EO4GHRO: A MULTI-SENSOR SYNTHESIS FOR THE SPATIOTEMPORAL QUANTIFICATION OF NEAR-SURFACE DENSITY ACROSS THE GREENLAND ICE SHEET

Living Planet Fellowship research project carried out by Kirk Scanlan.
Expected sea-level rise through the remainder of the 21st century has been primarily attributed to the continued melting of the Greenland Ice Sheet (GrIS). This loss of ice not only directly threatens coastal infrastructure around Europe but precipitates second-order effects such as proliferation of diseases, crop instability and increased health stressors. Therefore, there is a necessary and increasingly urgent need to improve our ability to monitor and project the evolution of the GrIS through both time and space. Near-surface density is the means through which spaceborne radar altimetry-derived changes in the surface elevation of the Greenland Ice Sheet (GrIS) are converted to changes in ice mass; the loss of which then contributes to global sea-level rise. Conventional GrIS surface density estimates are derived using numerical regional climate models, whose outputs (e.g., precipitation, temperature, etc.) serve as inputs in firn densification models. These numerical models underlie both calculations of current mass losses and associated sea-level rise from the GrIS (i.e., those observed during the satellite era) as well as projected future mass losses in the face of an ever-warming climate. As such, uncertainty in the nearsurface density of the GrIS directly contributes to uncertainty in projected global sea-level rise. While validated using individual in situ point measurements, there is currently no pan-GrIS observational timeseries against which the modelled near-surface density structure of the GrIS can be compared. The purpose of this work is to fill this fundamental observational gap using novel data analysis algorithms and synthesizing data from multiple European Earth Observation (EO) assets. Timeseries of spatiotemporal changes in the near-surface GrIS dielectric properties will be estimated through the quantitative analysis of Ku-band ESA CryoSat-2 and EC Copernicus Sentinel-3 as well as Kaband CNES/ISRO SARAL radar altimetry data products. These results will then pre-condition the inversion of ESA SMOS passive radiometry measurements in order to produce a final, synthesized, quantitative timeseries of near-surface density across the GrIS. Airborne ESA CryoVEx radar altimetry and swath LiDAR data over the GrIS will be used to validate the joint recovery of both surface roughness and density from radar altimetry and in situ density measurements will be leveraged for both calibration and validation efforts. The state of the European spaceborne EO infrastructure has never been as sophisticated and comprehensive as it is today. And while there are more missions/instruments collecting more data than ever before, the synergistic analysis of these data remains under-developed. This research will synthesize a decade’s worth of EO data in order to produce a new observational dataset aimed at addressing a primary source of uncertainty in projections of global sea-level rise due to melting from the GrIS. Increasing confidence in future sea-level rise estimates will enable more robust assessments of coastal infrastructure vulnerabilities as well as the development, review and refinement of adaption and mitigation measures.

Project outcomes

The main outcome from my time as a LPF has been greater insight into the derivation of Greenland Ice Sheet surface properties (namely surface roughness and density) from Earth Observation datasets. What I discovered from comparing surface roughness derived from both airborne (ESA CryoVEx) and satellite (ESA CryoSat-2, ISRO/CNES SARAL/AltiKa, and NASA ICESat-2) altimetry was how sensitive they both are to the scale over which the roughness is quantified. Scale-dependent surface roughness is not something typically considered when predicting the long-term behavior of the Greenland Ice Sheet and how it’s surface will evolve under a warming climate; an important point considering surface melting contributes roughly half the mass loss feeding sea-level rise. Now that a framework exists for characterizing this scale-dependency across the entire ice sheet from observations, the opportunity exists to revisit some of these underlying assumptions and make Greenland Ice Sheet mass loss projections more robust. Simultaneously, I investigated how the density of the Greenland Ice Sheet surface (another important parameter in ice sheet mass balance calculations) can be derived from the synthesis of ESA SMOS passive microwave and ESA CryoSat-2 and ISOR/CNES SARAL/AltiKa radar altimetry datasets. What I discovered was a complex

situation where ESA SMOS brightness temperatures are more strongly impacted by the density heterogeneity (i.e., layering) in the near-surface than the actual densities themselves. By fixing the very near-surface conditions with densities derived from ESA CryoSat-2 and ISRO/CNES SARAL/AltiKa, I was able to develop a proxy metric for the degree of vertical heterogeneity in the Greenland Ice Sheet firn layer. After validating the approach with in situ density measurements, applying the technique to the full ice sheet and for the period 2013-2023 revealed complex spatiotemporal patterns of firn rehabilitation (associated with increasing the firn layer’s capacity to internally store melt) and heterogenization (associated with promoting run-off contributing to sea-level rise). Firn rehabilitation or heterogenization is linked to weather patterns and there is now have a way to reliably characterize the impact of atmospheric forcing on the Greenland Ice Sheet near-surface in across the full ice sheet based solely on observations.