Inferred basal friction and surface mass balance of North-East Greenland Ice Stream using data assimilation of ICESat-1 surface altimetry and ISSM

Abstract

We present a new data assimilation method within the ISSM framework that is capable of assimilating surface altimetry data from missions such as ICESat-1 into reconstructions of transient ice flow. The new method relies on algorithmic differentiation to compute gradients of diagnostics with respect to model forcings. It is applied to the North East Greenland Ice Stream where surface mass balance and basal friction forcings are temporally inverted, resulting in significantly improved modeled surface heights that match existing altimetry. This new approach allows for a better quantification of basal and surface processes, and a better understanding of the physical processes currently missing in transient ice flow models to better capture the important intra and inter-annual variability in surface altimetry. It also demonstrates that large spatial and temporal variability is required in model forcings such as surface mass balance and basal friction, variability that can only be explained by including more complex processes such as snowpack compaction at the surface and basal hydrology at the bottom of the ice sheet. This approach is indeed a first step towards assimilating the wealth of high spatial resolution altimetry data available from EnviSat, ICESat-1, Operation IceBridge and CryoSat-2, and that will be available in the near future with the launch of ICESat-2.