The Determination and Error Assessment of the Earth's Gravity Field Model

Precision orbit determination methods, along with a new technique to compute relative data weights, have been applied to the determination of the Earth’s gravity field and other geophysical parameters from the combination of satellite ground-based tracking data, satellite altimetry data, and the surface gravimetry data. The University of Texas Earth's gravity field model, TEG-2, was determined from data sets collected from fifteen satellites, spanning the inclination ranges from 15° to 115°, and surface gravity data. The satellite measurements include laser ranging data, doppler range-rate data, and satellite-toocean radar altimeter data, which include the direct height measurement and the differenced measurements at ground track crossings (crossover measurements). The terrestrial gravimetry data used in the TEG-2 solution were provided from the Ohio State University 1°×1° area-mean free-air gravity anomaly data and elevation data. The estimated parameters in the TEG-2 solution include GM - gravitational constant of the Earth, geopotential coefficients complete to degree and order 50, selected dynamical tide coefficients, coordinates of the doppler tracking stations, two separate sets of surface spherical harmonics coefficients complete to degree and order 10 for Seasat and Geosat altimeter-derived quasi-stationary sea surface topography. Error assessment and calibration of the TEG-2 solution and formal error covariance have been performed. Observation residual combined with covariance prediction analysis indicate that a level of internal consistency has been achieved in the TEG-2 solution, and a calibration factor has been determined for the TEG-2 formal error covariance matrix so that the resulting covariance presents a realistic accuracy of TEG-2 solution. The improvement of the TEG-2 geopotential solution presented in this investigation is due in large part to the optimal weighting technique applied to the simultaneous solution for the topography, the gravity field, and the satellite orbit. The accuracy of geographically correlated mean for Topex radial orbit predicted from the 36×36 part of the TEG-2 geopotential error covariance is about 11 cm, and the global accuracy of geoid undulation predicted from the 50×50 part of the TEG-2 geopotential error covariance is about 1.6 m.