 stations with distances of about 50 km. In addition, 8 stations have been occupied for overlapping connections, and traverse links were established for connecting the fundamental stations Wettzell (VLBI and SLR) and Onsala (VLBI). Final results show a GPS observation precision of a few cm for loops of some 100 km circumference. After transformation of the GPS results to geoid heights using the leveled heights, comparisons with different existing gravimetric geoid determinations including geopotential models were performed. In addition, new geopotential models complete to degree and order 360 tailored to gravity data in Europe, and gravimetric geoid solutions using 6 x 10' mean gravity anomalies were investigated. The comparison with GPS and leveling yields rms discrepancies of + or - 0.1...0.2 m over 1000 km traverse sections for the best solutions, but a strong slope is existing in Sweden and southern Norway in almost all solutions, which is probably caused by systematic errors in the available gravity data for Scandinavia. This is confirmed by a new geoid computation at the Danish Geodetic Institute where the slope has disappeared. If this new solution is taken for the northern traverse section and the best solution for the central part, the rms discrepancy reduces to approximately + or - 0.2 m over 3000 km. Thus, a + or - 10 (exp 7) relative height accuracy seems to be achievable over long distances with the GPS/leveling and the gravimetric geoid calculation techniques, applied in this experiment.