| 1 | Coverage area | Lat: 16° - 33° Lon: 34° - 56° | Lat: 10° - 35° Lon: 30° - 62° | Larger coverage area |
| 2 | Standard accuracy of geoid height | ≈ 2–3 cm near benchmarks in eastern Saudi Arabia; ≈ 10–20 cm elsewhere in the Kingdom. | ≈ 1.5 cm near leveling baselines; ≈ 2.5–3.5 cm elsewhere in the Kingdom, particularly in the mountainous southwest. | More uniform accuracy across the Kingdom, averaging better than 2 cm |
| 3 | Global Gravity Model (GGM) used as reference | EIGEN6C4 | XGM2019e | The latest available GGM was used as the reference model |
| 4 | GGM used to fill gravity anomaly data outside the Kingdom | None | EGM2008 (372793 points) | Edge effects and oscillations were avoided |
| 5 | Marine gravity data used | 139331 | 245167 final points | More marine gravity data were used |
| 6 | Airborne gravity data used | None | Grid 1’ x 1’ (1800 x 1800m) - Line data - | Additional data covering a large area were used to validate terrestrial gravity measurements |
| 7 | Satellite-derived gravity data over water used to fill marine areas without gravity measurements | DTU15 (301101 points on land) | DTU18 (raw 147480 near-shore points up to 20 km); SIO S&S29.1 (raw 626913 off-shore points >20 km); total points: raw 774393; final 771500 | 2.5 times as much data as in the previous geoid model |
| 8 | Ultra-high Earth Residual Terrain Modelled (ERTM) gravity effects | SRTM30 (the model has resolution 30” (900 m)) | dv_ell_EARTH2014 (up to spherical d/o 2160 (spherical d/o 2190) for scales >10 km) ERTM2160 (for spheroidal d/o 2160 to 90000 - latest model (released in 2014) taking into account effects at scales from 10 km to 250 m) | Higher accuracy in areas with varied terrain |
| 9 | Digital Elevation Model (DEM) used to calculate precise quasigeoid-to-geoid separation | None | SRTM3 (the model has resolution 3” (90 m)); GASGI DTM (resolution 0.1” (3 m)) | Separation was measured accurately |
| 10 | Number of points with raw data | Final number of points: 890975 | Raw points: 2 383 559; final points: 2 010 766 | Larger number of observed points |
| 11 | Observation data validation | | | Two additional validation methods were used: Least Square Collocation (LSC); airborne data downward continuation to the land surface. |
| 12 | Number of benchmarks (BMs) with heights measured using GPS | 280 benchmarks from GEOSA | 3522 benchmarks from GEOSA; 13206 from Saudi Aramco; 801 from other organizations; total: 17529 benchmarks | More uniform fitting of the geoid model to benchmarks in Saudi Arabia |
| 13 | Difference model between the gravity-only GRAV-GEOID21 model and benchmark geoid heights derived from measured ellipsoidal heights when fitting the gravimetric geoid to benchmarks | None | combined deterministic and stochastic model | The approach provides a more uniform and accurate distribution of fitting errors |