Ground-Based Assessment of Contemporary Global Geopotential Models for Geoid Applications over the North Indian Region
Abstract
In recent years, many global geopotential models (GGMs) have been produced by various international organizations by combining satellite, airborne, terrestrial gravity, marine gravity anomaly, and altimetry data. These models provide a consistent mathematical description of Earth’s gravity field, enabling the calculation of geoid undulation and the conversion of global navigation satellite system (GNSS) ellipsoidal heights to orthometric heights. Over the past few decades, continuous advancements in gravity field modeling have significantly enhanced our understanding of the Earth’s heterogeneous mass distribution, both at the surface and within the subsurface. As a result, GGMs have emerged as a valuable resource in geodesy and geophysics, providing reliable insights into global and regional gravity variations while supporting high-accuracy geoid determination and the unification of height systems. Therefore, this research investigates the efficacy of 19 state-of-the-art and widely adopted GGMs and identifies the appropriate GGM and truncation degree for free-air gravity anomaly reduction using newly acquired 190 GNSS/leveling points and 3,102 gravity anomaly points covering an approximately 363,000 sq. km region with varied topography in the northern Indian region. In this study, the performance of different GGMs was assessed using two independent datasets augmented with the Residual Terrain Model (RTM): geoid heights obtained from GNSS/leveling observations and terrestrial gravity anomaly measurements. XGM2019e_2159 yields the best geoid undulation results over northern India, with an root mean square error (RMSE) of ± 0.199 m at degree 2,160, reduced to 0.143 m after applying a four-parameter corrective surface to account for datum biases, tilts, and drifts inherent to GNSS/leveling comparisons. The spectral analysis further shows that model ranking varies across the low, medium, and high frequency bands, and no single model exhibits uniform superiority across all degrees, while EIGEN-6C4 resulted in the most reliable model for filling data gaps, with an RMSE of ± 29.16 mGal (1 Gal = 1 cm/s 2 = 0.01 m/s 2 ; therefore 1 mGal = 1 × 10 − 5 m/s 2 ).