Testing mascon regularization strategies for simulated future gravity-field recovery

Abstract

Mascon solutions estimate mass changes directly on predefined surface elements, rather than only representing them by spherical harmonic coefficients. They are useful for regional water storage, ice-mass, and ocean applications, and are already provided by several groups such as GSFC, JPL and CSR. In the framework of GRACE-C Science Data System preparations at GFZ, a project funded by DLR, we are developing and testing mascon processing strategies using realistic full-noise simulations of monthly gravity-field solutions.We first consider a GRACE-C-like single polar-pair scenario, representing the continuation of the US-German GRACE mission series after GRACE and GRACE-FO. We then compare this with the proposed MAGIC double-pair constellation, an ESA-NASA concept in which the NASA/DLR GRACE-C polar pair is complemented by ESA’s inclined NGGM pair. By comparing the two scenarios, we investigate how the observation geometry affects the recovered mascon fields and the required regularization.The processing starts from monthly spherical harmonic solutions. These are regularized in coefficient space before being transformed to the mascon grid. We test different spherical-harmonic regularization settings and inspect their effect on striping, ocean RMS, and land signal preservation.In the mascon step, we compare two ways of building the regularization matrix. The first uses an intermediate mascon solution to construct a time-variable regularization matrix. The second derives the time-variable regularization matrix directly from synthesized spherical harmonic fields, avoiding the intermediate mascon step.The recovered fields are compared with the known input signal to evaluate noise reduction and signal preservation. Preliminary results show that time-variable mascon regularization can reduce errors compared with the first mascon solution, but the best balance between noise reduction and signal preservation is still being evaluated. This work presents the current workflow, first results, and open questions for discussion.