Numerical experiences on using the RTM method in quasi-geoid modeling

In local and regional quasi-geoid modeling, the residual terrain modeling (RTM) method is often used to remove the short-wavelength gravity field signals from the measured gravity both on the ground and up in the air, in order to obtain the regularized and smooth gravity field which is suited for field interpolation and modeling. Accurate computation of RTM corrections requires a set of fine-tuned parameters in terms of the gravity forward modeling technique, digital elevation model (DEM), reference topography, and integration radius for the inner zone and outer zone. To our limited knowledge, this has not been systematically documented, albeit its importance is obvious. This work aims to clearly investigate the impact of these factors on the RTM correction computation and their effects on the quasi-geoid modeling so to provide practical guidelines for real applications. Two gravity forward modeling techniques, i.e., the prismatic approach and the tesseroidal approach, are employed to investigate the following issues existing in the practical use of the RTM method: ① can a combination of a high-resolution DEM and a DEM with a lower resolution replace the use of a single high-resolution DEM for the RTM correction computation without loss of accuracy? ② how to properly choose the integration radius for the inner zone and outer zone while costing less time and keeping the accuracy? ③ what are the performances of using the reference topographies obtained by the direct averaging, the moving averaging, and the spherical harmonic analysis and synthesis on the RTM correction computation and quasi-geoid determination? For obtaining objective findings, two research regions are selected for this investigation. One is the Colorado area (USA) having quite rugged terrains and the other is the Auvergne area (France) with moderate terrains. The numerical results show that, in the computation of RTM corrections to gravity anomaly and height anomaly, the combination of a dense DEM and a coarse DEM can replace the use of a single dense DEM without loss of accuracy. The increasing and decreasing of the integration radius for the inner zone and outer zone do not influence the RTM correction computation much. The recommended values are 10 km for the inner zone and 111 km for the outer zone. The use of different reference topographies changes RTM corrections, however, the final quasi-geoid models are at the similar accuracy level.