Firn Modelling and its Effect on Satellite Derived Basal Melt Rates

Abstract

Basal melt, mass loss due to ocean-induced melting of floating ice shelves, is one of the main contributors to Antarctic mass loss. Basal melt rates can be measured in situ, or through remote sensing with satellite altimetry measurements. The satellite-derived basal melt rates have a much larger temporal variability. Long-term melt rates of the two methods also disagree in sign at several measurement sites. A possible large source of the uncertainty in the satellite-derived basal melt rates is due to incorrect modelling of densification processes, which stem from firn densification models. In this thesis, basal melt rate time series are recalculated for the output of two other firn densification models to assess the uncertainty due to these models. This results in spatial variability in average basal melt rate changes across ice shelves. 3-monthly basal melt rates and the 1-year running means of basal melt rates are also affected significantly. The calculated changes in the average basal melt rates are also compared to a newer dataset of 3-monthly basal melt rates, which showed much larger changes than what can be explained by the uncertainties due to firn densification models alone. This implies that perhaps other factors, like the velocity product and the reference frame, come with larger uncertainties than FAC corrections.