Assessment of EGM2008 over Sri Lanka, an area where 'fill-in' data were used in EGM2008

Abstract

The tide-free EGM2008 combined global geopotential model is compared with land and marine gravity observations and co-located GPS-levelling on and around Sri Lanka (formerly Ceylon). Not all these data are in the public domain, so offer an informative test of how the 'fill-in' methodology used in EGM2008 performs versus observed data. Sri Lanka is also in an area where the geoid exhibits its lowest elevation with respect to a geocentric reference ellipsoid. A -1.75 m bias between the GPS-levelling and EGM2008 led to an investigation into the Sri Lankan geodetic datums, showing a bias in the ellipsoidal heights. After rejection of 15 outliers, the standard deviation of the difference between 207 Sri Lankan GPS-levelling points and EGM2008 is 0.184 m. The difference between the gravity anomalies and EGM2008 showed that the Sri Lankan gravity data is based on the old Potsdam datum. The Sri Lankan land gravity data, after rejection of outliers, yielded standard deviations of 6.743 mGal for 20 GPS-coordinated gravity points on fundamental benchmarks, 14.704 mGal for 42 gravity points on fundamental benchmarks but with coarse locations, and 6.367 mGal for 1032 digitised and reconstructed free-air anomalies from a Bouguer anomaly map. The ship-track gravity data have not been crossover adjusted, and yield a standard deviation of 43.683 mGal. Importantly, the ability of EGM2008 to identify datum deficiencies is an implicit validation and leads to its applicationin other areas to search for datum deficiencies.