Three different methods of handling topography in geoid determination were
investigated. The first two methods employ the residual terrain model (RTM)
remove-restore technique, yielding the quasi-geoid, whereas the third meth
od uses the classical Helmert condensation method, yielding the geoid. All
three methods were used with the geopotential model Earth Gravity Model (19
96) (EGM96) as a reference, and the results were compared to precise global
positioning system (GPS) levelling networks in Scandinavia. An investigati
on of the Helmert method, focusing on the different types of indirect effec
ts and their effects on the geoid, was also carried out. The three differen
t methods used produce almost identical results at the 5-cm level, when com
pared to the GPS levelling networks. However, small systematic differences
existed.