Filtering techniques and their potential application for generating new products for GRACE / GRACE-FO

Jürgen Kusche
Institute for Geodesy and Geoinformation, Bonn University

Experience with GRACE has shown that scientific applications of the standard Level-2 products (monthly geopotential models) require a spatial averaging. This is most likely due to background modelling errors in combination with the measurement geometry, and with an amplification of errors which is inherent to the problem of converting potential change at altitude into surface mass change.

Spatial averaging is usually implemented by isotropically or anisotropically filtering the coefficients, i.e. forming linear combinations of the original coefficients. A number of strategies have been proposed for finding weighting factors. The same effect can be achieved through regularizing the solutions on the level of normal equations.

We will review these techniques, and the new products that were generated by their application, for GRACE. We will consider their potential application to products that might be derived from GRACE-FO missions.