Recovery of long-wavelength mean gravity anomalies for the dedicated gravitational satellite (gravsat) mission

This paper describes a computer simulation study that was undertaken to determine how well long-wavelength variations in the Earth's gravitational field can be recovered using data from the Dedicated$\text{Grav}$itational $\text{Sat}$ellite (GRAVSAT) mission. This mission is to consist of two low altitude (160 km) spacecraft in essentially the same orbit but separated in phase by 100–300 km. Geodetic data are measurements of the relative range rate to an accuracy of about 1 μm/s at 4 sec intervals. Specifically, a Bayesian covariance simulation was used to investigate simultaneous recovery of the spacecraft ephemerides and a global distribution of 20° × 20° mean gravity anomaly blocks. Sources of errors considered were tracking station positions, gravitational constant, Earth body tides, tropospheric modeling and measurement noise. It should be noted that this simulation does not include as an error source variations in the gravity field that have a character different from what was modeled. Consequently, this study demonstrates the potential of the low-low system as configured to recover the long-wavelength variations in the gravity field.Using only one days worth of data, the mean of the standard deviations of the 162 20° × 20° gravity anomaly blocks is about 1 μgal. For a 6 month mission (assuming a reduction proportional to the square root of the data intervals) this projects to < 0.1 μgal. Because of the potential of increased measurement precision at shorter separation distances, and the relative insensitivity of the recovery process to separation distance, it should be possible to recover both long and short wavelength variations with a modest distribution of separation distances tailored primarily to the short wavelength recovery. Effects of the uncertainty in the gravitational constant and Love's numbers are negligible. In a simulation not reported on, increasing the altitude of the orbit to 200 km from 150 km, degraded, as expected, the accuracy of the recovered parameters by only 7%.