Investigating effects of solution design choices on L2-based mascon estimates

Spatially resolved, or spatially constrained, mass changes are the primary science application from GRACE and GRACE-FO. Mascon solutions provide a framework for such mass change solutions, and Mascon solutions based on the analysis of L1 data have become convenient and popular for many users. However, design choices inherent to L1-based solutions are difficult to assess, or to adapt, by users with regard to their specific applications. Mascon solutions based on L2 gravity field solutions allow more access to, and control of, design choices by a wider range of scientists. Such L2-based globally gridded mass change solutions require to find a compromise between propagated GRACE L2 solution errors and leakage errors.

Here we continue previous work on assessments of L2-based mass change solutions. We extend the method of tailored sensitivity kernels, which was originally derived for regional mass changes of the ice sheets, to derive globally distributed mascons based on GRACE Level-2 spherical harmonics. This method allows for a fast solution iteration for assessing different design choices. The resulting sensitivity kernels, which describe the weighting functions that are used to integrate the input data, allow for a direct interpretation of the mass integration step of individual mascons. We present the impact of design choices on ocean mass change and signal leakage over the land-ocean margin. We focus on two design choices: (a) the amendment of  a-priori mascon patterns by their sea-level fingerprints and (b) the choice of signal variances and covariances.