Gravity field processing towards future LL-SST satellite missions

This study focuses in aspects concerning gravity field processing of future LL-SST satellite missions. Closed-loop simulations taking into account error models of new generation instrument technology, are used to estimate the gravity field accuracy that future missions could provide. Limiting factors are designated, and methods for their treatment are developed. The contribution of all error sources to the error budget is analyzed. It is shown that gravity field processing with double precision may be a limiting factor for exploiting the nm-level accuracy of a laser interferometer. An enhanced precision processing scheme is proposed instead, where double and quadruple precision is used in different parts of the processing chain. It is demonstrated that processing with enhanced precision can efficiently handle laser measurements and take full advantage of their accuracy, while keeping the computational times within reasonable levels. However, error sources of considerably larger impact are expected to affect future missions, with the accelerometer instrument noise and temporal aliasing effects being the most significant ones. The effect of time-correlated noise such as the one present at accelerometer measurements, can be efficiently handled by frequency dependent data weighting. Residual time series that contain the effect of system errors and propagated accelerometer and laser noise, is considered as a noise realization with stationary stochastic properties. The weight matrix is constructed from the auto-correlation functions of these residuals. Applying the weight matrix to a noise case considering all error sources, leads in reduction of the error levels over the complete bandwidth. Co-estimation of empirical accelerations does not show the same efficiency in reducing the propagated noise with the applied processing strategy. Temporal aliasing effects are reduced essentially by adding a second pair of satellites at an inclined orbit. Compared to a GRACE-type near-polar pair, a Bender-type constellation delivers solutions with major improvements in terms of de-aliasing potential and recovery performance. When the integrated effect of all geophysical processes is recovered, the maximum spatial resolution of 11-day solutions can be increased from 715 to 315 km half-wavelength. A further reduction of temporal aliasing errors is possible by co-parameterizing low resolution gravity fields at short time intervals, together with the higher resolution gravity field which is sampled at a longer time interval. One day was found to be the optimal sampling period for reducing the error levels in the solutions. A uniform sampling at the co-parameterized short periods, is a prerequisite for an efficient reduction of aliasing errors. High frequency atmospheric signals are captured by daily solutions to a large extent. Hence co-parameterization at daily basis results in big reduction of aliasing caused by their under-sampling. This enables future gravity satellite missions to deliver the complete spectrum of Earth's geophysical processes. The corresponding by-products of daily gravity field solutions are expected to be very useful to atmospheric science and open doors to new fields of applications.