Characterising ice slabs in firn using seismic full waveform inversion, a sensitivity study

The density structure of firn has implications for hydrological and climate modelling, and ice-shelf stability. The structure of firn can be evaluated from depth models of seismic velocity, widely obtained with Herglotz-Wiechert inversion (HWI), an approach that considers the slowness of refracted seismic arrivals. However, HWI is strictly appropriate only for steady-state firn profiles and the inversion accuracy can be compromised where firn contains ice layers. In these cases, full waveform inversion (FWI) may yield more success than HWI. FWI extends HWI capabilities by considering the full seismic waveform and incorporates reflected arrivals. Using synthetic firn density profiles, assuming both steady- and non-steady-state accumulation, we show that FWI outperforms HWI for detecting ice slab boundaries (5-80 m thick, 5-80 m deep) and velocity anomalies within firn. FWI can detect slabs thicker than one wavelength (here, 20 m, assuming a maximum frequency of 60 Hz) but requires the starting velocity model to be accurate to +/- 2.5%. We recommend for field practice that the shallowest layers of velocity models are constrained with ground-truth data. Nonetheless, FWI shows advantages over established methods, and should be considered when the characterisation of firn ice slabs is the goal of the seismic survey.

Automated summary

The density structure of firn plays a crucial role in hydrological and climate modelling, as well as ice-shelf stability. Traditional methods like Herglotz-Wiechert inversion (HWI) can evaluate firn structure by analyzing depth models of seismic velocity. However, these methods are only suitable for steady-state firn profiles and may encounter difficulties when there are ice layers present. Full waveform inversion (FWI) overcomes these limitations and shows better performance in detecting ice slab boundaries and velocity anomalies within firn.