Hydrological de-aliasing model data for GRACE/-FO gravity field processing

Abstract

Temporal aliasing artefacts arising from unaccounted high-frequency mass variations in the geophysical fluid layers of the Earth are one of the most important error sources in present-day GRACE/-FO monthly-mean gravity field time-series. While non-tidal atmosphere-ocean signals are routinely corrected with the AOD1B product, there is no commonly agreed strategy to cope with aliasing effects from rapid changes in terrestrial water storage (TWS).In this contribution, we investigate the impact of introducing hydrological background model data into GRACE/-FO processing on the quality of the resulting monthly solutions. For the hydrological mass signal, we use sub-monthly TWS output from the recently improved open-source global hydrological model OS LISFLOOD, which we introduce into the ITSG-Grace processing scheme operated at TU Graz. We evaluate the effects of (i) deterministic information about TWS variability only, and (ii) additionally stochastic information on the remaining uncertainties in the modelled TWS data, which are obtained by performing a small ensemble of OS LISFLOOD model runs forced with different atmospheric reanalysis data.Even though temporal aliasing artefacts associated with TWS variability are substantially smaller than those originating from atmosphere and ocean, we conclude that the incorporation of hydrological background model information as derived from OS LISFLOOD into GRACE/-FO processing is advantageous for the quality of monthly solutions and should be considered to be included into the next major reprocessing of the satellite data record. With the provision of the new hydrological de-aliasing product at GFZ we establish the basis for this.