Development of a hybrid geoid model using a global gravity field model over Sri Lanka

Abstract

Sri Lanka is still in need of a well-defined local geoid model. This geoidal void has made present-day Global Navigation Satellite System (GNSS) surveys heavily dependent on Global Geopotential Models (GGMs) for height determination. Further, in many instances, the accuracy of GGMs have shown drawbacks in elevation determination over Sri Lanka. Therefore, the study focused on developing a hybrid geoid model (HGM) for Sri Lanka by integrating the available GGMs. Five high-resolution (2190°) GGMs; EGM2008, EIGEN-6C4, GECO, XGM2019e-2159, and SGG-UGM2 were employed to extract GGM-derived geoid undulation for 21 Fundamental Benchmarks (FBMs). The residuals (geoid height deviation) were calculated relative to the observed geoid undulation using GNSS/leveling on the FBMs. The data set was clustered based on topography, and residuals were adjusted using weighted least squares adjustment (LSA). The uneven distribution of the FBMs promotes topography-based clustering. EIGEN-6C4 is found to be the robust GGM for Sri Lanka to develop a hybrid approach, with a 0.001 m RMS value of estimated residuals in LSA. The resulting HGM was interpolated at 1 arc-second grid resolution (30 m × 30 m) using the Inverse Distance Weighted Interpolation. Regression lines were generated for the interpolated HGM with respect to the interpolated observed geoid undulation for 9 transects along the parallel passing through Mount Pedro and for the 16 transects along the meridian. The coefficient of determination on both lines is 0.999. HGM generated by EIGEN-6C4 has shown reliable RMS gradient and intercept values of 8.860078 × 10−9 and 0.0039239, respectively, in first-order polynomial fitting.