Novel insight into the spatiotemporal distribution of Greenland ice sheet surface densities from eleven years of satellite radar altimetry

Greenland Ice Sheet near-surface density is a key parameter as it facilitates converting changes in volume from repeat satellite altimetry to changes in mass. However, the need to continually quantify density through both space and time means it is conventionally derived using numerical models. Recently, new focus is being placed on directly deriving ice sheet surface properties from satellite observations. Here we show how 11 years of surface densities derived from ESA CryoSat-2 and CNES/ISRO SARAL radar altimetry compare against in situ and regional climate model results. We find that a consistent interpretation of the satellite near-surface densities is strongly modulated by the presence of melt. For periods when the ice sheet surface can be considered frozen, satellite-based densities best represent in situ conditions at different intervals in the top 1.25 m. At the same time, while long-term pan-Greenland observation- and model-based mean densities are similar, the observations reveal spatial patterns absent in the models and consistently greater densities in the ice sheet interior. In months with strong melting, the satellite results return anomalously low surface densities, though accounting for associated changes in radar backscattering phenomena and stratigraphic interpretation are limited by the lack of contemporaneous in situ data. Overall, we establish a quantitative approach for interpreting novel densities derived exclusively from satellite observations and present a comparison to state-of-the-art model results.