A height datum for Uganda based on a gravimetric quasigeoid model and GNSS/levelling

Abstract: This study is devoted to the determination of a high resolution gravimetric geoid model for Uganda based on the optimal combination of terrestrial and satellite gravity anomalies using the method of Least Squares Modification of Stokes’ formula with additive corrections. Specifically the study investigates the current status of the existing Uganda Vertical Network relative to the requirements of a modern height datum and includes a detailed evaluation and validation of terrestrial gravity data, several digital elevation models and some recent global geopotential models. Finally a new height datum based on a gravimetric quasigeoid model and Global Navigation Satellite Systems (GNSS)/levelling is proposed.In this thesis, the Uganda Gravimetric Geoid Model 2014 (UGG2014) is computed from several datasets which, include 7839 terrestrial gravity data points from the International Gravimetric Bureau, the 3 arc second Shuttle Radar Topography Mission digital elevation model and a recent Gravity field and steady-state Ocean Circulation Explorer-only global geopotential model. To compensate for the missing gravity data in the target area, the surface gravity anomalies extracted from the World Gravity Map 2012 were used. Outliers in the terrestrial gravity data were detected using the cross-validation technique which, also estimated the accuracy of the remaining terrestrial gravity data as 9 mGal. Based on 12 GNSS/levelling data points distributed over Uganda, the root mean square fit of UGG2014 before and after the 4-parameter fit is 16 cm and 9 cm, respectively.The study has revealed that the heights of the Uganda Vertical Network are normal-orthometric heights for which the quasigeoid is the closest approximation to the zero reference surface. Consequently, the Uganda Gravimetric Quasigeoid Model 2014 (UGQ2014) was derived from the UGG2014 with the quasigeoid-geoid separation computed from the Earth Gravitational Model 2008 complete to degree/order 2160 of spherical harmonics. The root mean square fit of UGQ2014 versus GNSS/levelling is 15 cm and 8 cm before and after the 4-parameter fit, respectively, which shows that the quasigeoid model fits GNSS/levelling better than the geoid model. Thus a new height datum based on UGQ2014 and GNSS/levelling was determined as a practical solution to the determination of heights directly from GNSS. Evaluated with 4 independent GNSS/levelling points, the root mean square fit of the new height datum is 5 cm better than using the quasigeoid model alone. With an average parts-per-million of 29 in the relative test, the new height datum satisfies the precision and accuracy requirements of third order precise levelling. Overall, the results show that UGG2014 and UGQ2014 agree considerably better with GNSS/levelling than any other recent regional/global gravimetric geoid models. Therefore, both gravimetric solutions are a significant step forward in the modelling of a “1-cm geoid” over Uganda given the poor quality and quantity of the terrestrial gravity data used for computation.