ANU RL03 mascon solutions: regularisation based on range acceleration signal content

The ANU mascon monthly solutions are derived from least squares inversions of inter-satellite range acceleration signals, using irregularly shaped mass concentration (mascon) elements that cover the Earth. As with all GRACE and GRACE-FO monthly solutions of the temporal gravity field, regularisation of the data inversion must be applied to mitigate the noise in the solutions to permit the geophysical signals to be detected. Previous releases of ANU mascon solutions have used a “regional” regularisation approach, whereby continent-scale areas have been assigned the same mascon parameter uncertainty, which has been used for every month, in what was an attempt to use a “standard” (rather than a “tailored”) regularisation. While individual monthly solutions are smooth and free of striping errors, this standard approach led to noisy time series, with unrealistic variations in mass from one month to the next.

In our ANU RL03 mascon solution we use the level of signal in the range acceleration residuals to determine a unique regularisation sigma for each mascon for each month. While much more time consuming to generate, the RL03 time series of monthly temporal gravity field estimates are more realistic in mass change variations, while maintaining similar long-term trends as the RL02 solutions. Comparisons of degree variance of our RL03 solutions show a comparable level to the other mascon solutions (CSR, JPL, GSFC).

In this presentation we will explain the process of deriving the “tailorerd” regularisation matrix for each month and compare our RL03 solutions to the other mascon solutions. We will show some interesting differences between our and the other solutions, in particular for the polar ice sheets.