In this study, the GIS-based raster analysis principle was applied as a trial to enhance the local performance of geopotential models. Specifically, harmonic models-derived geoidal height raster layers, with respect to the WGS-84 reference ellipsoid, were investigated. As an illustrative case study, the Egyptian Territory was used as a test region. A set of geoidal height grids was synthesized from five geopotential harmonic models. The synthesis window was organized so as to avoid any subsequent edge effects during the analysis process. After importing such raster layers into a GIS-platform, an iterative series of linear high-frequency filters was applied. The progress of the spatial quality of the filtered layers, at independent GPS/Leveling points, was interactively monitored. Then, the optimal filtering radius and the relevant best attained spatial quality were picked out for each of the five raster data. The results revealed that the applied GIS-based sharpening algorithm works well, even for a satellite-only model. In general, large spatial quality enhancements were achieved. The maximal attained improvement was 21 and 29 %, in terms of standard deviation and range of differences, respectively, at the control points. So, it is recommended to try the GIS-based raster analysis scheme over other geographical windows. Also, linear and non-linear forms of high-frequency filters may be applied and tested, regarding the improvement of other types of synthesized gravitational signals.